Research Archive Citation for published version: Dennis Collopy, ‘Social media’s impact on intellectual property rights’, in Peggy E. Chaudhry, ed., Handbook of Research on Counterfeiting and Illicit Trade (Cheltenham: Edward Elgar Publishing, 2017) DOI: https://doi.org/10.4337/9781785366451 Document Version: This is the Accepted Manuscript version. The version in the University of Hertfordshire Research Archive may differ from the final published version. Copyright and Reuse: © 2017 Peggy E. Chaudhry Content in the UH Research Archive is made available for personal research, educational, and non-commercial purposes only. Unless otherwise stated, all content is protected by copyright, and in the absence of an open license, permissions for further re-use should be sought from the publisher, the author, or other copyright holder. Enquiries If you believe this document infringes copyright, please contact Research & Scholarly Communications at rsc@herts.ac.uk Social Media’s impact on Intellectual Property Rights 1. Introduction and background The pervasive nature and impact of social media on modern life is widely recognised especially what Rifkind 1 describes as its “open, universal, and transparent nature “, which has created essentially a “battle of paradigms” with governments “ caught in the middle attempting to serve two masters, one dedicated to a capitalist model and the other to a commons model”. It should not be surprising that social media has had an effect on Intellectual Property Rights (IPR) although there has until now been very little research around this effect. The UK’s Intellectual Property Office (IPO) commissioned research in 2015, by a team led by the University of Hertfordshire on the impact of social media on IPR, specifically to assess the ways in which social media platforms might facilitate IPR infringement in relation to physical goods (‘counterfeits’). The range of trade bodies and sectors involved in the research was shaped by those whose goods were most widely impacted by the availability of illicit goods through social media. Discussions with key UK trade bodies 2 helped determine the sectors most impacted to enable the appropriate level of focus and supporting data required. This paper is a summary of the key observations and findings emanating from that stud; it highlights the challenges involved in researching a contested market with divergent motives apparent across all the sectors involved. 2. Aims, Objective, methodology and research structure 2.1 Aims and Objectives Employing the required methodology for assessing the extent of social media’s effect on IPR in physical goods meant there were two key aims; firstly, to compare data and insights from industry, government and consumers to produce a representation of recent levels of counterfeiting within the UK and secondly to assess the extent to which this kind of illicit behaviour is moving online and is being facilitated by online social media platforms. The more specific objectives of the study involved assessments of the scale, impact and characteristics of infringements, as well as opportunities for IPR. The scale of infringement entailed assessing the extent to which social media facilitated infringement and how infringement was distributed between different sectors, products and types of IPR. The impact of infringement included assessing the likely costs (and benefits) to IPR-intensive businesses of social media facilitated infringement and indications of how social media has changed how IPR is used, promoted and enforced. This called for an examination of the impact on IPR intensive businesses’ reputations and the health and safety of consumers as well as an analysis of the scale and influence of closed groups (i.e. invite-only groups, created on social media platforms) on IPR infringement. The characteristics of 1 Rifkin, J. (2014) “The Zero Marginal Cost Society” Palgrave Macmillan New York Page 259 2 Alliance for Intellectual Property, Federation Against Copyright Theft (FACT) and the Anti-Counterfeiting Group (ACG) 1 infringement aimed to increase understanding of the placement of IPR-infringing content (closed groups, adverts on social media pages and links to sites/proxies) and highlight the types of infringing products provided and the formats they are provided in. The final specific objective, opportunities for IPR, was intended to include an exploration of current initiatives used to counter infringement, enforce rights, promote respect for IP, and assess the effectiveness of these initiatives. 2.2 Methodology The methodology recommended in our 2014 study ‘Measuring Infringement of Intellectual Property Rights’, called for a “multi-tiered but blended approach” involving the collation of data from multiple sources, including government and industry, as well as fresh data on consumer attitudes and behaviours.3 This methodology was designed to create a data set for measuring counterfeit trade by counting industry and customs seizures, as well as by deriving statistics from consumer, producer, distributor, and, where possible, retailer surveys, along with mystery shopping. The focus of the 2015-16 study was on the distribution and sale of counterfeit goods through social media and we felt it appropriate to prioritise industry and government seizure and enforcement data along with consumer surveys. We also anticipated securing further intelligence data on illicit activities on social media from industry and various government enforcement agencies. The most novel element of the 2015-16 methodology was creating an online tracking process to benchmark data from the consumer survey. 2.3 Research Structure The research was structured into three parts 1. Literature review/government and industry data This was intended to produce a benchmarked summary of the emerging identified trends and comprised a literature review, a survey of businesses and the collation of recent industry and government seizure statistics. We aimed to provide a a top-down as well as a bottom-up approach to government and industry sources, the latter involving interviews with key representatives in each impacted sector, attending industry events and liaison with the UK’s Intellectual Property Office’s (IPO) intelligence hub. Initially we aimed to find data sources covering periods of time sufficient to reveal meaningful trends to evidence claimed increases in counterfeited and pirated goods being sold as a direct consequence of consumers and retailers moving online. The structured surveys with a number of trade body members during the project focused on the key sectors identified from the official data set6 and those most impacted by counterfeiting according to industry figures. The choice of sectors involved comparing the different methods used within the 2014 UK ‘IP Crime Report’, sectors suggested by the trade bodies and also those identified by Google search 3 ‘Measuring Infringement Of Intellectual Property Rights – Executive Summary’, page 3: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/325020/IP_Measuring_Infringement.pd f 6 Such as the ‘IP Crime Reports’ for 2014 and 2015. 2 data. The final element of the industry surveys involved the three main technology platforms Google, Facebook and Twitter, to elicit their responses to the issues raised by industry and government agencies. Phase 2: Tracker methodology and consumer survey Tracker methodology The online monitoring process involved tracking six sectors with two products within each chosen sector using the above method as well as utilising data from the structured business interviews conducted with organisations from the same sectors. We developed an appropriate infrastructure for sampling, tracked a bulk process and then sampled a proportion of the references to follow links to determine their nature and what proportion of links were directed towards either legitimate or infringing resources. This enabled a scale up of the sample to the size of the bulk processing. Tracker tasks The team designed and developed monitoring agents to locate, track and trace the possible infringement of physical products on Facebook, Twitter and Google.7 A database was developed to support the monitoring agents and their results, along with a website to configure and monitor the agents and review results. The tasks included the development and implementation of the sampling structure for assessing the types of sources promoted via social media and, secondly, the analysis and scale estimation of the impact of social media on infringing material. The development, testing and production implementation work was spread over a six-week period. Consumer survey (3,000 respondents) This element was conducted online and complemented by an offline survey to estimate the scale robustly. The focus of the survey was on the proportion of social media-triggered purchases that were ‘at risk’ of being infringing, by working from the opposite direction. This meant tracking, through the survey, the proportion of those actively engaging with social media and using these insights to seek recommendations and respond to opportunities. Once we identified the size of this group, we asked them to recall the last purchase that they made that had been directed by a recommendation on social media and ask how confident they were that the purchase was legitimate. If they claimed to be confident, we asked what methods they used to gain confidence, what led them to make the purchase, and how many times they had declined to purchase anything because they thought there was an issue with IPR. This involved a series of questions ‘funnelling’ down to an individual influenced by social media when making purchase decisions, including which social media platforms they used, and then four/five questions that covered their experience of and level of sensitivity to IPR infringement. We aimed to check whether the purchase was in one of the’ target-selected sectors within the online tracker 7 For the sake of clarity, Google was included as the most likely entry point for the discovery of counterfeit or pirated goods. 3 Phase 3: Economic assessment We considered a methodology based around separate interviews with different industries to develop a more nuanced approach to levels of piracy and counterfeiting around ‘hit’ mid-range successful products. The industry survey included questions designed to assess this aspect of their market. While concerned about repeating the exercise started by Rand in 2012,8 we also aimed to develop a methodology that industry would both comprehend and support. However, given the nature of the products chosen for the tracker and consumer survey, we recognise that assessing the impact on ‘hit’ products might prove too cumbersome for firms. Nonetheless, we aimed to ask industry respondents this question during our structured surveys. At the outset, we hoped to be able to focus on the expected performance of the products monitored to provide an additional counterfactual position and enhance analysis of product performance versus expectations, and compare performance to industry norms or averages. Given the reluctance of industry to share the kind of insight needed, the economic assessment focused on understanding the direct and indirect economic harm of counterfeiting (on industry, government and consumers), as well as the role of social media and other online platforms. The assessment aimed to provide insights into some of our research objectives, including the impact of social media (particularly on producers’ reputations) and the characteristics of infringement (such as how certain social media channels are used over others/how particular sectors or goods are targeted). In addition, given the limitations of the forecast approach, we anticipated examining other methods to assess the impact of social media on levels of counterfeiting and relating this to the primary (non-obvious copy) and secondary (obvious copy) markets. 3.0 Initial Observations 3.1 Industry Insights and Views From the 2014 Study At the outset, we noted a number of pertinent key findings from two industry stakeholders featured in our previous research. Anti-Counterfeiting Group (ACG) Our review of their literature and interviews with their staff revealed key barriers to developing an understanding of the full scale of counterfeiting, which included a lack of cross-sector measurement along with reluctance from individual companies to share their in-house data (including seizure stats, online takedowns, and annual cost of enforcement). It was apparent that the industry focuses on day-to-day anticounterfeiting tools but is looking for more support and reporting from lawenforcement agencies. Federation Against Copyright Theft (FACT) 8 We expressed doubt on whether the Rand Model would meet industry’s needs and expectations in ‘Measuring Infringement Of Intellectual Property Rights’, pages 55–57: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/325020/IP_Measuring_Infringement.pd f 4 This enforcement agency acts for the UK entertainment industries and uses a twin approach to assess the current state of the market, which includes their annual IPSOS survey plus day-to-day infringement intelligence. FACT were vocal about the increasing move of counterfeit sales from offline to online, although this was mainly seen as relating to the dissemination of such goods on Amazon and eBay. While offline counterfeiting had diminished in importance in their planning, they were the first to mention encryption as a hurdle to monitoring online infringement. Another factor was the authentic feel of many counterfeit sites, including fake ads purporting to be from brand holders, which is a significant contributor to the levels of deception alleged by certain brands and private IPR enforcement agencies. 3.2 Previous Distinctive Features We noted several distinctive features from our 2013 study in both the industry and government literature that are relevant here. Industry literature This revealed a reliance on consumer surveys and recommended the adoption of the Centre for Economics and Business Research’s (CEBR( 2000 ‘omnibus approach’, although even that has limitations. The earlier review showed industry focusing on the broader issues of counterfeiting and piracy, as well as the standard measurements of counterfeiting activity post-TRIPs,9 based on customs seizures. MarkMonitor’s approach was the best of the industry research, even with their skewed sample. We saw this as being suitable for individual brands, even if not for an overall assessment of the market. There was little, if anything, within the literature at the time about the role of social media; the focus was instead on eBay and other online auction houses. Government and academic literature This was heavily characterised by reliance on customs seizures data even though this was, at best, a mere fraction of the total and open to bias. The OECD had correlated industry and seizure data but suggested combining more objective and robust methodologies. Crucially, they argued that there is little systematic collection and evaluation of data accompanied by on-going reliance on anecdotal and fragmentary information. The OECD advocated a mixed approach based on hard data from government, industry and consumers, and this underpinned our methodology. The EU’s Office for Harmonisation in the Internal Market (OHIM) also criticised most studies as ‘snapshots’, an opinion relevant to our belief in the need for a longer-term view of the market. Rand’s (2012) assessment of this issue included four solid approaches, but their fifth approach (an economic model) was based on what we felt was an unreliable data set, not least for industries launching new products. This was an important consideration for the economic assessment part of our methodology. 3.3 Initial Observations and Assumptions 9 The common abbreviation for the World Trade Organization’s Agreement on Trade-Related Aspects of Intellectual Property Rights: https://www.wto.org/english/tratop_e/trips_e/intel2_e.htm 5 Following our initial discussions with industry bodies and official agencies, we were made aware that certain social media platforms were producing new offerings (‘store’ and ‘buy and sell’) that had the potential to exacerbate the sale and distribution of counterfeit goods. We noted the distinction between the business models of online platforms (e.g. Google and eBay/Amazon Marketplace) and social media platforms (Facebook, Twitter) and the different challenges in monitoring the online behaviour of buyers and sellers, especially within ‘closed groups’. The initial contact with the three social media/online platforms, Google, Facebook and Twitter was positive and encouraging, but we soon became aware that there were potential issues involved in our dealings with them, not least following the Anderson report on surveillance,10 as well as online grooming 11 issues, all of which we believed might impact the cooperation of Google, Facebook and Twitter. 4.0 Research Outcomes 5.1. Scale of Infringement This section aims to assess the extent to which social media is used to promote IPR and the extent to which it facilitates infringement. Significantly, it also aims to enable an understanding of how infringement is distributed between different sectors, products and types of IP. To determine the landscape for assessing the scale of infringement, we have broken it down into reviews of industry (brands and trade bodies), government, academic, and technology industry literature. This is followed by the responses of industry, government enforcement agencies and technology firms to our survey questionnaires. 5.1.1. Literature and Media Review Industry sources Bryce and Rutter’s (2005) ‘Fake Nation’12 focused on the demand side of counterfeiting, and their report argued that consumption of fakes was commonplace in the UK. Even though the authors could not take account of the impact of the Internet at the time, their study influenced part of the design of our consumer survey.13 MarkMonitor’s (2012) ‘Shopping Report’14 claimed that 20% of bargainseeking shoppers were deceived into buying fake goods, and this was especially relevant to the UK consumer given the much higher rates of online purchases in the UK, which are ahead of Germany and France and double the European average of 10 Home Secretary on publication of the Anderson report: https://www.gov.uk/government/speeches/homesecretary-on-publication-of-the-anderson-report 11 ‘The use of social media for online radicalisation’: https://www.gov.uk/government/publications/the-use-of-socialmedia-for-online-radicalisation . This gave guidance to schools on How social media is used to encourage travel to Syria and Iraq 12 Bryce, J. and Rutter, J. (2005) ‘Fake Nation – A Study into Everyday Crime’ [Online]: https://www.academia.edu/597794/Fake_Nation_A_Study_into_Everyday_Crime http://webarchive.nationalarchives.gov.uk/20140603093549/http://www.ipo.gov.uk/ipresearch-fake-2005.pdf 13 See 5.5 (main report) and Appendix 6.4. 14 MarkMonitor (2013) ‘Shopping Report 2012’ [Online]: https://www.markmonitor.com/download/report/MarkMonitor_Shopping_Report-2012.pdf 6 22%. Their more recent ‘Global Consumer Shopping Habits Survey’15 indicated that 24% of consumers had (willingly or unintentionally) bought a product online that turned out to be a fake. NetNames’ report16 on the cost of counterfeiting described the “extraordinary” growth of global counterfeiting and, in common with most private enforcement agencies their assessment relied on official government or industry trade body estimates. These included reports such as a recent industry study from the US Chamber of Commerce (USCC),17 which cited the headline data from the latest OECD/EUIPO report to highlight their $461 billion global estimate of the global counterfeiting market, which is more than double the prior estimate from 2005. The Global Intellectual Property Center (GIPC) for the USCC also released a ‘Measuring the Magnitude of Global Counterfeiting’18 report, which analysed the 38 individual economies that made up 85% of the world’s economy. This study claimed that customs authorities were only seizing a tiny fraction of the value of the total estimated counterfeits (as little as 2.5%), and pointed to the ‘dearth’ of seizure data,19 arguing that the scale of the global counterfeiting problem has significantly increased, “fuelled by the proliferation of Internet use and social media platforms”. The GIPC authors quoted Chaudhry and Zimmerman’s assertion that the actual scope of counterfeiting is not “fully known”, with current estimates ranging from $200 billion to over $1.7 trillion. These differences are apparently attributable to the varying approaches to counterfeiting adopted by authorities, as well as the paucity of reliable industry data and the diverse methodologies used to estimate the market. Another illustration of the divergent estimates comes from the 2011 US TV programme Trademark Counterfeiting,20 which quoted extensively from the International AntiCounterfeiting Coalition’s (IACC) ‘Get Real’ campaign, citing the “$600 billion per annum” problem. Despite all of the various data sources, including those referenced above, we could find no reliable industry sources that estimated the scale of IPR infringement related directly to social media. Another, more recent, source 21 highlighted problems allegedly affecting Amazon Marketplace (which accounts for 40% of Amazon’s unit sales), which has “morphed into the world’s largest flea market” following the firm’s efforts to “openly court Chinese manufacturers”. These efforts led to sales from Chinese-based sellers more than doubling in 2015, without, according to the report, the installation of the checks needed to cope with the ‘influx’ of counterfeits. The article suggested that the scale of the problem of social media is part of a much greater problem involving wider ecommerce platforms. This confirms claims made by certain private enforcement agencies, like Yellow Brand, about the dangers posed by the largest online marketplaces. 