Capturing Value From IT Infrastructure Innovation
Capturing Value From IT Infrastructure Innovation
Capturing Value From IT Infrastructure Innovation
b u s i n e s s
t e c h n o l o g y
o f f i c e
Innovation is a critical part of the mandate for enterprise infrastructure functions. When we interviewed senior infrastructure executives about their innovation priorities and capabilities,1 many emphasized that constant innovation represented their only chance of meeting user expec tations while supportingwithin budgetary constraintsever-increasing business demands for computation, data storage, and connectivity. Cloud computing was the most prevalent
However, most leaders of infrastructure functions appear less than satisfied with their units ability to innovate. While they report significant progress over the past few years, most say their organizations do not have a repeatable process for extracting value from innovation and do not devote enough resources to it compared with running day-to-day operations. Moreover, many IT functions are not seen within the company as a source of innovation (Exhibit 2). This article presents seven key findings on how infrastructure leaders view their innovation priorities and how they are managing their organizations and vendor relationships to support their agendas.
innovation theme among the executives we interviewed (Exhibit 1). But many infrastructure organizations are also aggressively seeking to support data analytics and provide richer end-user experiences. Whats more, many of the infrastructure functions responsible for security technology are prioritizing investments in ways that will better secure their environments from cyberattacks.
described in this article, we interviewed 30 executives with enterprise-wide infrastructure responsibilities for Fortune 500 companies or institutions of equivalent size. The executives represented companies in the following sectors: banking, consumer electronics, e-commerce, health care, insurance, media, and retail.
Dan Page
Takeaways
IT infrastructure innovation is increasingly critical to meeting business expec tations, but selecting the right technologies is only half the job. Integrating technologies into services that meet business-user needs and building scalable processes for provisioning and support represent the more compli cated and challenging part of capturing value from innovation. Cloud computing, data analytics, and end-user productivity are the most frequently cited innovation priorities. Most infrastructure executives dont believe they have the right set of capabilities in place to get the most value from innovation.
when providing service and conducting dayto-day operations for their organizations. While many of these leaders summarized their inno va tion agendas using technological terms like cloud computing or desktop virtualization, they point out that selecting the right technologies is only half the job. Integrating technologies into services that meet business-user needs and building scalable processes for provisioning and support represent the more complicated and challenging part of capturing value from innovation. In addition, some infrastructure organizations have an innovation agenda that is more operational than technological. They are focusing on changing their interaction model with the business, making end-to-end processes lean, and automating critical functions like change and event management.
Infrastructure executives voiced a fair bit of discomfort with how cloud computing has been presented in the business and trade press. Several described their private-cloud programs as the natural continuation of years of consolidating, standardizing, virtualizing, and automating their server environments. A few infrastructure leaders described cloud as a marketecture term and said they preferred to talk about utility computing or virtualization 3.0. That said, the largest infrastructure functions, particularly in heavily regulated industries like banking and pharmaceuticals, plan to build private-cloud environments over the next three to five years rather than procure publicly available cloud services. Why? As might be expected, there are concerns about application availability, data security, and regulatory com pliance. Several infrastructure leaders said that they believed public-cloud providers have not yet developed the capabilities required by the most complex infrastructure functions. Data location, for example, is a key regulatory requirement. One infrastructure head said, We have data that cant leave the United States, and we have data that have to stay in Europe, but we havent seen how the public providers can support that. Other infrastructure heads pointed out that the process of contracting for enterprise cloud services was not yet mature and that the contractual terms and conditions required for a successful agreement would only emerge over time through trial and error. (See sidebar, Building a private cloud, on page 7.) A few infrastructure organizations, especially those with annual budgets in the hundreds of millions rather than billions of dollars, did express openness to using public-cloud offerings for infrastructure services. Infrastructure leaders considering this option believed that they could
58
Low-cost data storage and access platforms for big data and massive analytics Shared data services Integrated communications, messaging, and presence across voice, desktop video, e-mail, and chat Shared collaboration and knowledge-management platforms and tools, including social networking Deployment of tablets Adoption of bring your own device strategies for smartphones and tablets Thin client devices for the end-user desktop Stateless desktop or provision of browser-based desktop to a wide range of devices Enhanced instrumentation of technology environment to provide insight into emerging attacks Data-loss protection and enhanced identity and access management Automation of core operational functions Migration to a product- or service-based operating model
53
Enterprise mobility
42
Virtual desktop
32
Security
26
16
Low latency
11
not get top-management support for the capital required to build out a private-cloud environment and did not believe they had the skills and talent in their organizations to manage a private-cloud program. In some cases, they saw the use of publiccloud services potentially leading to a fundamental shift, with the infrastructure organi zation becoming much more of a vendor-management and governance function for external services selected and paid for by business unit IT.
