Falling in love (again) with Digital Analytics in Higher Education (PSEWEB 2015 Keynote)

Editor's Notes

1. Hello everybody, I gave this keynote at the PSEWEB conference on July 28 and I thought that you guys might be interested in it. So, this is the reason for this screencast – recorded just for you – since context is always important, I chose to edit out the slides I used to break the ice with the audience at the beginning of this talk. Let’s start then!
2. Before meeting digital analytics, my life was sad and gray. I worked for a radio station – which was actually quite fun and even geeky – but when it came to measurement and analytics, those were the Dark Ages
3. Day in day out, we would be behind the mic putting some content for an audience we couldn't see And, the only way to find out how we were doing was to wait for the audience phone surveys that were done only twice a year and that would give us 1 metric: the number of listeners
4. Fast forward the new millenium + 6 and google gave us the most beautiful flower bouquet: Google Analytics – finally, it was possible to search, find and share measurement data in real-time. For me, it was love at first sight.
5. And, my love for digital analytics in higher education quickly turned into obsession. I wrote dozens of blog posts & articles, conducted surveys on how we measure in higher education, gave presentations, organized online conferences – including the higher ed analytics conference taking place in February every year. – And, more recently developed a curriculum of professional certificate courses to help others fall in love with analytics.
6. While I don’t think my relationship with analytics is totally healthy – and my therapist agrees. It seems that we all have some issues to work on when it comes to analytics in higher ed And, I'd like you to take 10 seconds to think about it.
7. If you could display your analytics relationship status on your Facebook profile – what would YOU pick: married for life, engaged, separated, divorced... Or it’s complicated – like many of the digital professionals working in universities and colleges I’ve met at conferences or online over the years.
8. All these conversations I had taught me that the relationship we have in higher education with analytics and measurement is indeed often complicated
9. We don’t want to talk about it in public. I made the mistake over the past few months to ask web teams about the bounce rate value of their homepage. And, I got the type of answer you usually get when you ask a lady about her age. In higher education, analytics feels strangely personal, deep, private. Photo Credit: Soumyadeep Paul https://flic.kr/p/dTSnUc
10. Our relationship with analytics is often complicated because it can be complex, and you know, numbers can give us headaches. Photo Credit: [3/4 of zer0] https://flic.kr/p/5o4TqX
11. Yet, vanity metrics are so tempting, they are easy and makes us feel SO good about ourselves. Photo Credit: Anna Gearhart https://flic.kr/p/9cF3hd
12. Our relationship to analytics in higher education is complicated, because size seems to matter SO MUCH. And, it looks like the bigger the number, the better – yet, deep down, we know it's not always the case and can feel something is just not quite right Photo Credit: Jeff https://flic.kr/p/5vM8NW
13. But, then why is it SO tough to get it right? Like for any other relationship issues, there isn't just one culprit. Let's talk about these for a minute.
14. It's not them, it's us. In higher education, we are somehow wired to experience more issues when it comes to measurement and analytics.
15. Because most institutions are public and non-profit, accountability hasn't always been high on the priortiy list of decision makers. Universities don't get closed because they didn't meet their sales quotas – although this is starting to happen for some for-profit in the US. In our line of work, budgets aren't lost because they were wasted on marketing channels that are not reaching our audience. Non profit also means no big pressure to perform, and no need to measure performance Photo Credit: KMR Photography https://flic.kr/p/q9W2Eu
16. While this might not be true for the developers or the more IT oriented professionals in this room, most communicators and some marketers are not crazy about maths. And, there's a little bit of it involved in digital analytics. Which can feel intimidating when you chose to study communications and not engineering. Photo Credit: Kevin Harber https://flic.kr/p/5wh8KN
17. With the explosion of digital over the past decade, there is also a question of time involved. Engaging in a committed relationship with analytics takes time and a certain degree of commitment – that not everybody has.
18. Then, we (or our bosses) are often too scared to care about analytics and measurement. What if we end up seeing we failed – failure is scary, we don’t want to know whether or not we were successful, what if the numbers are bad? What if? Photo Credit: Paolo Margari https://flic.kr/p/432C1w
19. So we have issues we need to work on, but we are not the only ones. It's not always us, it's them too
20. Numbers drive action(s)
21. Want an example? Any fitbit users in the room? I don't know about you, but When I first got my fitbit, I worked very hard to get my daily step count up. So, I started walking more, then running, then running faster, then running faster for longer periods of time – anything that would impact the step count. Biking – even if it's good exercise and can be really nice – didn't automatically impact that metric, so I stopped biking. – and, then my fitbit battery died and I stopped everything – but that's another story
22. Speaking of biking – the bike sharing service in Montreal, Bixi, does provide some analytics to members – which is a nice gentle way to keep people engaged and to push them to use the service more. Numbers drive actions indeed.
23. Even vanity metrics, which is why they might be vanity metrics for your school, but they are/were key performance indicators for social networking websites early one. That's why Facebook at the beginning was a friendsracing game, the more friends the better – not for facebook users, who can keep up with 500 friends – but for the platform itself
24. Because when you want to sustain your business by selling advertising, the size of your network and the total numbers of eye balls do matter a lot.