15 MarkMonitor (2015) ‘Global Consumer Shopping Habits Survey’ [Online]: https://www.markmonitor.com/download/report/MarkMonitor_Online_Barometer-2015.pdf 16 NetNames (2015) ‘Counting the Cost of Counterfeiting – A NetNames Report’ [Online]: http://www.netnames.com/blog/wp-content/uploads/2015/10/NetNames-Counterfeiting-Report2015_REVISEDFINAL.pdf 17 Elliot, M. (2016) ‘New numbers don’t lie: counterfeits pose a growing threat’. US Chamber of Commerce [Online]: https://www.uschamber.com/op-ed/new-numbers-don-t-lie-counterfeits-pose-growing-threat 18 GIPC/US Chamber of Commerce (2016) ‘Measuring the Magnitude of Global Counterfeiting: Creation of a Contemporary Global Measure of Physical Counterfeiting’. 19 ibid., page 4. 20 https://www.youtube.com/watch?v=EnxY0a9Hin8 (25 August 2011) 21 Levy, A. (2016) ‘Amazon’s Chinese counterfeit problem is getting worse’ [Online]: http://www.cnbc.com/2016/07/08/amazons-chinese-counterfeit-problem-is-getting-worse.html 7 Government agencies International data Over the past five years, the two most widely used estimates of the global value of the counterfeit goods trade have been BASCAP (2011)22 and OECD (2009),23 with the former heavily reliant on the latter. The US’’s Government Accountability Office (GAO) 2010 report 24 noted that several official statistics used in the US were of uncertain provenance and, in the absence of a single official measurement, there was a tendency for global estimates and ‘rules of thumb’ to be employed when trying to assess overall levels of counterfeiting and piracy. Such estimates ranged between 1.8% (OECD) of legitimate global trade, which was the most frequently quoted measure, and 7% of legitimate global trade (International Chamber of Commerce ICC). BASCAP/Frontier 25 argued that the OECD’s 2008–2009 estimate only related to international trade and did not include domestically manufactured fakes or digitally pirated goods, the broader economic effects and employment losses. The two other segments of the global market and the associated economic losses were integral to BASCAP’s projected estimate of the global value of counterfeiting and piracy as standing between $1.22 trillion and $1.77 trillion. The higher figure is the one most widely cited by industry and the government enforcement community 26. The US GAO 2010 study27 identified the different criteria used for almost every sector, which make it almost impossible to arrive at one single figure to accurately measure counterfeiting. The GAO study suggested that this explained the reliance on anecdotal measures and ‘rule of thumb’, highlighting the difference between the International Trade Council’s (ITC) claims that counterfeiting and piracy accounted for 5% to 7% of world trade and the OECD’s 2009 estimate of 1.95%. OECD and EUIPO’s recent report, ‘Trade in Counterfeit and Pirated Goods: Mapping the Economic Impact’,28 confirmed our concerns about the limited value of seizure data. While confirming the effects of the globalisation of value chains and the rapid growth of e-commerce in enabling the global distribution and sale of counterfeit goods, the authors noted the discrepancies between EU, US and global customs data sets and the rapid growth, between 2011 and 2013, of seized postal shipments across the globe29 (a result of increased e-commerce). However, the study also argued that the e-commerce market is ‘nuanced’, dynamic and industry-specific.30 The report used seizure data estimating that counterfeit and pirated products 22 ICC/BASCAP (2011) ‘Estimating the global economic and social impacts of counterfeiting and piracy’. February 2011 [Online]: http://www.iccwbo.org/Data/Documents/Bascap/Global-Impacts-Study-Full-Report 23 OECD (2009) ‘Magnitude of Counterfeiting and Piracy of Tangible Products: An Update’ [Online]: http://www.oecd.org/industry/ind/44088872.pdf 24 U.S. General Accounting Office/GAO (2010) ‘Observations on Efforts to Quantify the Economic Effects of Counterfeit and Pirated Goods’ [Online]: http://www.gao.gov/products/GAO-10-423 25 ICC/BASCAP (2009) ‘The Impact of Counterfeiting on Governments and Consumers’ [Online]: http://www.iccwbo.org/advocacy-codes-and-rules/bascap/bascap-research/economic-impacts/ 26 WCO at ACG Conference, October 2015. 27 US GAO (2010) ‘Observations on Efforts to Quantify the Economic Effects of Counterfeit and Pirated Goods’ [Online]: http://www.gao.gov/products/GAO-10-423 28 OECD and EUIPO (2016) Trade in Counterfeit and Pirated Goods: Mapping the Economic Impact. OECD Publishing: Paris. DOI: http://dx.doi.org/10.1787/9789264252653-en 29 Ibid., page 55. 30 Ibid., page 57. 8 accounted for $461 billion, or almost 2.5%, of world trade in 2013, but acknowledged that this does not include domestically produced and consumed counterfeit and pirated products.31 However, there is little here that provides any reliable indication of the scale of social media within this trade. OHIM has also endeavoured to use a forecast model to assess the likely size of counterfeiting and piracy in the EU (in excess of €1.1 trillion). However, at these global and continental levels there was no data we could locate that scoped the levels of counterfeit trade attributable to online platforms, especially social media. UK infringement data The IPO Crime Report,32 although a highly regarded assessment of counterfeiting and piracy, was still not able to provide a single figure to measure the overall market given the different methods used to assess impact and value. A significant challenge for our study was establishing a meaningful trend from the data made available, as much of what was included was based on snapshots (rather than measurements over time), or methodological issues undermined the data. With that said, the last two versions of the ‘IP Crime Report’ clearly identified the “growing threat from social media” and indicated that sales of counterfeit goods via social media rose by 15% in 2013–14. Yet the 2015 report also claimed that; “online sale of counterfeit items remains a significant problem […] [but] it has not increased significantly from 2013/14 after a significant increase in recent years”. This was a rare example of an estimate of the scale of IPR infringement attributable to social media. The Trading Standards (National Trading Standards Board) annual survey was a potentially useful metric. Its methodological issues, due to the response rates varying considerably across the survey since its inception, made meaningful comparisons difficult to justify and prevented a robust snapshot of the trend being developed. However, the survey yielded data on the most investigated products, which we were able to compare with Google search terms to support the choice of goods for the consumer survey and tracker parts of this research. EU infringement data The European Commission (EC)’s (2009–2014) ‘Report on EU customs enforcement of intellectual property rights – Results at the EU border’ (2008–2013)33 provided a data set on the scale of infringement, breaking down data for each member state and enabling a view of the trend over time. We established that there had been a noticeable recent decline in articles seized (and the domestic retail value of seizures), as well as a tapering off of cases at the EU border when compared to the period from 1999–2008. The decline in articles may have been a result of better enforcement, especially given increased cooperation between rights holders and customs authorities over the 31 Ibid., page 68. 32 IPO (2015) ‘IP Crime Report 2015’ [Online]: https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/461792/ip-crime-report-2014-15.pdf 33 European Commission (2014) ‘Report on EU customs enforcement of intellectual property rights – Results at the EU border (2008–2013)’ [Online]: http://ec.europa.eu/taxation_customs/resources/documents/customs/customs_controls/counterfeit_piracy/statistics/2 014_ipr_statistics_en.pdf 9 past 10 years. It is just as likely that it could reflect changes in counterfeiting traffic, not least an increase in the domestic manufacture of counterfeit goods. Claims that the decline in seizures indicates a drop in counterfeiting traffic are in contrast to assertions made by industry about the rise of illicit online traffic. As such, we believe that the decline in articles seized evidences the limitations of customs seizure levels as a reliable indication of the trend. OHIM reports assessed the impact of counterfeiting across different sectors and recently formed part of a programme developed with the OECD. We had concerns with the model adopted for several recent case studies (recorded music, cosmetics and clothing) given the reliance on the Rand 2012 methodology to estimate counterfeits as the difference between actual and forecast sales.35 The rest of their methodology appeared robust and used innovative approaches to assess the likelihood of counterfeits in different industry sectors. Their segmented approach had much to offer and echoed comments made by the US GAO report, which holds that “effects vary across industries”. Developing a data set that could more accurately assess the impact on individual sectors while still using a single methodology to measure the impact across the industry seemed to have considerable merit. There was limited data available from the ‘IP Crime Reports’ that we could use for this study, and there was no official data at a suitable scale to support claims that social media plays a significant part in facilitating IP infringement and the sale and distribution of counterfeit goods. The on-going reliance on customs seizures as the official measurement tool to estimate levels of counterfeiting was of little value when assessing claims about the part played by social media. The private IPR enforcement agencies that work directly for many of the brands were potential sources of relevant and current information and were likely to be able to provide the kinds of insights and data needed to evidence the scale and impact of social media on IPR. Industry suggested that there was little confidence that official data would reflect changes in fast-moving markets, least of all in real time. Official data is, by definition, out of date and the speed of changes in the online world demonstrates a need for more frequent and current research. Academic Bates 36 noted that EU counterfeiting data was limited to data on seizures and pointed to the role of compromised and complex supply chains, as well as the internet, in the fake drugs trade. Naim 37 highlighted the growth rate of the trade in fakes (eight times the rate of legitimate trade since the early 1990s) and how China has become the leading exporter of counterfeits. Chaudhry and Zimmerman’s 2009 work The Economics of Counterfeit Trade 38 criticised official data, tracing estimates of the size of the global counterfeit market since the early 1980s to conclude that it was still not clear “what the real magnitude is”, noting unclear metrics and data falling 35 This was extensively reviewed in our 2013 study ‘Measuring Infringement of Intellectual Property Rights’, pages 55–56. This may also explain our reluctance to employ this type of model in our economic assessment. 36 Bate, R. (2008) Making a Killing: The Deadly implications of the Counterfeit Drug Trade. AEI Press. 37 Naim, M. (2005) Illicit: how smugglers, traffickers and copycats are hijacking the global economy. IDB Cultural Centre, Anchor/Random Books. 38 Chaudhry, P.E. and Zimmerman, A. (2009) The Economics of Counterfeit Trade: Governments, Consumers, Pirates and Intellectual Property Rights. Springer Verlag: Berlin. 10 short of what was required for policy making. They suggested that the total global counterfeit market had a collective worth of around $500–600 billion, but argued against the use of customs seizures as indicators of counterfeiting levels, claiming that such proxies represent (at best) a tiny fraction of illicit activity. Such flaws in the supply-side data led Chaudhry et al. to call for more demand-side research, yet there is a paucity of such research, with most of the extant work reliant on convenience samples of consumers within single-country markets and few empirical studies being conducted across country markets.39 The kind of demand-focused research they envisaged would involve investigation of consumer behaviour to better understand what motivates people to purchase illicit goods (beyond the incentive of a low price). Academic research on counterfeiting trends carried out at Michigan State University A-CAPP (Center for Anti-Counterfeiting and Product Protection) included papers by Wilson and Sullivan40 that highlighted the complexity of counterfeiting, and the methodological shortcomings of both government and industry estimates. These also 41 featured comments from brands about the problems of measurement and issues with identifying the number of counterfeits. This was much harder for firms “with sectors across different parts of the supply chain” and was particularly difficult for multinational companies. Elsewhere,42 the same authors stated that while product counterfeiting “is a global problem that is a growing concern for consumers, government entities, law enforcement, and businesses”, current assessments of the nature and extent of the problem are generally unreliable and use methodologies with significant limitations. We did not include Instagram when this study was first designed, but we came across research via Turner43 that focused on Instagram. The new study ‘Social media 44 and luxury goods counterfeit’ indicated that 20% of the luxury goods tagged on Instagram were fakes. This was based on their examination of around 150,000 posts tagged with luxury good brand names, mostly from accounts based in China, Russia and Malaysia. Analysis of the fake Instagram accounts showed postings of more than 140,000 images, displayed to 700,000 followers, over a three-day period. Technology literature sources We looked for data to evidence the reach of social media, given the perception that it has become the dominant feature of the online world. Within 2015’s global population of 7.3 billion, there were 3.17 billion Internet users and 2.3 billion active social media users.45 The importance of social media for the business community was evidenced 39 This was, of course, before the more recent OHIM and EC studies on IP rights perceptions. 40 Wilson, J. and Sullivan, B. (2016) ‘Brand owner approaches to assessing the risk of product counterfeiting’. Journal of Brand Management, May 2016, Volume 23, Issue 3, pages 327–344. 41 Wilson, J. and Sullivan, B. (2016) Measuring Product Counterfeiting: Insights from Current Research and Practice. Michigan State University A-CAPP (Center for Anti-counterfeiting and Product Protection). 42 Wilson, J., Sullivan, B. and Hollis, M. (2016) ‘Measuring the “Unmeasurable”: Approaches to Assessing the Nature and Extent of Product Counterfeiting’. International Criminal Justice Review. 1057567716644766. 43 Turner, K. (2016) ‘That Chanel bag on your Instagram feed may not be a Chanel bag’. The Washington Post. Online]: https://www.washingtonpost.com/news/the-switch/wp/2016/05/26/that-chanel-bag-on-your-instagram-feedmay-not-be-a-chanel-bag/ 44 Stroppa, A., di Stefano, D. and Parrella, B. (2016) ‘Social media and luxury goods counterfeit: a growing concern for government, industry and consumers worldwide’. [Online]: https://www.washingtonpost.com/blogs/theswitch/files/2016/05/IG_A2016_ST2.pdf?tid=a_inl 45 Brandwatch (2015) ‘Amazing Social Media Statistics and Facts for 2016’ [Online]: https://www.brandwatch.com/2016/03/96-amazing-social-media-statistics-and-facts-for-2016/ 11 by almost all (91%) retail brands using two or more social media channels. Internet users, on average, use 5.54 social media accounts and there were 1.65 billion active mobile social accounts globally. Increasing numbers of firms used social media to advertise, with 38% planning to spend more than 20% of their total advertising budget on social media channels. Only 4% of the Fortune 500 companies engaged with their customers on Facebook, although 83% had a presence on Twitter. Consumers’ use of brands’ social media accounts showed that those aged 55–64 were more than twice as engaged with branded content than the 28 or younger age group. Significantly for brands’ online strategies, however, 96% of the people that discussed brands online did not actually follow profiles those brands owned.46 However, users were increasingly adopting multi-networking behaviours,47 and the average number (5.54) of social media accounts did not correlate to the active use of platforms, which was much lower (2.82). There were clear demographic trends indicating that the average number of accounts reduces as age increases. The 25– 34 demographic were most likely to use Facebook and was the most engaged set of social media users. While it is clear that social media has become a significant channel for marketing products and, as part of that transformation, there is a risk that counterfeit goods could be made available through social media platforms, the reach of these platforms has not extended to the whole population. A 2012 article, ‘Dark Social: We Have the Whole History of the Web Wrong’,48 challenged the notion that social media was in fact all-pervasive or represented all online social activity. Madrigal noted how email programs, instant messages, some mobile applications and moving from a secure site to a non-secure site were invisible to most social media analytic tools, referring to this as the ‘dark social’. The social networks we were familiar with represented around 43.5% of social traffic, yet this other ‘unmeasured dark net’ represented 56.5% of activity. In fact, Madrigal claimed that almost 70% of social referrals were ‘dark social’, with Facebook the next main source, achieving 20%, and Twitter just 6%. While this data from 2011–12 may not fully reflect online social activity in 2015– 16, it reminds us that social media does not represent all internet-based social engagement. The scale of the profound changes brought about by social media was highlighted in the Pew Center’s 2015 report ‘Social Media Usage: 2005–2015’.49 The report confirmed the US demographic picture over the past decade, noting that in 2015 65% of American adults, up from 2005’s 7%, used social networking sites. While confirming young adults as the most likely (90%) age group to use social media, the report noted that 35% (against just 2% in 2005) of those 65 and older now use social media. Social media habits within different age groups relevant to estimates of the 46 Smith, K. (2016) ‘Marketing: 96 Amazing Social Media Statistics and Facts for 2016’. BrandWatch [Online]: https://www.brandwatch.com/2016/03/96-amazing-social-media-statistics-and-facts-for-2016/ 47 Mander, J. (2016) ‘Internet users have average of 5.54 social media accounts’. Global Web Index [Online]: http://www.globalwebindex.net/blog/internet-users-have-average-of-5-social-media-accounts 48 Madrigal, A.C. (2012) ‘Dark Social: We Have the Whole History of the Web Wrong’. The Atlantic [Online]: http://www.theatlantic.com/technology/archive/2012/10/dark-social-we-have-the-whole-history-of-the-webwrong/263523/ 49 Perrin, A. (2015) ‘Social Media Usage: 2005–2015’. Pew Research Center [Online]: http://www.pewinternet.org/2015/10/08/social-networking-usage-2005-2015/ 12 scale of social media’s role in IPR infringements included indications that women and men used social media at similar rates and suggested that socioeconomic differences, such as higher education level and household income, were important. There were also consistent indications that ‘rural’ residents used social media less than those in suburban and urban communities. Ofcom’s April 2013 report titled ‘Adults Media Literacy50 was part of a series of media literacy trackers and quantitative surveys to have shed light on certain trends in internet usage and online attitudes since 2005. It identified significant traits, such as on-going smartphone growth and older users driving the increase in social networking. The report indicated declining trust in what users saw or read on social networking sites, and claimed that more Internet users were ‘critically aware’. It also underscored the potential influence of social media/networking