Yesterday my old buddy Dave Wright told me about Bill James on 60 minutes - for those of you who don’t know, Bill James is the Red Sox stats guy who, like Billy Beane at the Oakland As, uses data mining and analytics to drive recruiting, game planning etc. Dave’s comment was “now that’s Decision Management” but, as you can tell from the headline, I am going to disagree with him.

Don’t get me wrong - I am very impressed by the changes Billy Beane and Bill James have made. It takes a lot of skill both to do the math they have done and get traditional organizations like baseball teams to adopt these approaches. The use of data and data analytics/data mining to understand players and games is great and clearly a very powerful tool. But, like other examples of data mining in healthcare (to see which techniques have better success rates) or similar, they are about creating individual nuggets of insight. In a business context, this is like using Business Intelligence (BI) or Performance Management tools, visualization software or even data mining tools to decide where to put the next distribution center or when to add a new store in a particular location. It’s powerful and potentially very profitable, it’s not decision management.

Decision management, enterprise decision management or business decision management, is about the automation and improvement of operational decision making. It is about a focus on the micro-decisions within operational systems, day to day transactions. Each such decision has a small value, but the volume is such that it really makes a difference if you can make them all more precisely, more consistently, faster and cheaper. It also makes a difference if you can change and adapt them quickly - if they are agile. To apply decision management to baseball you would have to have pitchers, catchers, fielders and batters whose decision making could be controlled and analytically improved. We may one day get to robot teams, but so far this is still an area of human judgment and skill. If you think about it, the folks delivering these operational decisions (baseball players) are highly paid and very skillful. Like most highly paid, skillful employees they need tools to help them analyze their performance, ways to track improvement and warnings when things look like they may start to slide - they need BI and performance management, in other words.

But what if it was a system that was making each pitching, hitting or fielding decision? What if they people “delivering” these decisions were not skillful and highly paid? Well, then you would need to manage these decisions in software - you would apply business rules to automate the basic decisions, apply predictive analytics to leverage data in these rules, use adaptive control and optimization to constantly improve them and so on. You would, in other words, apply decision management.

So, if these wonderful stories about data mining and analytics in sports are not decision management, can we learn anything from them? Well certainly. For instance:

• Using data and analytics to change behavior is a massive organizational change. Baseball teams may be very traditional and superstitious, but changing to an analytic culture is really difficult in business also.
• The illustration of needing to measure “on base percentage” is a good one as you might need to count or measure things differently to successfully apply decision management. Like win-loss records for pitchers, some of your traditional measures and KPIs may not, in fact, work.
• It is important to remember that observations of results can fail to differentiate between luck and skill. Two customer service representatives may have very different calls on a given day and that may explain the difference in results, not their ability.
• As soon as you count something it can become a fetish. Just because you are counting or measuring something for a KPI, does not mean you should. Going back to first principles and proving connections and testing assumptions is critical.
• Change is constant so this kind of analytical decision making, like decision management, is not static but must be constantly tested, refined, used and assessed. It’s not a one time thing, but a change to a whole new way of working.

You can learn from these stories but, sorry Dave, they are not what we mean by decision management.

If you are interested, I blogged about Michael Lewis, the author of MoneyBall, when he spoke at a Fair Isaac conference.