Skip to main content
"Blogging Live" from InterACT Lisbon: "How EDM Really Works"

(Posted by Guest Blogger, and Corresponding Secretary of the James Taylor Fan Club, Ian Turvill.)

I attended James Taylor's session "How EDM (Enterprise Decision Management) Really Works "yesterday."  First, I should state that it was exceedingly well attended, with probably more than 100 people in the room - which is a very positive sign, given the total number of people overall at the conference.

There were a number of key points he made, and which I thought were worth repeating here.  In addition to my commentary below, I've also included most salient slides from James's presentation.  I'm not going to reproduce the whole deck here - you have to pay and attend if you want to see the whole thing!  ;-)  James will be repeating the presentation at InterACT San Francisco, so you can catch the whole thing there.

One of the most important elements in James's presentation was the concept of "Hidden Decisions".  This will be one of the central themes in James's planned new book "Smart (Enough) Decisions", and it encompasses the idea that many companies are - for a variety of reasons - simply unaware of the complete range of decisions they make.

For example, decisions may be hidden, because they are not automated, but instead made on an almost unconscious basis by employees at the front line of interactions with clients: think about tellers at banks or ticket agents at airport gates.  (See Slide 1 to see further examples of "Hidden Decisions".)

A further key point in James's presentation was why Enterprise Decision Management is becoming increasingly necessary for companies wanting to maintain their competitive edge:

First, the range of available analytical capabilities is evolving and maturing rapidly.  (See Slide 2.)  Where companies once relied on straightforward descriptive and predictive analytics, there are now evolving to use of "decision optimization", where the profit maximizing action for a specific set of circumstances can be readily prescribed.

Second, there's no point developing models if you can't actually deploy them within systems quickly: the perfect model is of no use if you can't actually deploy it as part of your operations.  The range of technologies that companies rely upon to put analytics into action is growing, but the problems associated with each are growing.  (See Slide 3.)  James would - obviously, but correctly - state that EDM can be the "salve" for these problems.

Third, many managers might believe that, if an IT system has to change, then there must have been something wrong when it was created.  The problem is that most decisions - and therefore the systems that support them - should not be like that: they should change as a matter of course, because the decision making approach should evolve to reflect the changing circumstances of the market.

EVEN IF the environment were static (an unlikely situation), you would STILL want your decision making approach to change to see if you can improve outcomes.  There are two standard approaches companies use to adapt their decision-making strategies over time (see slide 4): Champion/Challenger testing, which pits new tactics against existing operational methods, and Learning Strategies (sometimes known as Experimental Design) which tests many different alternative approaches in a systematic way.

Of course, Enterprise Decision Management provides a high level of automation and support to help users to evolve to the best possible decision making strategy.

related posts