on internet users’ behaviour, not least their willingness to treat such sites as the main gateway/platform for their online experience. The regular use of social networking sites was higher in younger age groups (such as 16–24), but in terms of socioeconomic categories those in classes DE and C1 were more frequent users than those in classes AB and C2. Facebook 51 was the sixth-most-valuable public company in the world, with more data about more users than almost any other company in history. It had 52 1.55 billion monthly active users globally in 2015, up 14% from a year earlier, with 1.39 billion now using the service on mobile devices. With somewhere between 30 and 33 million users in the UK, Facebook remained the default social networking site for almost all (96%) of the UK adults online.53 Brandwatch noted that 72% of all online adults visited Facebook at least once a month, but stated that there were 81 million fake Facebook profiles; additionally, of the 40 million active small business Facebook pages, only 2 million of them pay for advertising. The 62% of social logins to brands’ and publishers’ websites and apps evidenced Facebook’s central role in the online ecosystem.54 Having grown from a single website and app, Facebook has cemented its status in the social network order. It now includes WhatsApp (1 billion users), Instagram, Messenger (900 million users) and Groups; as such, it now runs four of the six biggest social media platforms. Instagram’s typical user profile was 90% under 35 years old, and it was the favourite platform of 32% of US teens.55 Facebook, in common with Google, has received an increasing number of requests for users’ personal data; according to Titcomb,56 such requests were up by 60% in 50 Ofcom (2013) ‘Adult Media Literacy’ [Online]: http://stakeholders.ofcom.org.uk/binaries/research/medialiteracy/adult-media-lit-13/2013_Adult_ML_Tracker.pdf 51 Economist (2016) ‘How to win friends and influence people’ [Online]: http://www.economist.com/news/briefing/21696507-social-network-has-turned-itself-one-worlds-most-influentialtechnology-giants 52 Abutaleb, Y. and Maan, L. (2015) ‘Facebook revenue, profit beat forecasts; shares hit all-time high’. Reuters [Online]: http://www.reuters.com/article/2015/11/05/us-facebook-results-idUSKCN0ST2VF20151105 53 McGrory, R. (2015) ‘Social Media Statistics for 2015’. Social Media ltd [Online]: http://www.rosemcgrory.co.uk/2015/01/06/uk-social-media-statistics-for-2015/ 54 Brandwatch (2015) ‘Amazing Social Media Statistics and Facts for 2016’ [Online]: https://www.brandwatch.com/2016/03/96-amazing-social-media-statistics-and-facts-for-2016/ 55 Abutaleb, Y. and Maan, L. (2015) ‘Facebook revenue, profit beat forecasts; shares hit all-time high’. Reuters [Online]: http://www.reuters.com/article/2015/11/05/us-facebook-results-idUSKCN0ST2VF20151105 56 James Titcomb (2015) ‘Facebook “snooping” requests increase 60 per cent in UK’. Telegraph, 11 November 2015 [Online]: http://www.telegraph.co.uk/technology/facebook/11989781/Facebook-snooping-requests-increase-60-percent-in-UK.html 13 the UK, with the social network fielding 3,384 demands in six months in 2015 from government and law enforcement authorities. Facebook57 claimed to have received more government requests from the UK than from any other countries besides the US and India, and complied with 78% of requests. 5.1.2. Industry and trade bodies’ responses An appeal at the ACG’s road shows for ‘real’ rather than anecdotal evidence demonstrated a fundamental problem facing the various sectors and brands, namely the lack of robust, current data to illustrate the scale of IPR infringements. The ACG’s call aimed to collate brands’ confidential counterfeiting information to share it with government agencies, and it reaffirmed a major challenge for our study in accessing this kind of industry-specific data. Entertainment industry enforcement agency FACT informed us that their current main focus was on dealing with online digital piracy, even though they were still active in the physical goods market, which they described as ‘dwindling’. Uniquely among ACG member responses, one firm claimed that social media accounted for around a fifth of all their IPR infringement. We gained some insight from various presentations from private enforcement agencies, including NetNames, which claimed that one in every five websites is fake and as many as 40,000 websites were compromised every week. According to them, this meant that, on average, at least 20% of a brand’s online traffic could be diverted away from its websites. NetNames also claimed that one in every six products sold online was counterfeit, and 30% of EU counterfeit seizures were linked to Internet distribution channels. China expert Yellow Brand Protection claimed that the counterfeiting industry accounted for 8% of China’s GDP. Apart from Alibaba, there are a number of Chinese sites, such as Makepolo, with global reach. Online marketplaces were apparently the #1 online sale channels, with more than 700 active online marketplaces on the Internet and 150 in China alone. However, none of these agencies gave specific data at scale on social media’s role within the fakes market. Their focus appeared to be on the wider issues of traffic diversion and website impersonation, where social media is just one aspect of a wider problem. 5.1.3. Government agency responses The data available from HM Revenue and Customs (HMRC) illustrated the limitations of relying on seizures. Despite cooperation with industry and other agencies, HMRC had a ‘hit’ rate of 6% on IPR-related consignments and over 50% of these contained misleading goods. The targeting of large-scale shipments also suggests that they have not been able to deal with the shifting pattern of delivery of counterfeit goods in small packages, evident from the EU border results between 2008 and 2014. Our efforts to review the long-term trends from the various ‘IP Crime Reports’ indicated that data was not always comparable and the IPO Intelligence Hub confirmed that the data set was unlikely to improve in the absence of statutory reporting on IP crime. According to the enforcement agencies, rights holders and Trading Standards members had complained about the frequency of sales of counterfeit goods on social media sites. 57 Facebook (2016) ‘UK Government Requests for Data’ [Online]: https://govtrequests.facebook.com/country/United%20Kingdom/ 14 Trading Standards acknowledged the problem of accurately measuring the scale of online illicit activity, not least due to disparate recording of data and intelligence. The situation is made worse by the significant underreporting of illicit activity (only 5– 10%), meaning anecdotal evidence still dominated. They relied on data from Citizens Advice,58 who handled complaints on ‘scams’ (including counterfeit goods), and passed on IPR-relevant cases to Trading Standards. The current data showed they have seen a marked increase in the scale of infringement on social media. Since 2010, there has been a 400% increase in complaints attributable to Facebook. Recent data from one region indicated that social media-related IPR infringements far exceeded those of the sale of infringing products on eBay. Intelligence data available via Trading Standards suggested that the sale of physical goods was a dominant factor on social media, with clothing, fashion accessories and DVDs forming the largest categories, accounting for approximately 60% of counterfeit sales. The remaining 40% was made up of footwear, electrical products, toys, toiletries and computer software. 5.1.4. Tech firm responses Google argued that only a small percentage of ‘bad actors’ misused legitimate online services to try to sell counterfeit goods. Google’s assertion about the low levels of complaints (a small fraction of 1% of advertisers in the past year) was a bold claim and one that called for a response from industry and enforcement agencies. At face value, it suggested that counterfeiting is not as big an issue for search engines as it might be for social media platforms. Twitter referred us to their ‘Transparency Report’, which details the number of requests received by them from government agencies, as well as industry IPR infringement notices. Their ‘Transparency Report’ on trademark notices for the six months ending 30 June 2015,59 and also December 2015,60 indicated that a relatively low proportion (6–11%) of accounts had been affected by alleged trademark violations. They noted in the most recent report that the number of trademark notices received for Twitter and Vine had, in fact, declined by 33% (8,588 versus the 12,911 in the previous January–June 2015 report). 5.2 Impact of Infringement This section aims to assess the costs and benefits to IP-intensive businesses of social media and indicate how social media has changed the ways IP is used, promoted and enforced. While there was the question of how IPR were being applied, the two key elements were: 1) examining the impact on IP-intensive businesses’ reputations and consumers; and 2) assessing the scale and influence of closed groups (i.e. invite-only groups, created on a social media platform) on IPR infringement. As before, this is broken down into a review of industry (brands and trade bodies), government, academic, and technology industry literature, followed by the responses of industry, government enforcement agencies and technology firms to our survey questionnaires. 58 Facebook is now second only to eBay in consumer complaints to CA, with trends suggesting that Facebook will soon overtake eBay. 59 Twitter (2015) ‘Trademark Transparency Report’. June 2015 [Online]: https://transparency.twitter.com/trademarknotices/2015/jan-jun 60 Twitter (2016) ‘Trademark Transparency Report’. December 2015 [Online]: https://transparency.twitter.com/trademark-notices/2015/jul-dec 15 5.2.1 Literature and media review Industry While brands and the ACG recognised the value of social media in promoting and marketing their products online, it was equally clear that they saw the dangers within the online ecosystem, and especially within social media. The focus of much of the industry literature was on harm to brand integrity and reputation. According to the ACG’s Lewis,61 if the products were poor or substandard, consumers were likely to “blame the brand”, a description echoed among much of the literature and industry sources. NetNames’ recent report,62 highlighting the cost of the “extraordinary” growth of global counterfeiting, noted that the pharmaceutical sector was most affected. GIPC described the negative economic effects on “consumers, trademark owners, companies […] and retailers, as well as the economy at large”. The impact of the lower quality ‘fake’ goods was to undermine brand integrity, reduce revenue, decrease market share, dilute and damage the brand, and require financial investment to cover the costs involved in enforcing their IPR.63 Another concern, voiced by MarkMonitor, was that the free speech environment online allowed for negative experiences to impact brands. The damage and harm from these sites came especially from fake special offers that spread very quickly and were not easily identified. The BBC’s Fake Britain (2014–2015) series focused on the impact on consumers of “easily set up dodgy websites and fake identities causing problems […] with lifethreatening consequences”. The qualitative insights available on the YouTube channel used to access the entire series included many critical consumer comments on the perceived excess profits made by brands on goods manufactured in Southeast Asia, suggesting that this justified purchasing counterfeit goods. This highlighted a significant problem for brands in combating counterfeit goods when the reputation of brands is so easily assailed on social media platforms. 2008’s The Fake Trade 64 described how a “new form of capitalism has emerged, bordering on anarchy,” one that impacted the popular brands because of the increased “appetite” for imitations of their products. According to the programme makers, unseen Chinese factory workers made global brands’ products as well as their competitors’ (i.e. ‘grey goods’), creating a major problem for enforcement. Such grey goods were shown to be an unintended consequence of the wide-scale outsourcing to China by “even the most expensive brands”. Government agencies The US GAO’s 2010 study65 criticised all existing and widely used estimates of economic losses that cannot be substantiated and critiqued the assumptions used for 61 Lewis, P. (2015) ‘Anti-Counterfeiting Group: fighting the good fight’. Trademarks and Brands Online [Online]: http://www.trademarksandbrandsonline.com/article/anti-counterfeiting-group-fighting-the-good-fight 62 NetNames (2015) ‘Counting the Cost of Counterfeiting – A NetNames Report’ [Online]: http://www.netnames.com/blog/wp-content/uploads/2015/10/NetNames-Counterfeiting-Report2015_REVISEDFINAL.pdf 63 GIPC/US Chamber of Commerce (2016) ‘Measuring the Magnitude of Global Counterfeiting: Creation of a Contemporary Global Measure of Physical Counterfeiting’. Page 10. 64 Hornby, N./Blast Films (2008) Fake Trade [Online]: http://topdocumentaryfilms.com/fake-trade/ 65 US GAO (2010) ‘Observations on Efforts to Quantify the Economic Effects of Counterfeit and Pirated Goods’ [Online]: http://www.gao.gov/products/GAO-10-423 16 substitution rates for fake, as opposed to legal, goods. Their report pointed to the broad range of effects on consumers, industry, government and the economy, arguing that the potential direct effects varied, with those on consumers being both negative and positive, whereas for industry the effect was mainly negative and the effects on government and the economy were entirely negative. The lack of data hindered efforts to quantify the economic impact of counterfeiting and piracy and this resulted in the use of assumptions to compensate, with most estimates being highly sensitive to the assumptions used (including the substitution rate, the value of fakes and the level of deception). The latter factor is of real significance to our study as it highlights the fight against counterfeits, which involves very different types of quality and levels of deception and impact, both on consumers and industry. The GAO claimed that no single approach could be used to quantify impact and drew attention to the difference between various estimates, especially those based on a ‘rule of thumb’. The EC’s ‘Impact Assessment’66 on OHIM and the EU Observatory in 2011 argued that the lack of reliable objective data within existing studies made it impossible to scope the impact of IPR infringement or measure trends. This assessment identified the main impacts of counterfeiting as reduced investment in innovation and jobs, threats to consumer health and safety, serious problems for SMEs and reduced tax revenues. OHIM’s 2015 ‘Economic Studies’67 identified four key impacts of counterfeiting as reduced sales by legitimate businesses, reductions in tax revenues, lower employment, and public and private costs of enforcement. OHIM’s two 2015 case studies on the ‘Economic Costs of IPR Infringement’ (on cosmetics 68 and subsequently on clothing and footwear) 69 reviewed the main direct and indirect costs. To assess direct costs to industry, the authors relied on Rand’s 2012 methodology (including the use of firm analysis), and to assess the indirect effects of counterfeiting and piracy they estimated how different sectors of the economy bought goods and services from each other for production. They also attempted to assess the impact on public finances. OHIM advocated a two-step approach involving estimated lost sales due to counterfeiting and piracy and then a translation of the lost sales into lost jobs and lost public revenue. The second study calculated that almost 10% of the EU market was made up of infringing goods and arrived at a headline figure of €26 billion in lost revenue, with €17 billion in losses to related sectors (indirect cost) and €8 billion in lost taxes. By 2016, the third report in this series featured an attempt to quantify music piracy; MusicAlly’s70 coverage of the (now renamed) EUIPO’s latest sectorial analysis71 was 66 European Commission (2011) ‘Commission Staff Working Paper: Impact Assessment on Proposal Regarding OHIM and European Observatory on Counterfeiting and Piracy’ [Online]: http://ec.europa.eu/internal_market/iprenforcement/docs/observatory/sec_2011_0612_en.pdf 67 OHIM (2015) ‘Economic Studies in the Observatory – Conference on Policy and Stakeholder Perspectives to Prevent Counterfeiting and Illicit Trade Brussels’. 16 June 2015. 68 OHIM (2015) ‘The economic cost of IPR infringement in the cosmetics sector’ [Online]: https://oami.europa.eu/ohimportal/documents/11370/80606/Quantification+of++IPR+infringement+cosmetics 69 OHIM (2015) ‘The Economic Cost of IPR Infringement in the Clothing, Footwear and Accessories Sector’ [Online]: https://oami.europa.eu/ohimportal/en/web/observatory/ip-infringements_clothing-accessories-footwear 70 MusicAlly (2016) ‘European Union report tries to quantify music piracy’ [Online]: http://musically.com/2016/05/25/european-union-report-tries-to-quantify-music-piracy/ 17 highly critical of the methodology. It questioned the reliance on the hypothetical additional revenue that the recorded music sector would have earned had infringement not taken place. Apparently, it neither estimated the value of illegally acquired music recordings nor the substitution effect, but the hypothetical estimates seemed less controversial than assessing the substitution effect of piracy. Even though the study looked at direct and indirect revenue losses, along with job losses and the impact on public finances, it only referred to production and did not consider the impact on distribution or retail. Academic The most significant academic source for understanding the impact of IPR infringement was Hopkins et al.’s Counterfeiting Exposed,72 which proposed a ‘harm matrix’ encompassing four different levels of deception and quality, from high deception/high quality (such as grey goods) all the way to low deception/low quality (e.g. cheap Rolex watches). It also identified the different ways that counterfeiting impacted brands, noting that fake luxury goods were common in seasonal fashion markets. The authors raised the question of what constituted a counterfeit product if the fake goods were made in the same factories as legal offerings, and this raised an important issue for enforcement as to whether resources are best deployed in contesting such ‘grey goods’ markets or focusing on those products that are most harmful. In relation to social media offerings, grey goods can involve both high and low deception but are always high quality and as such have a lower impact and level of harm than high-deception/low-quality goods, which appear to be the goods causing the greatest harm to consumers and potentially to industry, given the potential harm to brands’ reputations 5.2.2 Industry responses Among ACG members, the impact on brands of fakes sold on auction sites like Amazon and eBay was clear from the poor customer reviews resulting from bad experiences. There were claims that established brands, notably those that have to take down hundreds of listings a day, suffer the most. Newly launched goods were thought to be most likely to be impacted after a heavy ad campaign for a product launch, which would drive factories to make fakes more quickly. There was a clear sense that the impact of social media was increasing, with claims of platforms acting as shop windows for fakes, with Facebook in particular a favourite for counterfeit sellers. None of the firms surveyed were able to articulate the economic damage sustained from social media in relation to their IPR, regardless of how critical they had been of the social media platforms. The IP Crime Group’s Social Media Group discerned certain online behavioural changes, with one of these indicating that those buying counterfeits were not able (to afford) to buy the ‘real thing’; this is an issue we picked up on in the consumer tracker as typical of complicit consumers, whose purchases are of lower economic harm. 5.3. Characteristics of Infringement 71 EUIPO (2016) ‘The Economic Cost of IPR Infringement in the Recorded Music Industry’ [Online]: https://euipo.europa.eu/ohimportal/en/web/observatory/ipr_infringement_music 72 Hopkins, D.M., Kontik, L.T. and Turnage, M.T. (2003) Counterfeiting Exposed – Protecting Your Brand and Customers. John Wiley & Sons. 18 This section aims to evaluate the placement of IP-infringing content (closed groups, adverts on social media pages and links to sites/proxies), as well as increase the understanding of the types of infringing products provided and the format they are provided in. As before, it is broken down into a review of industry (brands and trade bodies), government, academic, and technology industry literature. This is followed by responses from industry, government enforcement agencies and technology firms to our survey questionnaires. 