« Progressive Financial Services Firms Demonstrate the Value of Predictive Analytics | Main | EDM and interesting business technology trends »
December 7, 2007
The future of BI had better include EDM
(Posted by guest blogger, James Taylor)
A number of posts on what the future of BI is likely to bring caught my eye this morning and I thought I should add my thoughts on how I see BI evolving into EDM. Several trends, I believe, are going to force people to move beyond thinking about BI to thinking about EDM.
- If BI is about actionable insight, EDM is about insightful action
BI practitioners are focused on delivering "actionable insight". Not data, not information, but information that adds new knowledge to the person consuming it and does so in a way they can take action. This is all fine and good but there is no guarantee that the action will actually be taken. The person seeing the insight may not understand or feel authorized to take the action. The action might need to be taken in an automated way to support a web or other real-time transaction. EDM recognizes this need to ACT not just to know. - Operational BI is (almost) an oxymoron
Operational BI has several problems. Firstly, BI tools are mostly focused on analyzing the past and so are insufficiently real-time to deliver something "operational". Secondly, operational systems are often used by people who do not have the analytic skills to use BI. Thirdly, operations are increasingly 100% automated or "straight through" and so there is no-one there to see the results of BI at all. EDM, in contrast, is tightly focused on the kinds of high-volume, rapid-response, operational decisions involved in operational systems. - BI tends to focus on the past
The use of BI to understand what has happened in the past and to be able to see both what happened and why is necessary to a business improving itself. But it is not sufficient. Effective decision-making requires predictions about the future not just understanding of the past. EDM revolved around the use of predictive analytics to help make decisions that take into account not just what is known about a customer or a transaction, but what can reasonably be expected of that customer in the future. - Business process management needs analytics but not BI
The functionality common in BI systems is not very helpful once an organization starts automating business processes using a BPMS. The ability to take process information (state, history) and data and help a process complete more effectively requires EDM not BI. BI might help you understand what happened in all the various instances of a process but it is not going to help you improve each instance, at least not if you have very many instances! - Traditional BI does not add enough to Enterprise Applications
With many of the big traditional BI vendors now bought by enterprise applications vendors we are going to see an interesting issue arise. While companies have been frustrated that they cannot use the information resources created by their enterprise applications, they are also frustrated by the inability of the enterprise application to learn from this data. The growing support for BI from the EA vendors will help companies understand the data in those enterprise applications but it won't make the enterprise applications any smarter. EDM, in contrast, is an effective way to inject more "intelligence" into those enterprise applications and so turn that data into better operations.
However hot BI is as a market, and it's pretty hot, it is not going to solve the operational issues that companies have. EDM builds on BI to take it to the next level and that's where companies need to be going.
Here are the links that prompted the post.
If you want more on BI v EDM there are many posts on this blog and my others (ebizQ and Smart (Enough) Systems). So many, in fact, I am not going to list them all!
Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the bookor visit the companion wiki.
Posted by James Taylor at December 7, 2007 2:06 PM