25. Even analytics tools don't make things easy. Google Analytics or the analytics feature of the social media platforms provide so much information. Too much data to actually be useful, or usable – especially when you are not working full-time on performance measurement. Those are really powerful, incredible tools but they are also quite complex. https://flic.kr/p/5RSJp Andy / Andrew Fogg
26. For all these reasons, the love for analytics in higher education doesn't come easy
27. But, when it happens, it results in beautiful love stories. And, more and more of your colleagues, in Canada and the US, are living proof that analytics love can overcome all these obstacles. So, now I'm going to share 8 of their stories. – to help you rekindle or spark your flame for analytics in higher education.
28. Let’s start with an issue many social media professionals face: the challenge to measure social media impact beyond reach and engagement on the platforms themselves.
29. Liz Babiak, social media community officer at Algonquin College in Ontario, managed to do it.
30. She tags ALL outgoing links on social media with utm parameters for Google Analytics – Hootsuite makes the whole process pretty easy as you can see.
31. Then, Liz can easily measure the impact of her work on website traffic – and down the line, segment what these social media visitors end up doing on the college website. It’s also a great way to see the type of content used to answer social media questions. When the college noticed they were answering many questions related about tuition, the team reevaluated how this information was presented on the web. And, made it more visible.
32. Another challenge with social media is often to communicate its impact on the institution at large – especially to decision makers like the president or the board of trustees – who don’t have the time to go over lengthy reports.
33. At the Ohio State University, Ted Hattemer, the AVP of interactive communications, found a way to communicate social media impact and progress to its board of trustees with 1 main metric, the social media score.
34. This high level score is presented as part of the social media report card you can see on the screen. Ohio State decided to focus on Facebook and Twitter for social media.
35. The metrics used on that report card were carefully selected to be aligned with Ohio State strategic goals for social media. So for example, in the case of sustaining dialogue, they chose to track specific metrics and give them a score on a scale from 1 to 10 – 10 representing the stretched goal – the target for their goal. A that is defined by taking into account their past performance and the performance of a selected group of peer institutions. So for twitter conversation, they would score a ten, if they had that month 14K tweets, replies and favorites. The score is really interesting because it gives an indication on how far they are from their goal.
36. The approach is quite thorough, yet the resulting social media score is easy to understand. It also communicate very well impact and progress to the busy members of the board of trustees.
37. Obviously in higher education, social media professionals are not the only ones facing challenges with analytics – people taking care of websites still have some serious battles to fight – especially when it comes to homepage real estate.
38. Well, Frederick Oliver, digital marketing analyst at Sheridan College – and a graduate of our program on advanced and predictive analytics for higher ed did win that battle.
39. I’m mentioning the program because the analysis he performed for his final project helped him win a prime spot on the homepage to promote Admissions Open House at Sheridan.
40. By segmenting and analyzing referral data patterns in Google Analytics, Frederik was able to prove that this homepage banner directly contributed to open house registrations: sessions including a click on this homepage banner had a conversion rate of 22% vs 0.5% for the other sessions. He also proved that the homepage does still play an important role in the registration conversion funnel – and that users do actually click on banner presented in a carousel 
41. Not all homepage features can directly be linked to a straight conversion like in the previous case. So, how can you show the economic value of homepage feature stories – when it’s time for example to justify or request more budget?
42. Todd Gregory , digital analyst at Penn State University managed to do just this.
43. How did he do it? First he looked for the market value of digital real estate to help him define the economic value of the homepage feature – in this case, he used the rate for digital advertising on the Chronicle of Higher Education website.
44. Then, he came up with a CPM (cost per thousand) rate for the feature.
45. Factoring in traffic, he ended up with an economic value for the homepage feature – in a way that made total sense to traditional communication decision makers because this is basically how we used to measure the economic value of PR or media relations.
46. Another challenge faced by web professionals is to sell change – any change. To get buy-in from decision makers and stakeholders
47. If you attended yesterday the presentation given by Adrian Liem, Senior Web Coordinator at the University of British Columbia, you have already heard about this story, but I really wanted to share an example from this group.
48. What Adrian and his team managed to do with analytics is to validate user behaviors noticed in usability tests. If you ever participated or conducted a usability test, you know that they can help you pinpoint issues – but, because they rely on a small sample (5 to 10 users), it can be hard to get buy in from their results only. This is the pre-redesign page. The tests found that users were doing some back and forth between this page and the homepage when they didn’t find the Admissions information they were looking for.
49. Adrian and his team knew that the second-level pages were causing the issue. But by looking at the visitor flow in Google Analytics they were able to confirm their hunch: the behavior observed in tests was indeed representative of the larger population. – which led and informed the redesign of these pages you can see.
50. In digital, anything can be measured – but sometimes the issue is to set things up properly in order to track and measure what can be.
51. At BCIT, Alan Etkin, senior systems analyst found the holy grail of higher ed analytics: measuring the direct and indirect impact of digital efforts on the bottom line – all the way to the tuition dollars paid by enrolled students.