5.3.1 Literature and media review Industry The 2015 Canadian film Counterfeit Culture 73 challenged; “consumers to take a deeper look at what appear to be harmless knock-offs at bargain prices” and was highly critical of Facebook, noting the increased sophistication of counterfeiters on the platform. The 2015 Anti-Counterfeiting Guide 74 noted that the key issues for brands in the online world included the increased use of branded content to distribute counterfeit goods; fake profiles being used to impersonate legitimate brands; the use of copyrighted images to promote third-party services; and misrepresentation of official company communications. Counterfeit and grey market channels often went hand in hand, according to KPMG’s 2005 report.75 MarkMonitor’s ‘Shopping Report 2012’76 assessed e-commerce, with research covering over 5 million shopping sessions, to better understand consumers’ motivations in deliberate or inadvertent purchases of counterfeit goods. The report emphasised levels of deception on online platforms and claimed that most shoppers impacted by the deception had assumed that online commerce would be cheaper and a source of discounted prices. According to the ACG’s Lewis,77 UK consumers were the most regular online shoppers in the EU, but the market was distorted by the change of distribution from containers to small packages, along with the impact of social media, where photos of genuine goods were used alongside high-quality fake labels and packaging. These methods, combined with prices close to those of the real products, were all used to deceive consumers. The market has seen a further shift in consumer buying habits towards mobile platforms and smartphones, as noted by MarkMonitor,78 which further enable deceptive purchases. 73 Canadian Broadcast Company (2011) Counterfeit Culture [Online]: http://www.cbc.ca/doczone/episodes/counterfeit-culture 74 World Trademark Review (2015) ‘Global strategies for protecting brands in the expanded online space’ [Online]: http://www.worldtrademarkreview.com/Intelligence/Anti-counterfeiting/2015/Industry-insight/Global- strategies-forprotecting-brands-in-the-expanded-online-space 75 KPMG (2005) ‘Managing the Risks of Counterfeiting in the Information Technology Industry’ [Online]: http://www.agmaglobal.org/cms/uploads/whitePapers/KPMG-AGMA_ManagingRiskWhitePaper_V5.pdf 76 MarkMonitor (2013) ‘Shopping Report 2012’ [Online]: https://www.markmonitor.com/download/report/MarkMonitor_Shopping_Report-2012.pdf 77 Lewis, P. (2015) ‘Anti-Counterfeiting Group: fighting the good fight’. Trademarks and Brands Online [Online]: http://www.trademarksandbrandsonline.com/article/anti-counterfeiting-group-fighting-the-good-fight 78 MarkMonitor (2015) ‘Global Consumer Shopping Habits Survey’ [Online]: https://www.markmonitor.com/download/report/MarkMonitor_Online_Barometer-2015.pdf 19 NetNames79 questioned consumer attitudes, arguing that while a minority look for counterfeit goods, many more are deceived into buying them. This finding contradicts MarkMonitor’s claims and the findings in our consumer survey and tracker. Examining different consumer behaviours, Spire (2011) 80 segmented counterfeits into deceptive, non-deceptive and clones/duplicates, and the types of consumers of counterfeit goods into ‘happy purchasers’, ‘struggling consumers’, ‘Robin Hoods’ and ‘innocent purchasers’. Consumer motivation was also central to the conclusions of Bryce and Rutter’s (2005) ‘Fake Nation’81, notably that consumers’ motives for buying counterfeits were not solely based around economic costs and consumers readily distinguished between different types of products. Government agencies The 2011 EC impact assessment 82 cited survey data indicating that many EU citizens have knowingly acquired IPR-infringing goods, with 25% believing it acceptable to buy counterfeits and one in three feeling justified if the price of the legitimate product was too high. The authors noted greater compliance in the UK, Ireland and Denmark than in the rest of the EU. However, OHIM’s 2013 ‘EU Citizens and Intellectual Property’83 claimed that only 10% of EU citizens openly admitted to IPR infringement but 33% tolerated IPR-infringing behaviours. The report saw a disconnect between support for IPR and personal choices and listed a number of different reasons for this, including limited buying power and protests against the prevailing market economy or premium brands, but concluded that many see IPR as benefiting business elites rather than consumers. The OECD’s 2015 ‘Enquiries Into Intellectual Property’s Economic Impact’84 noted the convergence of services and the rapid rise of mobile broadband, which is now more widespread than fixed-line broadband. This confirms the increasing challenges that mobile broadband creates for IPR enforcement, particularly in countries with a shortage of fixed-line broadband. The EUIPO’s 2016 ‘Intellectual Property and Youth’85 study researched demand for IPR-infringing goods, using both qualitative research (28 focus groups with respondents aged between 15 and 24 years old) and quantitative research (online survey with 24,295 15–24 year olds from across the EU28). The report covered both 79 NetNames (2015) ‘Counting the Cost of Counterfeiting – A NetNames Report’ [Online]: http://www.netnames.com/blog/wp-content/uploads/2015/10/NetNames-Counterfeiting-Report2015_REVISEDFINAL.pdf 80 ACC/Spire Research and Consulting (2011) ‘Counterfeits: Measuring it, Fighting It’ [Online]: http://www.slideshare.net/spireresearch/010216the-american-chamber-of-commercecounterfeit-measuring-it-fightingit 81 Bryce, J. and Rutter, J. (2005) ‘Fake Nation – A Study into Everyday Crime’ [Online]: https://www.academia.edu/597794/Fake_Nation_A_Study_into_Everyday_Crime http://webarchive.nationalarchives.gov.uk/20140603093549/http://www.ipo.gov.uk/ipresearch-fake-2005.pdf 82 European Commission (2011) ‘Commission Staff Working Paper: Impact Assessment on Proposal Regarding OHIM and European Observatory on Counterfeiting and Piracy’ [Online]: http://ec.europa.eu/internal_market/iprenforcement/docs/observatory/sec_2011_0612_en.pdf 83 OHIM (2013) ‘European Citizens and Intellectual Property: Perceptions, Awareness and Behaviour’ [Online]: https://oami.europa.eu/ohimportal/documents/11370/80606/IP+perception+study https://oami.europa.eu/ohimportal/en/web/observatory/ip_perception 84 OECD (2015) ‘Enquiries Into Intellectual Property’s Economic Impact’ [Online]: http://www.oecd.org/sti/ieconomy/KBC2-IP.Final.pdf 85 European Union Intellectual Property Office (2016) ‘Intellectual Property and Youth Scoreboard 2016’ [Online]: https://euipo.europa.eu/ohimportal/en/web/observatory/ip-youth-scoreboard?platform=hootsuite 20 digital content and online purchases of physical goods; in relation to the latter, clothes and accessories (including footwear) were the most common products. Despite finding that 12% of young people had intentionally bought a counterfeit product online, the study claimed that counterfeit goods had a bad image among this age group. The study indicated marked differences in young consumers’ ability to positively identify websites offering counterfeit goods. There were significant differences detected between markets, with the UK very close to the EU average in most metrics, with only 10% intentionally purchasing counterfeit goods, 12% doing so unintentionally and the remaining 78% either not buying or not knowingly buying.86 Ofcom’s April 2013 ‘Adult Media Literacy87 data on website terms and conditions/privacy statements confirmed an emerging pattern on consumers’ acceptance of platform’s terms and conditions that is relevant to understanding online consumers’ behaviour. The Ofcom data indicated that Internet users were less likely to say they had read website terms and conditions thoroughly, and were more likely to say they had not read them at all. Ofcom’s 2014 follow-up report 88 showed further changes, with more internet users (79% versus 75% in 2013 and 66% in 2011) likely to say they did not read/only skim read a website’s terms and conditions. Ofcom’s data indicated that consumers rarely read and understand website terms and conditions and this has implications for the social media platforms’ insistence that their stated terms and conditions are suitable/sufficient to discourage IPR infringements. Academic Poddar et al.’s ‘Exploring the Robin Hood effect’ 89 argued that demand for luxury brands in turn fuels demand for counterfeits. While there were some deceived consumers for luxury/prestige goods, they were often willing accomplices whose social goals underscored their behaviour. The authors claimed that consumer morality was grounded in reasoned action, with moral and economic dimensions that influenced their willingness to purchase fakes. The authors’ ‘fairness heuristic theory’ correlated brands’ negative motives with unfavourable consumer responses, which they describe as the ‘Robin Hood’ effect, in which consumers look for moral equity. Morales 90 also argued that the likelihood of the purchase of a fake was linked to the price of the original as opposed to the copy if the quality was ‘sufficient’ and there was no loyalty to the company making the original. Chaudhry and Zimmerman’s The Economics of Counterfeit Trade 91 emphasised consumer complicity in the sale of counterfeit products and the need for brands to better align their actions with consumer beliefs. Consumer complicity was one of the 86 These figures are similar to the results of our consumer data (see section 5.5 of the main report and Appendix 6.5). 87 Ofcom (2013) ‘Adult Media Literacy’ [Online]: http://stakeholders.ofcom.org.uk/binaries/research/medialiteracy/adult-media-lit-13/2013_Adult_ML_Tracker.pdf 88 Ofcom (2014) ‘Adult Media Literacy’ [Online]: http://stakeholders.ofcom.org.uk/market-dataresearch/other/research-publications/adults/media-lit-10years/ 89 Poddar, A., Foreman, J., Banerjee, S. and Scholder, E.P. (2011) ‘Exploring the Robin Hood effect: Moral profiteering motives for purchasing counterfeit products’. Journal of Business Research, Volume 65, Issue 10, October 2012, pages 1,500–1,506, JBR-07390 Elsevier. 90 Morales, A.C. (2005) ‘Giving firms an “E” for effort: consumer responses to high-effort firms’. Journal of Consumer Research, 2005, 31, pages 806–12. 91 Chaudhry, P.E. and Zimmerman, A. (2009) The Economics of Counterfeit Trade: Governments, Consumers, Pirates and Intellectual Property Rights. Springer Verlag: Berlin. 21 key reasons for the growth in counterfeit trade, with evidence suggesting that consumers were “all too willing to purchase counterfeit products even when they know the products are fake”. The internet was seen as “an outstanding opportunity for pirated product[s]”, allowing counterfeiters “to reach a nearly unlimited worldwide audience” and enabling the availability of almost “every type of product […] across the Internet”.92 The authors also pointed to the rise of powerful worldwide brands made possible by globalisation, with far greater numbers of consumers familiar with and interested in the brands but many unable to afford the real thing. The authors modelled consumer complicity as a function of both intrinsic (demographics, attitude towards counterfeits, cultural values and ethical perspectives) and extrinsic (social marketing communications and shopping) factors. This model is of value in establishing influences on consumer behaviour, with the extrinsic factors relevant to assessing the role of social media in facilitating IPR infringement. This is further explored in Appendix 6.6. Raustilia and Sprigman’s ‘The Piracy Paradox: Innovation and Intellectual Property in Fashion Design’93 pointed to an over-reliance on trademark law given that much of the wide-scale IPR infringement that took place actually involved design rights and copyright. The authors argued there was also an inherent induced obsolescence in fashion goods, with much of the value being status based – this is an aspect that has made them ideal for dissemination on social media. Naim94 pointed to the way brands infused desirability and attainment in their products, which made them easy targets for counterfeiters; the more successful a product was, the more likely it was to be faked. There has also been a rise in ‘super copies’, with Phillips95 arguing that no product is safe from being copied. Additionally, Prabhakar96 pointed to the widescale use of anonymity, which has made much of the illicit commerce untraceable. 97 Stroppa et al.’s ‘Social media and luxury goods counterfeit’ study claimed that social media was an important part of the counterfeiting market, enabling the fulfilment of illicit transactions off platform via online payment sites. The study identified how brands’ investment in Instagram to raise the profile of their products had created a ‘promo-friendly environment’ that attracted counterfeiters. The methods used by the counterfeiters on Instagram include the use of spambots, which enable them to manage thousands of accounts at the same time, as well as automatic account creation and postings of images that cannot be detected by Instagram’s systems. The nature of the infringement involved sites posing as legal discount sellers and the alleged international chain of activities confirmed claims made by various enforcement agencies. Stroppa et al warned that this global trend 92 Ibid., page 23 – they also cite OECD’s five major reasons for the attraction of the internet: anonymity, flexibility, market size, market reach and deception. 93 Raustiala, K. and Sprigman, C. (2006) ‘The Piracy Paradox: Innovation and Intellectual Property in Fashion Design’. Virginia Law Review, Volume 92, December 2006, No 8. 94 Naim, M. (2005) Illicit: how smugglers, traffickers and copycats are hijacking the global economy. IDB Cultural Centre, Anchor/Random Books. 95 Phillips, T. (2005) Knockoff: The Deadly Trade in Counterfeit Goods. Kogan Page Ltd. 96 Prabhakar, H. (2012) Black Market Billions: How organised retail crime funds global terrorists. Pearson Education Limited. 97 Stroppa, A., di Stefano, D. and Parrella, B. (2016) ‘Social media and luxury goods counterfeit: a growing concern for government, industry and consumers worldwide’. [Online]: https://www.washingtonpost.com/blogs/theswitch/files/2016/05/IG_A2016_ST2.pdf?tid=a_inl 22 has grown since their previous research paper, ‘Online Advertising Techniques for Counterfeit Goods and Illicit Sales’,98 which focused on counterfeited clothing sold through Facebook-sponsored ads. They argued that creative selling practices, such as those detailed in a 2015 overview by Bloomberg’s Roberts,99 showed counterfeiters targeting the same customers as the authentic brands with high-quality (and comparatively high-priced) deceptive and grey goods rather than focusing on low-price fakes. This is a finding that chimes with claims made by MarkMonitor and NetNames. Stroppa et al also noted that Google claimed to be cracking down on fraudulent bots that can be costly to advertisers, who pay Google for clicks on their ads. They cite a December 2015 poll, which revealed; “millions of online shoppers are being duped into buying counterfeit products with one in four being ripped-off”. Their June 2015 research (‘Instagram spam-bots and social media popularity’) claimed that Instagram is “infested with millions of spam-bots and fake accounts”100. Technology industry sources Bartlett’s The Dark Net 101 argued that everything, legal and illegal, is now for sale on the Internet, with consumers using trust cues to help them make decisions. Facebook’s first president Sean Parker criticised102 “the advent of social media and narcissism”, suggesting that social media behaviour differs significantly to that of the ‘real world’. This is a finding borne out in the social psychology literature.103 Facebook’s closed groups were the focus of concerns from Chivers104 ; a particular issue was the use of closed and secret groups on Facebook to host online arms bazaars, which is in contravention of Facebook’s policies. Facebook’s response was to describe their policies as evolutionary, reflecting shifts in their social media ecosystem. They acknowledged that they allowed users to process payments through the Messenger service, as well as providing other features to aid sales. Given criticisms from industry about Facebook’s cumbersome procedures, we noted concerns elsewhere about Facebook’s reliance on users to report wrongdoing at a time when the platform had the technology to identify user activity and even appearance. Lomas105 noted that Facebook had agreed to do more to combat hate speech on its platform, but only after German prosecutors launched multiple investigations into Facebook’s managers’ apparent failure to remove racist posts. Even as Facebook claimed that there was no place for such content on their platform, as with other forms of inappropriate and illicit behaviour, they looked to their ‘users’ to report infringing content. 98 Their paper was cited extensively in Lepido, D. (2014) ‘Fakebook: Site Touts Luxury Goods That May Not Be What They Seem’. Bloomberg [Online]: www.bloomberg.com/news/2014-11-13/fake-out-many-luxury-itemsadvertised-on-facebook-are-phony-researchers-say.html 99 Roberts, A. (2015) ‘Luxury Firms Fight Online Fraudsters Over Expensive Fakes’. Bloomberg [Online]: http://www.bloomberg.com/news/articles/2015-07-27/fakes-at-7-800-send-luxury-companies-online-to-fight-fraudsters 100 Seetharaman, D. (2015) ‘Fake Accounts Still Plague Instagram Despite Purge, Study Finds. Wall Street Journal [Online]: http://blogs.wsj.com/digits/2015/06/30/fake-accounts-still-plague-instagram-despite-purge-study-finds/ 101 Bartlett, J. (2014) The Dark Net. William Heinemann: London. Page 133. 102 Price, E. (2015) ‘Sean Parker on tech industry’s “tricks”: Social media feeds our narcissism’. The Guardian [Online]: http://www.theguardian.com/technology/2015/nov/09/sean-parker-social-media-facebook 103 Margalit, L. (2016) ‘The Psychology Behind Social Media Interactions’. Psychology Today [Online]: https://www.psychologytoday.com/blog/behind-online-behavior/201408/the-psychology-behind-socialmedia-interactions 104 Chivers, C.J. (2016) ‘Facebook Groups Act as Weapons Bazaars for Militias’. New York Times, 6 April 2016 [Online]: http://www.nytimes.com/2016/04/07/world/middleeast/facebook-weapons-syria-libya-iraq.html?_r=0 105 Lomas, N. (2016) ‘Facebook, Google, Twitter Commit To Hate Speech Action In Germany’. TechCrunch [Online]: http://techcrunch.com/2015/12/16/germany-fights-hate-speech-on-social-media/ 23 Twitter has also become a focus of concern and TechCrunch’s Lomas’106 critique of online abuse noted that Twitter had increased efforts to combat the spread of abusive tweets, including the development of a new range of tools for users, algorithmic filtering to limit the spread of abusive tweets, and the formation of a trust and safety council. Twitter readily admitted that “we are frequently not as good citizens online as we are offline”, but it is clear that its network structure “can more easily facilitate the spread of abuse as opposed to other more closed/siloed types of social networks”. The 2016 Demos’ social media study,107 which mapped thousands of aggressive and abusive tweets, noted the growing public concern about the reach and impact of hate speech and abuse on social media. Demos researcher Alex Krasodomski-Jones argued that even with the new opportunities for public debate and social interaction, the digital world has also provided a platform for the “worst aspects of human behaviour”. 