Skip to main content
Why do you need Adaptive Control?

  • Why do you need Adaptive Control? - This Post
  • What's a basic technology approach for Adaptive Control?
  • What is Champion/Challenger?
  • What is Experimental Design?
  • How do I do Decision Analysis?
  • The first issue to consider with Adaptive Control is why to invest in the necessary infrastructure to improve your decisions over time. Adaptive Control requires software assets and staffing to be applied not to building new Decision Services, but to improving old ones. At first sight this goes against the whole "EDM reduces maintenance" pitch (that I have made here, for instance). It does not. You should not consider improvement of your decisions over time "maintenance". Adaptive Control is about continuously improving the way you make decisions. Some of these changes will come from changing business conditions that force a change in the approach being taken to the decision. Mostly, however, it is a case of making a decision better and better over time to boost profits, reduce losses, or improve retention.

    After all, you constantly learn more about your customers and gather more information about their behavior. New insights and market trends come from you, your competitors and from third parties. A process for continual review and improvement of how you take a decision allows you to detect and respond to changes in the behavior of your customers without having to start a special project and helps you show an ROI for the data you collect and analyze. This is complicated by an interesting fact about business decisions - at the point of decision it is not known what the long-term outcome of that decision will be. Thus not only does the best decision change over time but you may not know what the best decision will be when you must make the decision. For instance, the graph below shows how various actions taken today influence your profitability over time in different ways. One causes a steady growth in profitability, one has good profitability but only after some time and the third has a rapid growth in profit followed by a falling off.

    Profitovertimecurve_2

    This illustrates two truisms about decisions. Firstly if you use a single approach for every decision then you will only plot one of these curves and will have no data about how other actions might have resulted in better (or worse) results. Secondly it shows how important it is to track your results and know what kind of impact you are looking for - short term or long term? Low risk or high risk? Only by building an Adaptive Control environment into your Decision Services can you create an environment where you can manage and improve these decisions over time.

    Tomorrow, an overview of the technology infrastructure for this.

    , , , , ,

    related posts