I am attending Gartner Symposium and ITxpo this week and blogging as I go.
I have been looking forward to this session - I know Kurt and Jim well and Donald and Ted somewhat. Each bring a different perspective - Kurt BI, Jim BPM and Donald Information Management. Ted started by putting up a Venn diagram showing overlapping between various elements in this space. These diagrams are always limited as they fail to show certain specific overlaps. In this case although BI Platforms includes data mining but how to link this to decisions in processes was not shown. Rules Engines also do not impact on Event Tools which seems bizarre also (see this post on EDM and BAM and this one on SOA, CEP, BPM).
Anyway, opening comments:
- Kurt started off by talking about BI. He made the excellent point that most BI groups are focused on delivering information to people. They need to think about how BI impacts business processes that drive the business. This is exactly how I feel.
- Jim talked about process management. Processes are key, he said, but there are lots of supporting players required. The focus on creating services for IT flexibility is a key trend and then there are trends to externalize and manage processes and rules. Business trends of increasing speed and responsiveness. Also the business people are starting to manage more of the flows and rules for reduced costs and greater agility (I wrote about the dirty secret of business user rule maintenance but if done right it can be very powerful).
- Donald pointed out that although data and databases and data warehouses are central to all this, they too cannot stand alone. If no-one uses it, what was the point?
In the end all three agree that the business, and the entities outside the business, drive all this. The business MUST drive IT spending.
In the end I think the three of them would find it easier to describe the overlaps and integrations if decisioning was a category on their chart. It ties together information, analytics and BI and delivers it to BAM, event tools and BPM. Having an EDM focus and a decisioning platform would tie all this together much more effectively. I think my previous posts on transaction-centric decisioning as well as this article on bringing BI into BP and this post on ways of bringing BI into BP .
BTW Fair Isaac is at booth 305 and I am presenting at 12:15 on Wednesday on Real-World Experience in Applying Business Intelligence to Business Processes
Some random thoughts based on questions answered:
- BI is like driving through the rear view mirror
- BPM (and rules) could be thought of driving while looking forward.
- The optimization tools for BPM are very like those used in BI but applied with a different perspectives.
- "Managing business processes with math" - become more accessible and usable. Enterprise Decision Management is focused on making this work for real processes and there's lots more on the blog.
- Predictive analytics are built on historical data - so it better be right and available. But if you apply the right expertise to the company you can turn this historical data into prediction.
- Kurt made two good points - most companies need to add expertise to deliver predictive analytics but once it is done, the results can be fed into BI. Thus a predictive scorecard might require new expertise and software to develop but the resulting score - predicting risk or churn say - can become part of the reporting and analysis infrastructure for knowledge workers.
- The huge volume of data coming down the street will require new technologies (like RFID) and approaches to turn them into something useful. Events are likewise a high volume problem. The probability of intelligent separation of feeds into need-now and need-later may also be important.
- Kurt things that pervasive BI will come to pass in particular through performance management and that this will increasingly involve companies sharing information and metrics.
- Jim made the point that web 2.0 might increase the delivery of mass personalization as well as collaboration. Interesting.
- The merging of structured and unstructured data came up too. Ted pointed out that there is relational, unstructured as XML documents, and real unstructured data like video or email. People are starting to do this - the database vendors are adding support for text and content managers are starting to build on databases. Definitely coming. Of course, from my point of view, if you can't act on all this data then why capture it? Is the most useful thing to turn unstructured data into structured data? Could be, certainly something Fair Isaac does. If you want to drive the process with unstructured content then you are certainly going to want to do that.
- One question came up as to how to manage the replacement of retiring expert staff and the possibility of using business rules to do this was widely discussed. Definitely agree
- Metadata repository for everything seems to be slow in coming. I agree - for now I think you have to manage multiple repositories and do the metadata integration they need.This is why business rules products, like Blaze Advisor, have both good repositories and extensible metadata. The BI vendors, and to some extent BPM vendors, are lagging here.
Wow - interesting questions.
Technorati Tags: agile, analytics, BI, BPM, BPMS, BRE, BRMS, business process management, business rule, business rules, composite application, data mining, decision automation, EDM, Enterprise Decision Management, Gartner, operational BI, predictive analytics