5.3.2 Industry and trade bodies’ responses Certain brands identified a number of features of current IPR infringement, including China as the origin of most counterfeit manufacture, with many products shipped outside of their central distribution network and/or shipped in plain packaging. It was also claimed that clothing and licensed products were most likely to be counterfeited, as was anything in loose packaging. There was general agreement that meeting the problem of counterfeiting required a closed manufacture and distribution network. The issue of increased supply chain vulnerabilities and the infiltration of legitimate intermediary channels was picked up on by BASCAP’s Chris Oldknow, who saw a clear distinction between physical intermediaries and online intermediaries (such as websites, platforms, portals and services, infrastructure providers, search engines, online advertisers and payment processors). One company described online and social media as a “new opportunity for criminals” and an “evolving platform for illegal suppliers to sell their products”. There were examples where social media sites were being used and illegal suppliers were brazen in their approach. The key issue within the tobacco market is the distinction between counterfeits (made without brand holders’ consent) and grey goods (known as ‘illicit whites’), which are legally produced primarily for smuggling purposes but sold without payment of tax; both are sold at very low prices. FACT noted that technology has changed the physical goods market and social networking websites are now an integral part of modern life, with Facebook increasingly used as an online platform for DVD sales. They argued that social media, together with direct sale websites, online marketplaces and auction websites, is replacing ‘traditional’ hard goods sales methods. FACT claimed that social media has an influence on almost every type of copyright infringement that it investigates, and many ‘pirate’ websites have associated Facebook or Twitter accounts. They pointed to the ‘open’ groups used to attract online users and saw social networking accounts being employed to promote, advertise and directly link to the ‘pirate’ 106 Lomas, N. (2016) ‘UK study quantifies Twitter’s misogyny problem’. TechCrunch [Online]: http://techcrunch.com/2016/05/26/uk-study-quantifies-twitters-misogyny-problem/?ncid=rss 107 Demos (2016) ‘New Demos study reveals scale of social media misogyny’ [Online]: http://www.demos.co.uk/press-release/staggering-scale-of-social-media-misogyny-mapped-in-new-demos-study/ 24 website. Since the advent of UK website blocking orders, they have seen discussions on Facebook of workarounds and suggestions of ways to circumvent blocks. Twitter is apparently used in a similar way, as users tweet and share links to content. FACT argued that social media has played a crucial role in assisting IP infringement in the vast majority of their investigations. In their view, the use of social media by individuals and groups infringing member content is continually rising and is usually adopted to enhance existing illegal services (hosted on third-party, infringing websites) by keeping users updated with new content and news. Social networking websites are now integral to modern life, and the global reach of the platforms provides the perfect opportunity for criminals to direct the public to infringing content hosted on other websites, advertise and sell infringing products, and provide ongoing ‘customer support’. FACT also identified problems with fake/hidden website registration details and overseas servers. Among the surveyed ACG firms, the key problems in dealing with Facebook resulted from the platform’s refusal to close down a ‘counterfeiter’s’ customer profile unless it breached their terms and conditions. There was no clear trend in counterfeiting of new as opposed to old products and purchasers of counterfeits were seen as people seeking ‘cheap’ products and generally willing to ‘delude’ themselves into thinking they had got a cheap deal rather than a fake. The key social media-related issue was thought to be that “the main sales happen in closed groups, which cannot be scanned easily” along with problems locating infringing content on social media because of restrictions on Facebook searches. Most respondents indicated that Facebook, and to a much lesser degree Instagram and Twitter, are being used to sell counterfeit products in local selling groups. One respondent claimed that Facebook has taken over from eBay and Gumtree as the major area of concern. Selling pages and closed groups were the main problems for most respondents, even though one saw Amazon and eBay as causing the most damage in the UK. It was often difficult to locate infringing content because it appeared in closed selling groups or hidden profiles, which are unavailable to the average user. Closed groups are often characterised by the use of fake posts and pages, making it harder than ever to catch them without investing a lot of time and resources. In the evolving landscape, one brand claimed that infringement has shifted from marketplaces like eBay to mobile selling apps and social media, which are now the hotbeds for counterfeit activity. Another firm believed that there has been a rise given the proliferation of social network services, auction and large selling sites (like Amazon). What the various ACG members’ responses illustrate is the divergence of social media’s effects on brands, with some able to identify a clear link to Facebook (and, to a lesser degree, Instagram and Twitter) and the role of closed groups and diversion tactics, whereas others still see eBay and Amazon as the main sources of concern. There were far more detailed explanations of the characteristics of infringement available from the three main private enforcement agencies. MarkMonitor described the chief online issues impacting brands as including impersonation, fan pages, social media pages transacting business, promotion, and the proliferation of websites selling counterfeits and offering fake special offers. Counterfeiters see social media as a haven and actively use both open and closed group pages, along with ‘likes’ and ‘retweets’, to disseminate their offerings. The social media platforms made it 25 easy to move channels by establishing fan pages and making it possible to carry out transactions on or off the social media platform. MarkMonitor also drew attention to the ease with which counterfeiters could clone brand pages on social media, with some brand-impersonating pages having more likes than the corresponding genuine brand pages. NetNames described fake web-shops as those with high visibility and traffic that sell counterfeits by using images of authentic brands and logos and working as part of a network. NetNames focused on the whole online, e-commerce landscape, of which social media was just a part. They noted that brand owners were increasingly confronting anonymous online counterfeit and grey market sellers using rogue ecommerce websites and online marketplaces. NetNames saw the B2B marketplaces as the most important, because these were the primary platforms for selling and shipping large volumes of counterfeit and grey market products directly to customers. They mentioned on-going problems with auction sites, but for them mobile apps were the fastest-growing online channel for counterfeit goods. They also mentioned the dangers of traffic diversion, which involves cyber-squatting and the manipulation of search engine optimisation, as well as the use of social networking sites, blog entries and review sites to divert consumers to rogue e-commerce websites. In NetNames’ analysis, social media was part of a range of online tools used by counterfeiters to divert traffic away from legitimate websites. Yellow Brand claimed to focus on social media and argued that such channels are increasingly popular targets for counterfeiters, with fake goods being sold on both global and local channels, particularly in China and Russia, as well as via highvolume platforms like Facebook, Twitter and Instagram. Yellow Brand confirmed that counterfeiters’ tactics for avoiding detection include securing content in closed groups. They also identified particular challenges for identifying infringing content, such as ads for counterfeits that omit a brand’s name and so do not show up in online searches, as well as the dominance of online sales by marketplaces that source stock for many web stores and carry out business as business-to-business and consumer-to-consumer online marketplaces. 5.3.3 Government agency responses The use of the ‘dark net’ and VPN has enabled considerable online activity to take place away from any possible oversight and with greater potential reach than social media, especially Facebook’s ‘closed’ groups. The privacy and anti-surveillance issues found with the technology platforms and consumer attitudes have shaped enforcement tools and the ability of rights holders to use them. In addition, there has been a problem of definition, as some used the term ‘counterfeit’ to describe ‘grey goods’, while others used it more literally to describe any goods that were copies of originals rather than the same goods only authorised for sale. This confirms the common ‘original’ versus ‘copy’ distinction, which may be an unhelpful binary. This is an area that requires greater levels of distinction for effective enforcement. Trading Standards have had difficulties dealing with Facebook-based traders and particularly the sale of counterfeits in ‘closed groups’. They believed the ‘buy and sell’ function for different towns and cities now saw Facebook competing directly with 26 eBay. Their greatest concern was the flagrancy of some of the sellers taking photos of their goods ‘in situ’ and posting them to Facebook, and also claimed that “sellers enjoy […] ever-greater access to new customers via closed social media groups”. The key obstacle Trading Standards faced was the lack of intelligence and data, which has allowed Facebook to ignore the problem. Trading Standards believed that social media was particularly attractive to users when ‘advertising’ IPR-infringing products because there were no fees or costs yet they were still able to find buyers and sellers in the local area; this made social media an attractive proposition when compared to more traditional online marketplaces such as eBay, Amazon and Gumtree. IPO’s Intelligence Hub described social media as the retail end of counterfeit goods, with Facebook seen as presenting the biggest challenge, since its scale and reach make it impossible to police and enforcement agencies need Regulation of Investigatory Powers Act (RIPA) powers to investigate. It was also claimed that Facebook was not a selling platform, meaning selling goods on the platform could breach their terms and conditions. Social media, it was argued, provided a relatively safe way for members of the public to trade in these goods and the majority of the public in the UK saw counterfeit goods as socially acceptable. Indeed, there were sectors of society where this translated beyond tolerance into actively seeking out counterfeit goods that had the appearance of luxury goods but were more affordable. The public’s tolerance of counterfeits, the ease of ordering via the internet, and the relative security of anonymous online entities have created a safe haven for people to trade in counterfeit goods. 5.4. Opportunities for IP This section explores current initiatives for countering infringement, enforcing rights and promoting respect for IP, as well as assessing the effectiveness of these initiatives. As in previous sections, we have broken this down into reviews of industry (brands and trade bodies), government, academic, and technology industry literature. This is followed by the responses of industry, government enforcement agencies and technology firms to our survey questionnaires. 5.4.1 Literature and media review Industry A useful study from Spire (2011)108 suggested that enforcement strategies aimed at purchasers of counterfeits should take account of consumers’ greater readiness to listen to victims and experts than authority figures. We saw evidence of improving awareness of IPR and increasing consumer knowledge of counterfeits within a number of TV programmes, some of which were directly influenced by a number of the books featured in this study. The BBC’s Fake Britain (2014–2015) series stands out, as it explained the danger to consumers as follows: “easily set up dodgy 108 ACC/Spire Research and Consulting (2011) ‘Counterfeits: Measuring it, Fighting It’ [Online]: http://www.slideshare.net/spireresearch/010216the-american-chamber-of-commercecounterfeit-measuring-it-fightingit 27 websites and fake identities causing problems – in this case, with life-threatening consequences”109. Government The IPO’s Intelligence Hub considered eBay to no longer be as big a problem as social media sites, as there has been increased cooperation between eBay and rights holders. This seemed to be borne out by comments from certain brand holders, as well as academic sources. The calls from the National Trading Standards Board (NTSB) in the UK to create a single IPR enforcement agency in the UK to improve IPR enforcement coincide with complaints made by Facebook about the current fragmented enforcement landscape in the UK and also mirror recent initiatives, like the creation in the US of a new National Intellectual Property Coordination Center.110 This initiative brought 23 partner agencies together to be run as a taskforce to enable the best use of the resources, skills, and authorities of each partner in order to provide a comprehensive response to IP theft. The EUIPO’s 2016 study Intellectual Property and Youth111 claimed that arguments relating to personal safety rather than moral values were better suited to convincing young people to think twice about buying counterfeit goods. This echoes suggestions from Spire (above) about consumers being more influenced by victims’ stories than authority figures. Academic Chaudhry and Zimmerman criticised the current tone of anti-counterfeiting messages, especially themes reliant on role models, peer pressure, education, fear, product quality and negative associations with organised crime and/or terrorism. They question whether firms (and government) actually test the salience of their advertisements to assess the influence of such messages on their target audience. The effectiveness of these messages could be developed as part of a future demand-side-focused research programme. Prabhakar112 argued that online enforcement has been made much harder due to the lack of policing within e-commerce sites. The purely reactive approach adopted by the platforms is made worse by a lack of cooperation and information sharing unless pressed by brands and enforcement agencies. Phillips113 described the historical problems eBay encountered in dealing with counterfeit goods on their platform, which led to the introduction of a verified rights owner (VeRO)114 programme. While he argued that eBay still has a less-than-perfect takedown system, it is clear that the relationship between certain online platforms and brand owners has improved. In 2005, Philips also suggested that the UK’s Trading Standards was unable to cope with offline counterfeits because of its limited powers and resources; this still seems to be the key problem for enforcement 10 years on. 109 BBC (2013) Fake Britain [Online]: https://www.youtube.com/watch?v=-ezrVVtx9uU (16 May 2013) 110 IPR Center (2016) ‘About us’ [Online]: https://www.iprcenter.gov/about-us 111 European Union Intellectual Property Office (2016) ‘Intellectual Property and Youth Scoreboard 2016’ [Online]: https://euipo.europa.eu/ohimportal/en/web/observatory/ip-youth-scoreboard?platform=hootsuite 112 Prabhakar, H. (2012) Black Market Billions: How organised retail crime funds global terrorists. Pearson Education Limited. 113 Phillips, T. (2005) Knockoff: The Deadly Trade in Counterfeit Goods. Kogan Page Ltd. 114 eBay ‘Verified Right Owners programme’ http://pages.ebay.co.uk/vero/about.html 28 Technology industry sources Concerns were expressed about the use of online platforms’ terms and conditions, as showcased in the 2013 film Terms and Conditions Apply,115 which indicated that acceptance of these rarely meant more than a tick-box exercise for users. Similarly, a more recent UK parliamentary select committee report116 strongly criticised the length and complexity of the terms and conditions used by social media firms. The very low number of consumers who read the online platforms’ terms and conditions was also highlighted in Ofcom’s ‘Adult Media Literacy’ studies, published between 2013 and 2015.117 Even the Electronic Frontier Foundation admitted118 that the terms are “one-sided in the service provider’s favor […] often designed to be beyond any judicial scrutiny […] most users never even bother to read let alone understand these agreements filled as they are with confusing legalese”. Recent criticisms 119 of the big tech firms’ promises to take online abuse seriously included arguments for the platforms to be held legally accountable for hosting such abusive content. Laville claimed that the solution proffered by the social media platforms was just a means of evading their responsibilities, and this echoed comments made by some of the brands struggling to cope with the lacklustre responses from social media firms in removing infringing content. The attitude of the key tech firms to the rights holders is borne out by the activities of the Electronic Frontier Foundation, which described brand holders’ efforts to curtail infringement of their IPR as the actions of ‘trademark bullies’.120 Implicit hostility to such IPR is similarly manifest in the Ranking Digital Rights (RDR) Corporate Accountability Index,121 which opposes certain rights deemed harmful to an open Internet. The same index confirms the emphasis on the technology platforms’ relationship with their users and the balance of their users’ rights and expectations against those of national regulators and disrupted businesses. Even so, platforms like Microsoft’s Bing have shown willingness to issue warnings122 to their users about the dangers of purchasing counterfeit drugs online; this is the only such example found among these firms that could be described as an attempt to counter IPR infringement. Stuart Dredge’s analysis of Facebook’s Rights Manager tool, their equivalent of YouTube’s Content ID system, emphasised its role in managing copyrighted content uploaded to the social network and tackling infringement.123 This demonstrated the 115 Terms and Conditions May Apply [Online]: ttp://tacma.net 116 Cellan-Jones, R. (2014) ‘Social media told to simplify terms and conditions’. BBC News, 28 November 2014: http://www.bbc.co.uk/news/technology-30234789 117 Ofcom (2014) ‘Adult Media Literacy’ [Online]: http://stakeholders.ofcom.org.uk/market-dataresearch/other/research-publications/adults/media-lit-10years/ 118 Electronic Frontier Foundation (2016) ‘Terms of (Ab)Use’ [Online]: https://www.eff.org/issues/terms-of-abuse 119 Laville, S. (2016) ‘Top tech firms urged to step up online abuse fightback’. The Guardian [Online]: https://www.theguardian.com/technology/2016/apr/11/facebook-twitter-google-urged-to-step-up-online-abusefightback 120 Electronic Frontier Foundation (2014) ‘Who has your back? Protecting your speech from copyright and trademark bullies’ [Online]: https://www.eff.org/pages/who-has-your-back-copyright-trademark-2014 121 Ranking Digital Rights Corporate Accountability Index (2015) ‘Research Indicators’ [Online]: https://rankingdigitalrights.org/project-documents/2015-indicators/#content (Nb. comments under Freedom of Expression, F3 and F6). 122 Bing (2015) ‘Bing to warn customers about the threats of fake online pharmacies’ [Online]: https://blogs.bing.com/search/2015/08/06/bing-to-warn-customers-about-the-threats-of-fake-online-pharmacies/ 123 Dredge, S. (2016) ‘Facebook gets its own Content ID with Rights Manager’. MusicAlly, 13 April 2016 [Online]: http://musically.com/2016/04/13/facebook-gets-its-own-content-id-with-rights-manager/ 29 platform’s ability to come up with IPR solutions when it was in their business interests to do so. Tiku124 described Facebook’s ‘playbook’ as one where they first explored a new technology, then built that technology into a product, and finally attached that product onto its existing ecosystem of a “billion or more people”. The IPR repercussions of its new technologies are almost certainly of secondary importance at the outset of their introduction, requiring the firm to play catch up and rely very heavily on the letter of the law. That said, Facebook’s sophisticated technologies, like their DeepText AI,125 permit the platform; “to sieve through and understand several thousand posts per second across 20 languages” and enable Facebook to ‘decode’ messages, comments and posts and make recommendations for individual users. This tool could also be employed to search for illicit behaviour. 