JT on EDM

David Greer had a cutely named post this week – The Engine That Can. David and I had a nice chat about eOptimize a few days ago and I thought I would respond to his post with some thoughts of my own. eOptimize’s product is unlike those often described as decision management applications – it is not a risk management or marketing application (which take up most of the space) nor is it a platform designed to build those kinds of applications. It also does not use predictive analytics, in the sense of executable models, but constraint-based optimization. That said, it does have some of the classic power of a decision management application and some of the challenges:

• It adds on to existing infrastructure, and processes, rather than requiring rip-and-replace.
• It is complex on the inside but relatively simple on the outside (the embedded intelligence is hidden behind a business-friendly exterior).
• While it uses the same core algorithms for every customer, the ability of those customers to specify their own rules and provide their own data means its behavior is customized in a very real way to each.
• It is designed with change in mind, being easy to alter and evolve.
• It does not attempt to handle 100% of decisions automatically but to handle a very large percentage automatically while supporting the people who must make the remainder.
• While it’s initial ROI often comes from cost savings (reduced headcount), its long term value comes from better scheduling decisions with less waste and less padding.
• It often “replaces” fairly expensive, knowledgeable and well-trained staff (because the decision it automates is a complex one) but these staff are typically redeployed to improve customer service or overall effectiveness of operations once they can be spared from the drudgery of ordinary decisions.
• It can take a while for the users of the application to trust its decisions, though they are typically much better.
• It enables much more what-if planning by empowering staff to quickly get decisions to a wide variety of possible scenarios.
• It can be hard to show the value of the application because it delivers all its results through other applications and use data provided by others – it is not a system of record.
• Once people “get it”, though, they REALLY get it.

There’s a growing number of these decision management packages out there – I like to think of these as the after-market turbo-chargers for traditional enterprise applications. eOptimize is, in that respect, a classic of the genre.