52. How did he do that? First, he got all the pieces in order – making sure he will capture what could be tracked in Google Analytics: digital advertising, social media posts, landing pages, submitted online forms up to the student registration system, Banner – this wasn’t easy, but he managed to convince the IT group in charge of banner to let him track what happened in that black box.
53. Setting things up and tracking everything was the tough part – once this was done, Alan was able to put a dollar value on many initiatives. Knowing - for example - how many students attended an information session, how many applied and how many registered – Alan could come up with an average dollar value for every attendee at these sessions. This type of information is really useful when you want to make a decision on how much should be invested to market these sessions.
54. If you work with academics or scientists at your school, I bet you’ve heard the line about correlation & causation. And, often with digital analytics in higher education we show strong correlation between what we did and what happened – but NOT causation. Yet, wouldn’t it be nice to prove with a high level of certainty that your work saved the day?
55. That’s what Joshua Dodson, who teaches Analytics for Higher Ed Experts did a few months ago when he was still working at Eastern Kentucky University.
56. With the help of predictive analytics techniques used for forecasting possible future outcomes with past data, Joshua was able to demonstrate the value of his work on a specific project.
57. An external partner was doing all the marketing and lead generation for a given program. It was decided to bring all this in house and increase the organic SEO efforts. After a few months, Joshua was able to demonstrate – using predictive analytics technique – the impact it had. By showing what would have happened had they kept working with the outside vendor – the dotted line at the bottom of this chart. While his intervention had resulted in the dark line at the top – the actual traffic data. A big win caused by his work.
58. As you can see the possibilities are endless. So, how can YOU build a healthier relationship with analytics?
59. Today, I'm not going to give you recipies. The goal of this keynote is more to inspire you and make you reflect on your own practice. So, I want to share guiding principles and then I'll get the love doctors – the 2 experts who teach our courses on analytics for higher education, to weigh in with their best advice.
60. Don’t measure for the sake of measuring. It’s pointless. As we saw numbers drive action, but you want to make sure to drive the right action.
61. That’s why it is so important to define and find your true love. What is the ultimate action you want your audience to take? If you removed everything else, what would you need to achieve? Once you have found how to measure your ultimate goal, measure and cherish it because it is your true analytics love.
62. For anything else besides your true love, always follow your heart not numbers. If you start with data, you will get lost -- it is addictive. So, the best way is to start with a question and then look for answers in the data
63. Put your flowers to good use. Tools don’t come first, but it is important to set them up and use them wisely. For example, switching to Google Tag Manager does make tons of sense if you want to get to the next level with Universal Analytics and tags or pixels from Facebook, Twitter or display ad networks. So, make sure that you have your tools ready and in order.
64. True love and flowers aren’t always enough. Analytics love can take time and efforts. So, if you feel like you’re going nowhere, get help. If you have time, learn on your own – Google Analytics does offer some great resources – even if they are not specific to higher education. Many of our alums in web analytics chose to take our professional certificate courses because they wanted more specific guidance. Alternatively, if you really need to deliver fast and budget is not an issue, find a knowledgeable consultant.
65. Once you are in love with analytics, share the love. Educate your boss, your internal clients, your stakeholders so they get it too and you can bring it to the next level with them.
66. Numbers drive action, but they can also be overwhelming, so be mindful of what you include in your reports. Kill the traditional report – or make it dead simple. Your report goal is to drive action: present data that matters, spell out your insights and wrap it up with action-oriented recommendations.
67. Now, it's time for the Analytics Love Doctors – who taught the art of higher ed analytics to many of your colleagues around the world in our professional certificate online courses.
68. We are going to start with Liz Gross. Liz is our social media analytics prof – and she recorded this advice for all the professionals in this room who are in charge of social media
69. Joshua Dodson is our Web Analytics Prof. He has been teaching our introductory course (as well as 2 other advanced courses) for 4 years. More than 200 of your colleagues around the world have taken his courses. Here's the piece of advice Joshua chose to share with all the digital professionals in the room today.
70. If you are already doing all what I have talked about – congratulations, and please come share your story with me so I can feature it in future posts or articles and maybe you can present at the next higher ed analytics conference in February. But, if you are still yearning for more, brace yourself for what's coming next. Analytics are coming to physical campuses. That's what Universal Analytics (the latest version of Google Analytics) is all about with the measurement protocol - integrating data from non-digital sources and measuring everything.
71. I wrote about the possibilities last year. But – a few weeks ago – this article in the Chronicle of Higher Education made it clear that the future is almost there as Google is funding this experiment with sensors on the Carnegie Mellon University Campus. Ultimately, it will be possible to measure activities taking place offline on campus and integrate them with online activities. Yes, the futur http://chronicle.com/blogs/wiredcampus/new-model-of-smart-campus-carnegie-mellon-to-embed-sensors-across-landscape/57079 http://collegewebeditor.com/blog/index.php/archives/2014/02/26/universal-analytics-a-game-changer-for-highered/
72. And, that’s it. If you have question, I think we have time for a few and you can always email me or DM me on twitter. I will also be around later today.