5.4.2 Industry and trade bodies’ responses The clearest description of ways to bolster enforcement of IPR came from a review of cross-sector best practices by BASCAP’s Oldknow, who called for new industry standards, to include the use of automated tools to identify transaction patterns, and the adoption of automation tools, as well as improvement to the transparency of notification, takedown and redress systems. This would require better coordination between intermediaries, government agencies and rights holders, as well as the adoption of preventative tools such as content filtering, verification, track and trace, and the improvement of the security of the global supply chain. Despite the different enforcement policies at the various online platforms (according to MarkMonitor), Yellow Brand claimed to be actively lobbying Alibaba and Facebook/Instagram in China to improve IPR protection for their clients. FACT argued that online behaviours demanded a wide variety of methods to deal with infringing websites, such as website closures, detection and removal of infringing content through takedown notices, and the use of auction website listing removal tools. The focus for social media was on reducing illicit websites’ popularity through search engine delisting and the removal of the offending pages on Facebook and Twitter, with the ultimate goal of restricting infringing websites’ revenues. FACT had a more positive view of the media platforms takedown policies’ than we encountered with individual firms, highlighting their procedures with eBay, Twitter and Facebook. The latter had an online reporting facility for rights holders and members of the public to report violations, including copyright and trademark issues. FACT had incorporated this procedure into their alternative to a prosecution strategy, with some success. They noted that Facebook would remove specific posts rather than the whole profile or community page, but would occasionally remove entire groups if repeat infringement could be shown. The ACG member rights holders we surveyed described having to monitor domains and social media online and then taking enforcement action against sellers and/or sites selling counterfeit versions of their products or using their imagery. Some used external private enforcement agencies, like MarkMonitor, but others made direct 124 Tiku, N. (2016) ‘Facebook Wants To Be The Layer Between You And The Future’. Buzzfeed News, 13 April 2016 [Online]: https://www.buzzfeed.com/nitashatiku/facebook-wants-to-be-the-layer-between-you-and-the-future 125Facebook (2015) Introducing DeepText: Facebook's text understanding engine https://code.facebook.com/posts/181565595577955/introducing- deeptext-facebook-s-text-understanding-engine/ 30 contact with the platforms to lodge complaints. Overall, firms used all the options open to them to enforce their rights online and offline. Some manually collected data from online listings and requested information from marketplaces such as eBay for their records, which they regarded as confidential information, while others relied on their external agents. eBay was considered easier to deal with than Amazon, but Facebook was not seen as helpful or supportive given “major criminality on their platform from sexual predators, terrorism and many other larger issues. Most of the platforms are out of their depth”. The perceived resistance from social media platforms involved a time-consuming process of finding, reporting and taking down infringing posts, along with a reliance on images used by sellers in the absence of test purchases. Concerns were expressed about Chinese B2B sites and problems investigating the supply chain, as well as determining the source of the counterfeits. There was less confidence in the current online enforcement process because “criminals are light years ahead of law enforcement” and due to the “high costs of maintaining and enforcing rights” and the “lack of cooperation from online platforms”. The IP Crime Group’s Social Media Group claimed that social media platforms were not subject to the RIPA 2000 Part III126 and that any compliance by them with the RIPA was voluntary. They argued that, despite some successes with Facebook taking down pages on sites identified by The Police Intellectual Property Crime Unit (PIPCU), closed groups had become the most challenging aspect of dealing with Facebook, requiring enforcement agencies to go undercover. In their experience, illicit offerings were made across various platforms, with clear links between the social media platforms and online marketplaces. A key insight was the emergence of Twitter as a new threat for digital media rights holders because of the speed with which infringers can publish infringing content, which serves as an illustration of the speed of change within fast-moving markets; even so, there was little sense that Twitter had increased the threat in terms of ‘hard’ goods. Facebook’s purely voluntary compliance with the RIPA made enforcement hard, as the platform required individual URLs to be forwarded for takedown. This timeconsuming and resource-intensive process was the result of the platform’s concerns about personal images being included in the takedown and meant that they were unwilling to take down whole albums of photographs. The group pointed to the process where Trading Standards’ forensics team monitored fake traders’ uploads, but brands were only able to report images impacting their own product. Since not every brand was a member of the ACG, this meant counterfeits of products from non-ACG members could still be offered. At the time, it was assumed that Facebook was not a selling platform, meaning counterfeiters could be in breach of the platform’s terms and conditions, which offered an easier tool for disruption than claiming IPR infringements. The main issues for industry were the platform’s refusal to accept bulk requests and establishing a more streamlined, coordinated approach to removing infringing content from the platform. 5.4.3 Government agency responses 126 Regulation of Investigatory Powers Act 2000 http://www.legislation.gov.uk/ukpga/2000/23/section/49 31 The existing complex and fragmented127 enforcement approaches seemed to partly explain the technology firms’ current cautious and limited cooperation, but effective IPR enforcement has been further hampered because while the nature of online IPR infringement crime was global, enforcement was local. The apparent lack of integrated approaches meant a need for greater coordination between stakeholders. Only multinational firms could truly adopt a multi-territorial approach, although Europol and Interpol could be part of the solution. Trading Standards issued a call for a single national body for IP crime that could be part of one of the existing agencies (such as immigration, customs or Trading Standards), and this recommendation had real resonance within the UK enforcement environment and chimed with recent changes in IPR enforcement in the US. The NTSB’s and Trading Standards’ current ‘IP drive’ focuses on Facebook and eBay and their national control strategy is about prevention, intelligence and enforcement, but this relies on cooperation with other agencies. Their interactions with Facebook had been difficult given the problems of evidencing the scale of illicit activity, and this has remained their biggest challenge in relation social media. Their reduced and limited resources mean they need more help, especially given the broad scope and scale of their work, with 12,500 feeding investigations, at a time when the local authority Trading Standards model is no longer fit for purpose. Trading Standards proposed a more regional approach as part of efforts to “join all this up”, and this meant being intelligence led and required greater cooperation from brands, industry bodies, and the IPO to tackle IPR infringement on social media. They admitted that their response to IPR infringement on social media platforms had been patchy across the UK, and any good practice and active enforcement work was limited in scale. They had encountered problems with identifying the ‘owners’ of social media profiles, with social media platform operators providing very little (if any) information to enable successful identification and location of offenders. They noted that there was no straightforward mechanism to identify potential IPR infringement on social media, but they had agreed a formal procedure with Facebook after extensive consultation between the NTSB’s eCrime Team and Facebook. The IPO’s Intel and Enforcement hub pointed to new initiatives aimed at joining up the current different approaches to reap the benefits of cooperation between industry and IPO/Trading Standards, such as the National Markets group and the Real Deal campaign. The establishment of the National Trading Standards Board in 2012 was supposed to have created a network of intelligence analysts and a national intelligence hub, but this has not been effective thus far. The current economic climate has resulted in a number of partnerships of necessity, but there has been little coordinated work undertaken in respect of social media due to the stance taken by the NTSB. According to the enforcement agencies, the way the social media groups were constructed required a surveillance authority under the RIPA 2000 to view any infringing goods. 127 Seventeen different UK agencies are involved in IPR enforcement. 32 5.4.4 Tech firm responses Google provided two key documents (How Google fights the advertisement of counterfeit goods and Google’s AdWords trademark policies) that set out their stance in relation to AdWords and counterfeited products. There was a clear line in their AdWords policy allowing trademarks to be used as keyword triggers in AdWords, and they claim to be unable to “arbitrate trademark disputes all over the world”. Google argued that determining infringement was a matter for the courts, especially as trademarks were territorial and applied to certain goods or services. They said they would, “as a courtesy to brand holders”, investigate reasonable claims about trademark violations in their ads and pointed us to their specific region/country policies and their “easy to use complaint form”. By contrast, Google’s stated policy on counterfeits had a very different tone and the firm claimed to have a zero-tolerance policy in relation to the advertisement of counterfeit goods. They claimed that ads on searches for trademarked terms were not confusing as the ads were very clearly delineated as ‘sponsored links’, but ads that were actually deceptive would violate their Terms of Service. Google argued that the Internet had created new complexities, and many stakeholders had a role to play in resolving this issue. More significantly, they claimed that brand owners and law enforcement must tackle counterfeiting at its source. However, they also clearly set out the limits of how online services could help given that they “are in no position to determine the authenticity of the millions of advertised goods, as they never even take possession of them, and fraudsters are always coming up with more sophisticated ways to game the system”. Facebook encounter a diverse range of crimes on their platform that impact public safety, but their priorities were whatever affected the safety of consumers, including combating child exploitation and terrorism. Any criminal activity was against their Terms of Service and they were able to ‘join the dots’ when law enforcement was looking for evidence of criminality. Their real-name policy meant that account holders had to use their full legal name, and this caused the platform problems with enforcement agencies, as they were aware that law enforcement had set up fake accounts, which Facebook could and would close down. In relation to rights holders, they had to accommodate all the various different laws and this required individual notice procedures. Because of this, they were unable to assess bulk processing. That said, they were able to ‘whitelist’ trade body reporting for members divided between copyright and trademark infringements, claiming their response turnaround time was within two hours. They had little contact with UK law enforcement apart from Trading Standards and did not encounter many impacted stakeholders, but the UK situation was apparently complicated. Facebook’s reaction to allegations of infringement was to produce a standard set of questions, as they occasionally received multiple requests. With multi-brand counterfeiters, these were often subject to Proceeds of Crime orders and on-going criminal investigations. There were many existing takedown requests that featured duplicates from rights holders and Trading Standards, usually involving test purchases. Facebook claimed that rights holders could report albums of photos of infringing content. Facebook is primarily a communication platform, so they are not involved with the online transactions and this is the chief reason for their need to establish the exact nature of any alleged infringement. Facebook had agreements in place with the 33 Home Office and the Information Commissioner's Office (ICO) to meet data requests, including supplying basic subscriber info. This had to come from their Dublin office, although they approved 75% of such data requests. As a US firm subject to US laws on divulging data, it could take six months to get the content of subscriber accounts, although they could volunteer subscriber information if there was a clear justification. They could provide the same information for Instagram where they saw certain levels of infringement. Facebook’s lawyers were perceived as risk averse and had difficulties understanding and managing the existing fragmented approach to the enforcement and diversity of stakeholders. Suggested efforts to streamline enforcement that might improve the process still have to satisfy Facebook’s requirement for proof. In their formal processes, Facebook claimed, according to their ‘Statement of Rights and Responsibilities’,128 that their users were prohibited from posting content that violated another party’s IPR. Additionally, they said that they offered tools to report potentially infringing content posted by users on their service. Their online reporting tool could be used to report both copyright and trademark infringements, and they pointed us to the Electronic Frontier Foundation (EFF) 2014 report entitled ‘Who has your back?’ 129 , in which the EFF argues that technology firms need to defend themselves against government requests for data. Apart from acknowledging that law enforcement agencies could request data relating to a criminal investigation, Facebook only disclosed account records in accordance with its Terms of Service and applicable law. As with both other technology firms, the platform published statistics130 on government requests for data and content removal. Twitter’s specific policy relating to trademark infringement covered anything considered a trademark policy violation and detailed their response to reports of trademark policy violations from holders of federal or international trademark registrations. When satisfied that there was a clear intent to mislead others through the unauthorised use of a trademark, Twitter would suspend the user account and notify the account holder. They distinguished between such accounts and those they determine are accounts that are confusing other users but “not purposefully passing itself off as the trademarked good or service” – in this scenario, they give the account holder an opportunity to clear up any potential confusion. They listed in detail how their counterfeit goods policy prohibited user attempts to pass themselves off as products of the brand owner. They noted how violations could be reported and, if their rules are broken, how such a violation will trigger “appropriate action”. Unsurprisingly, they asserted the following: “as is standard industry practice, we do not proactively monitor the content users post to Twitter”. 128 Facebook (2015) ‘Statement of Rights and Responsibilities’ [Online]: https://www.facebook.com/legal/terms 129 Electronic Frontier Foundation (2014) ‘Who has your back? Protecting your speech from copyright and trademark bullies’ [Online]: https://www.eff.org/pages/who-has-your-back-copyright-trademark-2014 130 Facebook (2016) ‘United Kingdom Requests for Data July–December 2015’ [Online]: https://govtrequests.facebook.com/country/United%20Kingdom/ 34 5.5 Consumer Tracker and Survey Results 5.5.1 Introduction The consumer data was based on the results of a three-month online tracker and consumer survey looking across Google/Facebook/Twitter/eBay. The data revealed different types of online consumer behaviour that can be segmented as featuring unexposed/unaware, compliant, deceived and complicit consumers. The research team cross-referenced the consumer data with customs seizures data, as these were the most consistent data available relating to the volumes of seizures, even if they were only available for intermittent very high-level information. The government seizure data evidenced a jump in reported volumes from 2005–2008 to 2010–2013, but within each group of data the trend was negative. The seizure data generally focused on high-volume opportunities, with each case covered during the initial period (2005–8) concerning over 2,000 items on average. However, during the second reporting period (2010–13), the average had declined to just over 270 items. Our report focused on six particular consumer sectors, but this specific sector data was only publicly available at an EU-wide level and volumes of seizures differed widely between the different sectors, even though normalised trends indicated that, at the EU level, the trends in seizures were downwards. There were a number of alternative interpretations of the emerging data, including: • Actual non-compliant behaviour might have been declining within these sectors, in line with the overall reduction in seizures in the UK. • Reductions in seizure activity might have been a result of changes in public policy and/or cuts in enforcement spending. This seemed unlikely as regards the UK, where seizure cases had increased but the average volume of items in a case had dramatically fallen. • A more likely interpretation was that non-compliant activity had become more diffuse and difficult to detect and capture • It can also be argued that the distribution of non-compliant products was being handled at a more granular level, which more easily evaded existing government enforcement activity. Connecting with government and sector-derived statistics As can be seen, the overarching trends in the available government and industry information are open to several interpretations, notably within individual sectors and official seizure levels. We therefore sought to complement these top-down data sets with bottom-up data. The approach to the latter started from the perspective of purchasing and counterfeiting online, as experienced by individual consumers. We sought to include both explicit experiences, as divulged through a consumer survey, and compare and contrast these with the consumer behaviours captured within the social media tracking study. 35 Limitations in the approach In reviewing the current estimates that are based on this bottom-up approach, it is important to keep in mind that there are a number of limitations. First and foremost, because of the focus on consumer-derived data and the deliberate use of ‘naïve’ researchers, this study focused on counterfeit goods that, at some stage in the buying cycle, became explicit. In the tracking study, the researchers were not even in a position to assess the actual products delivered by a service, but rather assessed them according the nature of the online endpoints. Further tracking studies would be required that completed the buying cycle and used detailed expert examination to identify the voracity of the products delivered before this would have a direct link to all counterfeit activity. Furthermore, in the case of the consumer survey, while there was an informal assessment of the products received, this was based upon consumer perceptions, which may well wrongly attribute products. Both options are possible, as properly licenced goods may, on receipt, be considered substandard or even counterfeit. Secondly, this initial investigation was designed as a proof-of-concept ‘existence’ test, to validate whether these consumer-based approaches could, at viable levels, actually detect significant levels of activity. As such, they were designed to give the activity the best options for detection. This was done by selecting sectors where counterfeiting was known to be prevalent elsewhere. Then, within each sector, there was a focus on brands that had high levels of online search activity. This made sense for this stage of a study, but it should be remembered whenever there is an attempt to extrapolate estimates from within the study to more general cases. In this report, the general estimates have been given as a sense of the potential order of magnitude of any impact and also as a demonstration of the types of methods that could be used in further studies, where more wide-ranging sampling methods are deployed. Thirdly, these initial investigations accepted the challenges involved in building a representative sample of consumers. As we were focusing on online purchase activity, we felt comfortable utilising online survey methods, and we deployed some methods of weighting to mitigate the potential for the behavioural characteristics of the study to be preferentially chosen within the sample. We have also sought to design the survey as carefully as possible to reduce response bias. However, none of these effects can be reduced to zero and, at some stage, it may be possible to generate complementary studies via different routes that allow for an unbiased estimate of these side effects. While there are clear extensions to the current method that can mitigate, to various degrees, the limitations outlined, the most straightforward approach is to introduce a regular sequence of similar studies. This requires a long-term commitment to the approaches, but allows the information provided to move from being a snapshot in time, of an ill-defined baseline, to provide a rigorous methodology that can detect changes in the economic system. This is the most straightforward way in which studies like this can begin to enter into mainstream macroeconomic studies. Once stable time series are derived, a series of tools can be used to begin to examine the 36 correlation, lead, lag and impact of different characteristics, in particular, the rise, fall and influence of social media on counterfeiting, as well as assess counterfeiting levels in general. This would be a similar approach to the differences found between the UK national crime survey and those statistics that are linked to police-reported crime. Trends in consumers’ online behaviour The longitudinal research of online behaviour conducted by the British Population Survey demonstrated a consistent trend of increasing rates of online purchase of non-groceries and, alongside it, the steady increase in engagement with social media This increase was consistent across social grades and age groups, with differences anticipated between each segment. Notably, those over 50 lagged significantly behind younger groups in the uptake of online purchasing, but arguably this difference could disappear, as the younger groups grow older. There has been an even more dramatic uptake of social media over the years since 2008, with the young adopting social media first; however, by the end of the research period (August 2015) this appeared to have reached a plateau, possibly as this age group had moved more to private messaging services like WhatsApp. There were less obvious differences in social media use between social grades AB and DE; indeed, social grade C1, by the end of the data collection period, showed the highest penetration of use. The data on trends in consumer behaviour online demonstrated the likelihood that increasing numbers of digital transactions were less susceptible to existing enforcement methods and monitoring for infringement. The research also indicated the overriding influence of past experience and advice from friends and family; these measures remained relatively stable throughout the whole research period, with around 70% of consumers claiming they are influential. By contrast, a small (less than 5%) but gradually increasing proportion of consumers was being influenced by social media recommendations and also by brands’ presence on social media. Tracking online promotion of brands The study aimed to establish the potential relationship between physical goods and social media. It was recognised that a lot of non-compliant behaviour could be attributed to the ease of passing off grey goods or near-perfect fakes in online environments. We segmented the respondents according to their behaviours: • complicit consumers – actively seek out ‘copied’ goods • compliant consumers – avoid copied goods • deceived consumers – mistake copied goods for the real thing • unexposed consumers – only find genuine goods 5.5.2 Tracker The tracker was designed to maximise the detection of interaction effects and worked across Google, eBay, Facebook Groups and Twitter, focusing on six different representative sectors, with two products in each: alcohol, cigarettes, clothing, footwear, perfume and watches. 