virtual-desktop services, and enterprise mobility. There are a number of reasons for this enhanced productivity focus. Technology-savvy employees are increasingly capable of using productivity tools effectively and are more and more confident about demanding tools that meet their needs. Highly functional, easy-to-use, and low-cost (or free) consumer-oriented tools like Dropbox, Facebook, and Skype raise users expectations. At many companies, the most critical employees (in functions such as sales, marketing, research, and design) depend on end-user technology tools rather than business applications to enhance their productivity. Despite the increasing focus on end-user productivity, there is much less consensus on what steps to take and what the value proposition will be than there is for cloud computing. Some organizations are implementing unified-communi cations platforms (integrating chat, e-mail, knowledge management, video, and voice for a seamless user experience) based on expectations that they will facilitate employee productivity. Others have not been able to make the business case to senior management. Some organizations are incorporating desktop video into their unified-communications plans; others do not believe that it justifies incremental bandwidth costs. Some organizations focus unified communications on chat, video, and voice. Others plan to integrate collaboration and knowledge-management capabilities tightly. Some organizations are partially funding unified-communications investments by removing telephone handsets from user desktops; others say their users will never stand for that. Perspectives are similarly split on how to deploy virtual-desktop services, which separate the end-user desktop environment from a physical machine and can potentially simplify support,
23 12 6 1 2 3 4 6 5
We have a well-dened process and organizational constructs to stimulate and capitalize on innovative ideas
30 18
30 18 6
We invest enough time, resources, and funds in identifying, prioritizing, and implementing innovative ideas
41 35
12 6 1
1Figures
6 2 3 4 5
Exhibit 3 Bring your own device strategies have rapidly gained traction.
% of respondents1 14 Wary 29 Evaluating Bring-your-own-device (BYOD) models being considered and used Companies may subsidize employee device purchases and services or simply allow users to access corporate IT services through their existing devices and services without reimbursements Rationale for BYOD 29 Planning or implementing Improves user experience: employees no longer need to carry 2 devices Eliminates an asset-management headache Reduces cost Issues and challenges 29 In place Getting appropriate security model in place Determining how much control to exercise over user device
1Figures
improve data security, and allow users to access their desktops from a range of devices. Most infrastructure leaders report good experiences with early virtual-desktop deployments and believe that they offer flexibility and security advantages that a traditional desktop environment cannot match. Much less consensus, however, exists about the economics of virtual desktops. Some infrastructure leaders report savings of 30 percent compared with traditional environments because of lower costs for moving, adding, or changing desktop equipment and providing user support. They therefore plan to migrate much of their user base to virtualdesktop services over time. Others see virtual desktops costing about the same as or a little bit more than traditional ones because of back-end hosting costs and therefore plan to limit virtual-desktop services to targeted user segments. It seems that the difference between these perspectives stems from variations in virtual-desktop solutions (for example, how much storage is allocated to each user) and the efficiency of back-end hosting.
procuring devices and services that duplicate ones that the employees buy for personal use. Even without subsidies, there seem to be a signi ficant number of employees who would still take advantage of the service. One institution launched a voluntary pilot that did not include subsidies but was nevertheless quickly oversubscribed. Most of the executives we interviewed are at some stage of planning and implementing programs to replace their corporate BlackBerry and mobile-phone support with a BYOD program. In almost all cases, the program is voluntary for employees, and uptake has been greater than expected. These programs most commonly provide e-mail and calendar synchronization for iPad and iPhone devices. Several interviewees are considering leveraging virtual-desktop technology to extend BYOD programs to laptops and desktops, though none of the infra structure groups we surveyed had rolled this out at scale. In part, they are considering doing so because they have users who work from home, and BYOD programs would eliminate the cost and complexity of managing devices in employees homes.
for possible innovations to business and application leaders. Two organizational constructs have been helpful in accomplishing this: A few infrastructure groups have set up small organizations, with funding, to work directly with business owners on figuring out ways to leverage technology in solving business problems and transforming go-to-market approaches. At least one is seeing fairly substantial change and impact in how the company engages its customers at branch sites through the use of wireless technology. Others have created product-driven models, in which infrastructure product managers oversee an overall road map for end-user services or appli cation hosting that integrates innovation initiatives with more tactical investments, for example, in expanding capacity and refreshing technology assets. This model enables infra structure organizations to innovate much more proactively, but it also requires a constructive working relationship with businessunit IT for success. In addition, some infrastructure functions have managed to be proactive at innovationwithout creating specific processes or organizational constructsby creating space for talented staff to focus on innovation. A couple of companies hold hackathons where for a day (or night) staff will form small teams on topics that are outside their typical day-to-day activities and build solutions that improve performance. Moreover, despite the wide range of funding models in use, most participants finance innovation as part of the regular investment process. In most places, dedicated innovation funds or budgets are limited to small investments for proof-of-concept activities.
