I attended the IBM Business Analytics Analyst Summit in Ottawa and while I can’t tell you much about what was discussed there due to confidentiality restrictions that will be released shortly, I can share with you some of my own observations regarding the state of BI, particularly what’s wrong with it. By “wrong,” I mean why aren’t adoption rates higher? Why aren’t users more satisfied? Our Ventana Research benchmark research on BI indicates that only 37 percent of organizations are satisfied or very satisfied with their BI efforts.
Here’s my theory on what’s wrong with BI: The basic problem is that software vendors are trying to bring users to BI rather than bringing BI to users. Do you remember the phrase “BI for the masses”? In a 2005 article Industry influencer Cindy Howson already described it as a “once revolutionary phrase that has become a cliché.” One way I have seen it used by SAP Business Objects and others was to describe new and improved user interfaces that would make BI so much easier to use that adoption rates were certain to increase dramatically. It also has been used to describe browser-based interfaces that would reduce the barriers to adoption by eliminating the desktop installation process for BI.
New interfaces and easier administration are both valuable advances, but they are not what’s needed to spur widespread adoption of BI. In short, many software vendors have been focused on solving the wrong problem. They’ve been working in the labs and doing research looking for a silver bullet that makes BI so easy to use that everyone will want it. Don’t get me wrong. I applaud these efforts, and there have been significant and valuable advances as a result of these efforts. In fact, our benchmark research on business intelligence indicates that usability is sufficient in over half the cases. But if usability is sufficient, why hasn’t there been more adoption?
Analysts and expert BI users certainly can do more types of analyses, quicker and over more data than they could just a few years ago. However, these changes haven’t solved the problems either. Adoption rates and satisfaction levels for BI tools are still far lower than most organizations would like. I think this approach hasn’t worked because BI is not a single universal function that can be delivered to everyone in every organization in a one-size-fits-all approach. In vendor speak, the marketing claims often sound something like this: “Deploy our tool and all your BI problems will be solved.” This simply is not the case.
For truly widespread adoption, BI needs to be packaged and delivered to business users in bite-size, pleasing morsels that are easy to swallow. Or using a different analogy, you might even say “invisible” – users shouldn’t think they are using BI. They should think they are just doing their job – whether that is managing payables and receivables, sales pipelines, new product research, workforce management or any other performance-related activity.
Where do most users spend their time? They spend it in e-mail, spreadsheets and the applications (ERP, CRM, finance, etc.) that are part of their daily jobs. According to our research on BI, in over two-thirds of organizations use of Microsoft Office is common. If they are like me, not only do they use e-mail, but they live by their Outlook calendar and maybe even use the to-do list/task features. That’s where BI needs to be – in the productivity tools and applications that are used every day whether they are desktop, browser-based or mobile. Asking someone to use a separate tool creates a barrier to adoption.
If you follow Ventana Research, you know we research how people, processes, information and technology all affect the performance of an organization. BI software vendors naturally gravitate to the technology elements, but BI is also a process – a decision-making process designed to improve the performance of an organization. BI vendors need to support that process, not just deliver numbers. When looked at as a process, the importance of collaboration becomes obvious. Our benchmark research indicates that for 64 percent of organizations collaboration technologies are important or very important as a supplement to BI – that’s higher than they rank search technology. And by the way, the term “collaboration” in the enterprise software market can mean many things, which causes confusion. I suggest we think of collaboration in terms of shared activities necessary to support a business process. It begins with the delegation of a task to an appropriate resource, monitoring the status of the task, capturing the history and discussion associated with the task, and knowing something about the people working on the task. When you turn BI on its side and look at it as a collaborative process rather than simply crunching numbers, several good things happen:
You understand that BI is a means, not an end.
The delivery vehicles for BI become more obvious, and they are not just reports or dashboards.
You can more easily support the governance, risk and compliance requirements of your organization.
I shared many of these same thoughts with the Business Analytics team at IBM. You will have to draw your own conclusions about whether it is because they have tackled some of these issues or if they still need to in its upcoming product releases. Regardless of whether you are an IBM customer, you should be asking your vendor about their plans for supporting these capabilities.
David Menninger – VP & Research Director