37 After initial periods of tracking, the study systematically queried the different sectors over a period of three months, covering June to August 2015. We automated a sequence of searches and captured the results for each hour for each brand; the links within each selection were then followed to the endpoints (where no further links were provided) and these endpoints were then manually reviewed to assess whether they were genuine or suspect. We then contrasted two characteristics: the proportion of communications that were suspect; and the high level of skew in user behaviour. This revealed that a small proportion of users generate a lot of the suspect activity. The volumes of search quantities (see Tab. 1.1) depended on the search algorithms within the platforms themselves, and this was an appropriate test as it reflected the experience of normal users of the each platform. The volumes were much bigger from Facebook and Twitter, due to the availability of automated Application Programming Interface API connections. Tab. 1.1 Volume of tracked online items by platform and brand Results of tracking online promotion of brands The volumes of search results varied considerably across the different search terms, with Google the most consistent, given that its content was less dynamic than the content of the other platforms. The largest group of individual communications were driven by search terms on Nike within the footwear sector. Alcohol had a relatively low presence and was completely undetected within Facebook, with only negligible content within eBay. Tobacco was similarly low profile, reflecting eBay’s terms and conditions. This may be because alcohol sales are universally prohibited and smoking is generally restricted according to country-specific rules and regulations. The highly concentrated nature of suspect behaviour was established within different social media channels. This can be seen in the two charts below (Figs 1.1 and 1.2). These highlight the proportion of suspect communications within the different channels and in particular show the exposure in closed and open Facebook groups. It was evident that the proportion of suspect classifications was much larger on 38 Facebook, within both ‘open’ and ‘closed’ groups. However, there was much higher exposure to suspect material (over 500%) in closed groups than in open groups. The data showed very different behaviours between the different platforms, which was demonstrated by detailed examination of the distribution of exposure. Fig. 1.1 The overall proportion of communications that led to suspect endpoints, by platform The Google platform kept the suspect behaviour distinct and diffuse from the other types of behaviour, and eBay also demonstrated an even more diffuse distribution of suspect behaviour, at 20% of suspect initiators. By contrast, the Gini curves in Fig. 1.2 show that both Facebook group types had highly concentrated suspect behaviour: 76% and 58% of suspect communications, within closed and open groups respectively, were generated by only 10% of users. Fig. 1.2 Gini curve of suspect communications, by unique username within the group 5.5.3 Consumer survey Challenges These included the online sampling selection, verification of suspect online personas, and handling the highly skewed nature of typical digital behaviour, as seen in the tracker. Self-selection in online surveys has presented something of a statistical challenge when using surveys to make general estimates of scale across the full scale. There was a need to minimise the sense of steering a response and the risk of sounding pejorative, hence the adoption of the phrase ‘copied products’. The responses were segmented using the same complicit, compliant, deceived and unaware matrix employed for the tracker. Assessing the distribution of behaviour online, and balancing the accuracy of specific responses with the opportunity to scale the behaviour, meant restricting purchases of specific products and focusing on the last time the product was purchased. This kept the study synchronised with the online tracking study. The questionnaire was conducted during July 2015, the middle period of the online tracking study, and the survey was concluded once 2,999 respondents had completed the study. Participants were selected to match the demographic distribution of the UK population. 39 Consumer survey results and insights This showed that there were more frequent purchases online for the sectors examined by females and the highest volume of activity was within the age band 30– 34, although the highest frequency of purchase was within the age band 35–39. There was a preponderance of activity within the London region that showed both higher volumes and higher frequency of purchase. Those respondents within social grades AB had a high capacity and frequency of purchase. Social media was the least common source of information for purchases, but was associated with the highest frequency of purchase. Where social media was used we noticed that the average frequency of purchase increased in the opposite direction to the levels of use of a platform. Facebook had the highest level of use, but the lowest associated frequency – this was a symptom of wider experience being associated with a wider repertoire of platforms used. It was noted that higher frequencies were associated with purchases that were knowingly connected with unbranded sites, probably hosted abroad. The vast majority of purchases online (over 75%) appeared to be genuine, although this may have included a proportion of undetected fraud. There was a low level of perception of the risk of harm in choosing ‘copied’ products. There were very distinct demographic differences. Overall, just over 71% of respondents were ‘unexposed’ to the suspect locations, but this varied between different demographic groups. The highest proportions of ‘complicit’ behaviour were identified within males, the 30–34 age group, and social grades AB; this is shown in the Sankey chart displayed in Fig. 1.3. There was also a suggestion that complicit behaviour was geographically concentrated within London. 40 Fig. 1.3 Sankey diagram of the age band and social grade distribution of consumer segments The segments most likely to be ‘deceived’ were males, the 25–28 age group, social grades DE, and those located in London, the West Midlands or Northern Ireland. The most compliant segments were males, the 18–24 age group, social grades AB, and those located in either Northern Ireland or Scotland. The strongest difference in locations was that activity that took place on unbranded sites was likely to be hosted abroad. These sites present clear risks for tracking, and also the complicit evasion of excise duty. The connection between social media engagement and ‘complicit’ behaviour was clear. For individuals who did not use social media at all, the complicit segment made up only 13.9% and the unexposed were 77.8%. However, for those who used social media the proportions were reversed, with 74.5% within the complicit segment and 9.1% within the unexposed segment. The data indicated that complicit behaviour was symptomatic of high levels of digital engagement 5.5.4 Consumer summary and conclusions The data was indicative of a certain type of behaviour in relation to popular brands, and we cannot claim that this was entirely representative of all online purchasing, as the test was designed to identify the most likely locations for copied products to assess the impact of social media. The test was about searches for particular popular brands, rather than being a representation of general purchasing behaviour. Nonetheless, the data shows a pattern that further surveys and tracked research 41 could build on. There are further caveats about those who take part in online surveys, as well as the use of ‘copied’ products within the survey, which has a broader scope than infringing or fraudulent products. Even though these findings have to be properly contextualised, it is clear that this data supports many industry and government agencies’ claims about the roles that the internet and social media have in enabling counterfeited products to be made available and purchased. We were concerned about the credibility of claims about high levels of complicit behaviour, but the study was designed to focus on areas where complicit behaviour was more likely, so that we had a chance to distinguish the specific role of social media. Contrary to our own assumptions before the study, this appeared to be considerable. The project’s aim was to demonstrate, through research, whether social media and online sites increased the scale and impact of counterfeiting and piracy. This data indicated how much easier it was to consume and supply fake goods through webbased marketplaces and social media, which is hardly surprising given the impact of social media and the internet on modern life. The counterfeit trade may well be little different from any other activity, but the change this data points to is one of scale. Overall, these figures were impactful and there were clear implications to this data. The 17.5% figure was almost double the highest modelled estimates from OHIM on levels of counterfeited products within the clothing (the most pirated) sector. The estimates also support the general thread of the major brands’ arguments that the web has accelerated levels of and opportunities for counterfeiting, but until now they have had no way of calculating this. Given the paucity of offline data from government in relation to social media and the attendant reliance on anecdotal or ‘rule of thumb’ evidence, this data marks the first time we have become aware of the estimated levels of counterfeiting and piracy activity, not through forecast models but through tracked behaviour and surveyed attitudes. • 17.5% of transactions online were found to be of copied products: o Of this, 15.4% of online transactions were conducted by a ‘complicit’ segment of consumers who willingly participated. o Only 2.1% of online transactions were accidental purchases by individuals that were ‘deceived’ and only found this out on receipt of the goods. • Social media was the most distinctive medium for communication on copied goods: o 46.1% of ‘complicit’ purchases involved social media. o In contrast, only 4.1% of ‘unexposed’ purchases involved social media. • Social grades AB were significantly involved in ‘complicit’ behaviour: o 24.5% of social grades AB acknowledged ‘complicit’ behaviour. o In contrast, only 12.7% of social grade C2 acknowledged ‘complicit’ behaviour. • Online communication of suspect products was highly concentrated within a very small proportion of participants, currently particularly located within Facebook: o 72.5% of the suspect communications within open groups were generated by 0.78% of promoters. 42 83.4% of suspect communications within closed groups were generated by 6.2% of promoters. • Facebook groups represented the most exposed location for suspect communications, with suspect activity being much more prevalent on closed groups: o 8.3% of communications within open Facebook groups were found to be suspect. o 40.8% (five times more) of communications within closed Facebook groups were found to be suspect. o Final insights from the consumer part of this study include: a) On the methodology, the online tracker only captures complicit behaviour and to capture deceived behaviour would require an augmented approach, starting with mystery shopping, to identify the relevant links and then track them. b) Online groups are self-organising, involving herding. This is comparable, from an enforcement viewpoint, to activities within terrorist cells. c) Despite the emphasis placed on the threats posed by closed groups, opportunities exist in open groups to secure new users and these represent the greatest threat from social media in amplifying the counterfeiters’ messages. Even if the open groups are shut down, they can easily be set up again. d) Social media amplifies the counterfeiters’ messages by increasing the connectivity of potential complicit consumers. Crucially, these connections do not have to be strong; as can be seen from network effect sources the threshold for connection on social media is low. 5.6 Economic Assessment 5.6.1 Assessing harm and impact This section relates to the impact of IPR infringement, as well as the characteristics of infringement, notably the consumer behaviour involved. There were a number of different methods employed to assess the impact of and harm stemming from counterfeiting, including the OECD’s primary versus secondary market segmentation, which was adopted by BASCAP to describe non-obvious versus obvious copies, Hopkins et al.’s ‘harm matrix’ and various studies from GAO and, more recently, OHIM (now EUIPO). These were all helpful in identifying the different effects, including damage and harm (direct and indirect) to different stakeholders. Stakeholders When addressing the issue of the economic impact of counterfeiting activity, there were four basic parties to be considered: a) Industry/manufacturers. These are the different sectors involved, including luxury brands, fashion goods, and alcohol and tobacco (both of which are impacted by high UK duties). b) Consumers. These are our complicit, compliant, deceived and unexposed consumers of the products. 43 c) The social media sector. These are all those firms involved in the online world. This sector mainly covers Facebook and Twitter, but also includes online marketplaces like eBay/Amazon and search engines like Google and Bing. d) Government. This is affected in terms of both its tax revenues and consumer interests. Damage The following were the key elements used when assessing damage and harm. a) Brand reputation. This is the extent to which, where the degree of deception is high, counterfeiting can cause harm to the manufacturer of the authentic product. According to our interviews with industry, this was a key concern, particularly as deception online was much easier to achieve than deception offline. It also featured as a concern in the private enforcement agencies’ reports. b) Primary versus secondary/non-obvious versus obvious. This illustrates the different impacts on consumers (BASCAP) and is crucial to distinguishing the different types of consumer behaviour involved. Here, the damage or harm relates directly to the consumer. According to the consumer tracker and survey, the number of consumers who were deceived was far lower than the number of consumers who were complicit, but MarkMonitor and others have suggested that the proportion of those who were deceived in online purchases is growing. c) The harm matrix. Hopkins et al.’s typology of harm expands on the above primary versus secondary market distinction and categorises ranges of harm as follows: high quality/high deception, high quality/low deception, low quality/high deception and, finally, low quality/low deception. For our purposes, we must assume that the greatest harm and damage to both consumers and brands is concentrated in low-quality/high-deception goods, with the consumer left feeling cheated and the brand, according to industry sources, being blamed. There is also the potential harm to the consumer with goods such as fake hairdryers, where there are clear health and safety issues. This typically occurs when the goods are sold in near-perfect packaging while the goods themselves are inferior and unsatisfactory, thus representing the greatest level of deception. We believe that high-quality/high-deception goods would have the greatest financial impact on brands, as these would most likely represent lost sales to the brand. High quality/low deception may be of most benefit to consumers and is most likely to appeal to complicit consumers; these are arguably (at least according to some) not lost sales for the brands, as these consumers would not have purchased the authentic goods at the authentic price in the first place. What is clear is that high levels of deception, rather than high levels of quality, are the cause of the greatest harm, with high-quality, deceptive purchases being harmful to industry, and low-quality deceptive purchases being harmful to both industry and the consumer. d) Price. Morales notes the impact of price, as well as quality differences when purchasing counterfeits, with consumers more likely to buy counterfeit products when the price of the original is significantly higher than that of the counterfeit and when the quality of the counterfeit is sufficient. He also suggests that the consumer’s feelings about the company making the original product are important. 