DM as in Decision Modeling and Decision Mining: looking below the surface
Jean Pommier 270001XBPR firstname.lastname@example.org | | Tags:  decision-modeling brms decision-management decision-mining business-decisions modeling
0 Comments | 2,799 Visits
I was going to blog this week about “bad decision management” or, more accurately, the management of bad decisions vs. the bad management of decisions. But before I do that in an upcoming post, we need to talk about decision inventory or decision mining because it’s unlikely we’re even aware of our bad decisions without a clear and comprehensive list of the decisions we make as individuals, teams, departments or organizations already.
We certainly know about some of our decisions, especially the ones we’ve defined and often instrumented through an IT initiative – the ones that are helping to efficiently drive core business processes, such as claim processing, loan underwriting, credit scoring, order verification, or network design. But what about the decisions that we’ve barely identified? The decisions operating below the surface of our current decision management practices, micro-decisions that often number in the thousands?
Well, we can’t manage what we don’t know. Like the M stands for either Management or Modeling in BPM, Decision Management requires a strong Modeling component to be effective. We also talk about Decision Analysis in the industry. However, most of the current decision modeling or analysis techniques and frameworks assume that decisions have been identified. Again, this is certainly the case for some obvious and major decisions. Furthermore, many of these approaches are specific to a particular implementation or instrumentation approach (e.g., business rule management, predictive analytics), leaving aside decisions which are not necessarily relevant to such technologies. In either case, we’re missing critical information about our businesses.
Whereas modeling for execution is a good thing in IT, it’s not so much for Business because we’re missing the bigger picture of decisions that are not yet mature for instrumentation, or ones that cross several paradigms such as analytics, business rules or events, business process management, collaboration. For this reason, I’m advocating that we pay more attention to upstream phases in Decision Management, such as decision inventory or mining. To do so, we need to establish the right balance between some formalism and some informality. Formalism in order to ensure consistency in the knowledge we capture and linkage to potential implementation. Informality to capture more knowledge than what we do today in IT-driven modeling exercise and get more people from across the organization to contribute to the mining exercise by bringing their own decision nuggets. Yes, there are many decisions to discover out there, good and bad. It’s the new gold rush!
During a road show across Europe two weeks ago, I met a satisfied client who had just completed their first decision management implementation, a combination of business rule management and BPM, but was having difficulty selecting the subsequent business cases without an inventory of their decisions to easily pick candidates for improvement and instrumentation. Conversely, I met another client whose IT department was having trouble coping with the 30 or so decision services that the business is begging them to instrument because the business had proactively done its job of mining their operations and decisions.
Join me at SemTech. I’m putting this very topic on the table for discussion at the upcoming 2011 Semantic Technology Conference this week in San Francisco (Decision Modeling: Why and How). I’m going to talk about why I believe decisions represent key semantic differentiators in the business world, and therefore should be properly mined and modeled as enterprise assets, and not just when IT is ready to instrument them. Hope to see some of you there next week. If you haven’t registered yet, and want to join the crowd, it’s not too late and you can get a 15% discount on the SemTech registration by using the code SPK15. And tweet #SEMTECH.