All have started by hosting development and All have also started with infrastructure as a test environments. Many are only now making McKinsey On Business Technology 2012 Infrastructure Innovation service (providing virtual services on demand) the transition to hosting production workloads. Exhibit 4 of 4 but expect to offer more bundled platform Most infrastructure heads believe that they can
Exhibit Most large enterprises are considering or using public or private clouds.
% of respondents 5 Posture No cloud plans Open to public cloud Description Have not moved material workloads to public or private cloud Not considering doing so Considering or starting to migrate workloads to commercially available public-cloud services Developing initial plans to build or source a private or virtual-private cloud Have a private cloudtype environment up and running; may be using it primarily for nonproduction workloads May be using a vendor to build and manage private-cloud environment May be considering using public-cloud services for noncore, noncritical workloads Typical rationale Cloud architectures cannot meet latency requirements for core workloads Do not have access to capital and skills or capabilities required to build a private cloud Private-cloud platforms can host 6070% of workloads at 3040% better efciency and vastly increased exibility Private cloud is more economical, given scale and need for vendor margins Public-cloud providers cannot at this time meet enterprise resiliency, security, and compliance requirements (eg, data location) Contracting model for enterprise cloud services still highly immature
25
10
60
services over time, with the ultimate ambition of moving to a DevOps-type1 model, which provides appli ca tion teams with a high degree of self-service. Several infrastructure leaders said that their cloud environments would have to do this in order to meet business expectations for speed and efficiency. Building a private-cloud environment does not necessarily mean building it yourself, and there are many ways in which large infrastructure organizations are leveraging vendors, including project-based system-integration arrangements and long-term outsourcing deals, to not only build a dedicated cloud environment but operate it as well. Getting the governance, operational, and service models right has been a bigger challenge than the technical engineering of the cloud environ ment. For many organizations, putting a cloud environment in place means completely recasting their service catalog or making the difficult transition from a project-based funding model to a service-based one. This means moving away from allocating capital to individual application-devel opment projects and toward making investments in shared technology services that will be leveraged by many projects.
Other organizations are wrestling with the operational changes required to extract value from the cloud. The manual change-management process required for migration to one-touch provisioning dramatically reduces the speed and productivity benefits that cloud computing promises. Experiences with migration have been mixed. In some organizations, developers are eager to use flexible, low-cost cloud platforms. In others, they have expressed wariness about no longer owning serversa sentiment that slows down adoption and migration. Most private-cloud programs use relatively higher-cost technologies from leading vendors, with infrastructure leaders saying that the service and support available justify higher hardware and software prices. A few institutions that have a high degree of confidence in their internal technical capabilities can leverage low-cost components like scale-out storage, commodity servers, and hypervisors with open-source software.2 One infrastructure head, for instance, is experi menting with using scale-out storage from a low-cost storage provider as well as opensource software stacks for development and test environments.
1 DevOps is a software-
development method that stresses collaboration, communication, and integration among software developers and IT professionals. 2 Scale-out storage enables organizations to improve application performance and scalability on an incremental, as-needed basis by adding multiple replicated database servers on low-cost commodity hardware. A hypervisor is a systems-software layer that allows multiple operating systems, termed guests, to run concurrently on a host computer.
commercial insight, communication skills, and creative problem-solving skills to drive innovation initiatives. A few leaders pointed out that their organizations could introduce only one or two real innovations per year because they had to be personally involved to ensure success.
Infrastructure organizations are experimenting with a variety of levers to overcome these challenges and build the talent required for innovation: E mpowering engineers by expanding their
heart of the infrastructure organization. Hiring managers from high-technology vendors brings insights about leading-edge technologies and commercial product-development processes.
10
sourced staff may have less insight into and connection with business needsinsights and connections that can drive innovation powerfully. In other cases, infrastructure leaders said it was hard to get past vendor sales teams to speak with vendor thought leaders about particular technologies. Another occasional hurdle was that sales teams tried to work with application developers to introduce technologies that either could not be supported or did not align with the companys technology. A few infrastructure leaders also expressed concerns that vendors devoted too much of their development resources to incremental features rather than advances at the underlying level.
need to bring innovations to the table. This model is more common in mature and wellunderstood areas that are outsourced in a managed-service modelfor example, end-user desk-side support.
...
Leaders of IT infrastructure functions face a formidable innovation challenge. To make the most of technology opportunities like cloud computing, unified communications, mobility, and virtual desktops, these leaders must integrate technology with new services and processes that meet business-user needs. That means engaging proactively with the business side, finding the right talent, and building vendor relationships that support innovation.
Henrik Andersson (Henrik_Andersson@McKinsey.com) is a principal in McKinseys Silicon Valley office, James Kaplan (James_Kaplan@McKinsey.com) is a principal in the New York office, and Brent Smolinski (Brent_Smolinski@McKinsey.com) is an associate principal in the Atlanta office. The authors wish to thank the infrastructure executives who participated in the interviews that this article is based on. Copyright 2012 McKinsey & Company. All rights reserved.