44 This means that, as the difference in price between the original and the counterfeit product increases, this will increase consumers’ readiness to buy the counterfeit. By contrast, as the difference in quality between the original and the counterfeit product decreases, this will also increase the likelihood of consumers buying the counterfeit. Morales also claims that the difference in quality moderates the effect of the difference in price on the consumer’s purchase intentions. In our view, the price consideration may apply mainly to complicit consumers, but in e-commerce, where consumers are looking for if not expecting an online discount, there is clearly scope for them to be deceived by clever pricing. In terms of relating this to social media, it could be argued that low-quality/highdeception goods are the ones industry and enforcement agencies have claimed are likely to be sold on platforms with near identical, if not cloned, images from authentic goods websites; these are used to deceive the consumer into purchasing them. These may just as easily be offered in open groups, which could add to the sense of authenticity. This may also be true of high-quality/high-deception goods, where the pricing may be closer to the authentic price to attract a purchaser looking for an online discount. This is typical of certain products where the reproduction of the goods is near perfect, such as DVD box sets. Equally, the kinds of goods often being disseminated across closed groups seem to be high-quality/low-price products, where damage may be limited. Impact The direct and indirect impact on the main three stakeholders – industry, government and consumers – from social media can be assessed as follows: • Direct impact. Industry has been seeing a loss of revenue because of the potential substitutional impact of counterfeits on authentic goods, particularly where there is a high degree of deception, although this is lower if the goods are non-deceptive. This represented the most likely impact of social media where the platforms enable the dissemination of deceptive counterfeit goods. • Indirect impact. For industry, this is the result of reputational harm from the lowquality/high-price goods that are common on social media (according to industry and enforcement agencies). • Indirect impact. This is the impact on government due to the loss of employee and corporate taxes and the impact on employment. There is a widespread belief that much of the activity emanates from and profits are made by counterfeiters in China and other Southeast Asian markets. There is also the cost to government of having to enforce IP infringements, such as the activities of Trading Standards. • Indirect impact. For consumers, there may well be a welfare benefit (recognised by GAO) for certain types of products (high quality/low price), but for almost all other types of products however the impact is direct. Initial summary of harm Harm can be expressed as a function of price paid and expectations across the three key elements of the market: consumers, brands, and government. At the extreme end, paying a low price for low expected quality is of little harm except for the 45 marginal loss in tax, assuming the same amount would otherwise have been spent in the legitimate economy. Paying a high price for high (expected) quality, albeit often receiving low quality, is clearly harmful and represents a loss for the consumer and the authentic product manufacturer, as well as a loss of tax. Any other combinations would have a different net position for each player, which can be simply defined and explained. In essence, this means that the economic impact of low-priced and low-quality goods is very low and the main damage would be in the high-price/high-expected-quality end of the market. This higher economic impact and harm are concentrated in what we describe as ‘deception transactions’ that, according to our consumer data, represent a small proportion of the total counterfeiting and piracy volumes. The other significant economic impact on industry is on low-price/high-quality transactions, typically of grey goods, although these attract complicit consumers who are actively looking for fakes and are unlikely to purchase the authentic full-price goods. 5.6.2 Other approaches Behavioural economics approach There may be an argument for further research into consumer decision-making and the underlying trends in consumer behaviour, such as obsessions with discounts and frequent sales. Thaler131 argues that all economic decisions are made through the lens of opportunity costs, and decision making involves acquisition utility (i.e. consumer surplus) and transaction utility, which is “the difference between the price actually paid for the object and the price one would normally expect to pay”. Social media, groups and networks’ influence on behaviour Kasteler132 notes how social media influence our shopping, relationships, and education, and cites research suggesting that most social networks support preexisting social relations, meaning Facebook is used to maintain or strengthen existing offline relationships, as opposed to meeting new people. However, social networks are often designed to be widely accessible and many attract homogeneous populations initially, making it easy to find groups using sites that are easily segmented. There are also new theories about the nature of groups, including group effects, selforganising groups and criminal behaviour. More connections within groups (known as a concentrated network) can reinforce behaviour in the groups, but more connections between groups (known as an integrated network) can open up a group to new behaviours and to behavioural change. In self-organising groups, there can be different outcomes between ‘open’ and ‘closed’ schools in terms of shaping attitudes. The changes in behaviour between the two beg the question of whether we can see appreciable differences between open and closed groups on social media platforms. There are also indications that social networks can contribute to criminal 131 Thaler, R. (2015) Misbehaving: The making of behavioural economics. Allen Lane. 132 Kasteler, J. (2010) ‘How Social Media is Influencing Your Behavior’. Search Engine Land [Online]: http://searchengineland.com/how-social-media-is-influencing-your-behavior-40615 46 behaviour, often as by-products of individuals acting with some self-interest. The interpersonal spread of criminal behaviour is an example of a bad network outcome, with evidence suggesting that, partly due to social interactions, criminal actions in a given place and time can increase the likelihood that others will commit crimes. The Economist recently133 defined the online platforms as a type of marketplace where people and businesses trade under a set of rules set by the owner. For ecommerce sites like Amazon and eBay, this means connecting buyers and sellers; for social networks, this means bringing together consumers, advertisers and software developers. Nonetheless, what these platforms have in common is that they are ‘multi-sided’ and exhibit strong ‘network effects’. Estimates of economic impact on industry Certain models have been used by OHIM and Rand and we found one (see Appendix 6.6.4) that could be usefully employed by adding social media as a variable, but only in relation to perfect copies where there is clear deception of the consumer into purchasing fake goods and where an argument could be made for losses to industry. This may have value for future research given that deception is made easier online because of the counterfeiter’s ability to employ authentic images and pricing close to that of the genuine article. It was easy to foresee an increase in the volume of deceptive purchases, which currently only make up a minority of purchases; accordingly, the model may still be useful in estimating the economic impact of such platforms on authentic firms’ legal offerings. Estimates of drivers for complicit behaviour This was based on Chaudhry’s model of complicit behaviour (see Appendix 6.6.5), which distinguished between extrinsic variables (product attributes, shopping experience and communications experience) and intrinsic variables (demographics, cultural values). Within the terms of the schematic model framework, the results suggested the dominance of extrinsic factors as drivers of complicit behaviour. In Fig. 1.4, only these extrinsic factors showed significant contributions to a stepwise linear regression model of complicit behaviour. These were the factors that showed significant effects and indicated how influential the social media experience was, with the impact being more than three times greater than any of the other effects captured in the study 133 The Economist (2016) ‘The emporium strikes back’ [Online]: http://www.economist.com/news/business/21699103-platforms-are-futurebut-not-everyone-emporium-strikes-back 47 Fig. 1.4 Relative significant contributions to the schematic model of complicit behaviour In addition, we examined the nature of fluctuations around an equilibrium-based model of economic impact. While we fully understand that the particular levels of harm that are derived from these equilibrium models are couched in too many caveats to be practically relevant, we suggest that the nature of the fluctuations might help to get a sense of the scale of the contributions attributable to social media. By following this approach, detailed in Appendix 6, we arrive at estimates of the proportion of counterfeit harm attributable to social media, as shown in Fig. 1.5. For the examined sectors, this impact ranges from 6.5–29%, demonstrating that while there is a material contribution from social media, it is unlikely to be a ‘silver bullet’. It should be recalled that the nature of this study, designed to be a test for the existence of an effect, means that these estimates are most likely to be at the upper end of the impact, rather than an unbiased estimate of the mean contribution. 48 Fig. 1.5 Sector-specific relative contribution of social media to the economic impact of counterfeiting 5.6.3 Overview According to our tracker and survey data, social media affected the sales volume of counterfeit goods. There could be little doubt that social media was having a substantial social impact and changing the way consumers interact with each other. That there is evidence that social media increases the likelihood of accessing counterfeit goods should not be surprising. Restricting the supply of ‘grey goods’, combined with ensuring that consumers can make an informed choice, are two key steps to limiting the negative impacts of counterfeit goods. This should form part of any education and awareness campaign and illustrates the need to better inform consumers. It also supports our conclusion that more resources are needed on the demand side of this issue to better understand consumers’ motives and attitudes. We also aimed to provide insight on other research objectives, including the impact of social media on producers’ reputations. However, this was not clearly demonstrable from the research despite claims made by industry (notably private enforcements agencies) about the damage caused by fake websites, including social media pages, on brands’ reputations. In relation to opportunities for countering infringement, we felt that this requires a great deal more work, not least in improving education and awareness, as well as in terms of social media platforms investing in efforts to actively counter IP infringement. The avowed zero tolerance attitude of Google to trademark infringement and the anti-fake drugs campaign by Microsoft’s Bing platform have shown what can be done to help industry and government counter infringement. Trading Standards provided a unique perspective on why particular social media channels are used over others (essentially, it was a matter of cost), and the consumer survey data highlighted how and why particular sectors of goods were targeted. 5.7 Summary, Recommendations and Conclusions 5.7.1 Research outcomes summary Scale of infringement Even though search engine Bing showed that there is a role for educating consumers about the dangers of counterfeit goods, there was little evidence that social media has been used to promote IPR. By contrast, there were many claims, from both industry and government agencies, about its role in facilitating IPR infringement, sometimes flagrantly. How infringement was distributed between the different sectors, products and types of IP was not always clear, but the survey and tracker indicated that certain goods (like tobacco and alcohol) were less likely to be promoted on social media. However, views expressed in the industry survey and road shows indicated that almost every ACG sector was impacted, just to varying degrees. The paucity of current scaleable official data, combined with the lack of current industry data and unverifiable industry claims, made it difficult to reliably assess the scale of infringement. This means we had to rely on data from the consumer survey and tracker to reveal how social media can contribute to facilitating 49 infringement. We note the high levels of suspect transactions revealed by the tracker, but this data needs to be supported by further regular frequent tracking of online consumer behaviour. Nonetheless, the scale of infringing activity indicated by the consumer data bears out many industry claims. Impact of infringement We explicitly included this issue within our questionnaire and the responses indicated that the impact varied across the different sectors, with some firms blaming the rise of social media for an increase in levels of counterfeiting and thus damage to their business. None of the firms surveyed were able to specify the actual costs to their business, and we attributed this to industry’s historical reluctance to share confidential financial information and a recognition that major brands readily engage with social media for sound commercial reasons, albeit sometimes as a defensive measure. It was clear that the social media platforms use similar ‘safe harbour’ defences to resist attempts by industry to get them to adopt more proactive policies for combating infringement, and this reactive attitude has created a climate of distrust and suspicion between these platforms and rights holders, which is made worse by what are seen as cumbersome takedown policies. The social media platforms argued that the IPR system within the UK is very fragmented and is part of an even more complex system across 150 other countries. While consumers who use social media are able to enjoy many positives, the dark side of internet-based commerce is shown by the ease with which both websites and social media pages can be manipulated to deceive consumers (although, from our findings, we still regard such deceived consumers as a small minority of those who use the platforms). The main focus for blatant infringement, according to industry and government agencies, is the proliferation of closed groups (i.e. invite-only groups, created on social media platforms). These groups clearly have a strong influence on infringement and this belief is borne out by our consumer research data, which indicates that IPR infringement is five times more likely in closed groups than in open groups. We consider this the most important finding of the project. Characteristics of infringement There were claims from industry and government agencies about the flagrancy with which IP-infringing content is placed on social media, although only FACT provided us with meaningful examples. We relied on explanations from the private enforcement agencies as to how counterfeiters were able to copy near-identical images from legitimate sites to deceive consumers. In certain cases, this involved near-perfect copies of certain products being priced close to the authentic article, completely bypassing the legitimate brand owners’ distribution and retail system. This infringing activity took place across myriad online platforms, not just on social media. The consumer data provided by us pointed to these deceptive copies as a growing threat, albeit one that still represented a small part of total infringing behaviour on social media. Deceptive purchases were more likely to occur with products like clothing, but were not characteristic of every impacted sector and product, least of all alcohol and tobacco. The bulk of infringing activity tracked in this study involved complicit consumers. However, we are aware that the use of VPN and the dark net means that much of the current online illicit activity is beyond oversight and reach. On social media platforms, the increased use of spambots and links to 50 various payment sources off-site makes it harder than ever to control the full scope of illicit activity. Opportunities for IP Microsoft’s Bing search engine has shown that online technology platforms can take an active role in combating IP infringement; in their case, this related to the offering of counterfeited drugs online. Google’s statement in relation to trademark infringement was compelling given its avowed zero tolerance for counterfeits, and we note the improved cooperation between eBay and rights holders. Recent changes to Facebook’s business model suggest that there may yet be opportunities to improve IP awareness, especially as they become more reliant on advertising from the brands whose goods are infringed within their platforms. It was evident that the online platforms are most likely to act against illicit activity on their sites if their own business interests are under threat. Education and awareness campaigns to date have illustrated the need to better inform consumers, but in relation to opportunities for countering infringement this area still required much more work and greater investment by the social media platforms in efforts to actively counter IP infringement. In the absence of greater cooperation from industry in supplying data, the focus of future research should be placed on disrupting the current levels of consumer complicity, and this is one area where the social media platforms could have a role to play. 5.7.2. Conclusion and recommendations Efforts to benchmark and compare data from the three key sources (government, industry and consumer) have only been partially successful. The methodological problems that beset most official data and estimates, as well as industry’s reluctance to share confidential and often real-time information, render these first two data sources inadequate measures of illicit activity in this market. None of the three main private enforcement agencies contacted were willing to supply us with more than headline data, even in anonymised form, that would have provided current insights on the scale of illicit activity on social media and other web platforms. This meant we had to rely on unverifiable assertions and claims made by these firms at conferences and within their published reports. That said, these ‘private’ enforcement agencies were better informed about current online (including social media) infringements of IPR and seemed best placed to provide current updates on their work. We recommend increased industry cooperation in supplying essential headline data for government and policy makers to better understand the trend in the market. This privileged and confidential information is always a more current and accurate reflection of the market than the data available from government and official sources, which are either out of date or methodologically unsound. Nonetheless, our tracker and consumer survey data provide meaningful, current (albeit snapshot) data, notably on segments such as levels of deception in online purchases and the ‘generation’ gap in online consumer attitudes and behaviours. There is also a strong argument for making more out of existing as well as new data sources, and the technologies for capturing digital activities. This is a point made by Coyle 134 in citing 134 Coyle, D. (2016) ‘The Sharing Economy in the UK’ [Online]: http://www.sharingeconomyuk.com/perch/resources/210116thesharingeconomyintheuktpdc.docx1111.docx-2.pdf 51 a key finding of the interim report of Sir Charles Bean’s Review of Economic Statistics.135, 136 As Chaudhry and Zimmerman argued, we believe that more reliable and meaningful insights can be gained from consumer data and there should, in the future, be a much greater focus on researching the demand rather than the supply side of counterfeiting. The consumer data presented in this study has shown how social media plays a role in facilitating IPR infringement, particularly in closed groups, but the data presented represents a mere snapshot from the middle of 2015. The lack of any other comparable data means this cannot be used to provide a definitive indication of the development of this phenomenon over time. A methodology that allows an assessment of both stated and revealed preferences, such as the one we have employed within this study, is, we believe, desirable as a more effective and reliable measure of illicit activity. We also believe that there is a need for a single methodology for more frequent, longer-term research (comparable with the Ofcom/Kantar survey for online copyright infringement) to provide a unique data set as the basis of an official national measurement. We would argue that developing a greater understanding of consumer motivation for purchasing counterfeit goods is pivotal at this time. Our findings suggest that consumer behaviour is nuanced and encompasses complicit behaviour (favouring non-obvious copied goods), which has a lower economic impact, and deceived behaviour, which tends to involve high-quality/highpriced goods and represents a greater potential threat to brands in the future (even though, at present, these deceived consumers make up a small proportion of the total consumers impacted by counterfeit goods online and through social media). The full extent of the challenge from social media may not be entirely clear from this study, but from industry sources we believe that this is growing and will include more deceived consumers as the sophisticated tactics from the counterfeiters become ever more elaborate. We considered other challenges that may yet exceed those posed by social media, and these include: the arrival of messaging platforms suitable for mobile use, where it is harder to track illicit behaviour given their embedded encryption technology; blockchain technology, which may enable counterfeiters to further hide the financial benefits of their activities; and 3D printing, where the potential exists to considerably increase infringement across all the main IPR. Most of this activity could occur outside of any kind of scrutiny. 135 UK Government (2016) ‘Press notice: “Take economic statistics back to the future,” says Charlie Bean’ [Online]: https://www.gov.uk/government/publications/independent-review-of-uk-economic-statistics-final-report/press-noticetake-economic-statistics-back-to-the-future-says-charlie-bean 136 Bean, C. (2016) Independent Review of UK Economic Statistics {Online}https://www.gov.uk/government/publications/independent-review-of-uk-economic-statistics-final-report 52