Are you ready for decision management? Part 3

How do you tell if your organization is ready for decision management? If not, what can you do to start laying the groundwork?

Third, Analytic understanding

A third area of readiness is going to be in terms of analytic understanding or sophistication. Delivering decision management solutions, especially the more advanced ones, requires analytic skills. There are a number of areas in which analytic readiness may be an issue.

Some companies will lack the necessary analytic skills completely. Where there is not a prior culture of analyzing data to try and understand the company and no history of using mathematical models there is unlikely to be any existing analytical staff. While one can be built it will probably be easiest to start by hiring external consultants to provide these skills. Fortunately many companies offer services to turn data into analytic models and, provided sensible due diligence is performed and the request is framed in terms of getting the analytic insight into a decision service not just developing a report explaining the model, all should be well. Over time, as analytic models prove their value, it may well be worth hiring in-house analytic skills. If you lack this kind of analytic skill in house then the first few projects should be those where rules (regulations, policies, expert judgment) dominate rather than data-driven insight.

Some companies will have plenty of analytic skills but will find them focused purely on reporting and analysis - Business Intelligence-style analytics. While these skills might be useful in an decision management context, a change in focus and an upgrading of skills will be required. Someone who can use a regression analysis tool to understand how data is trending has many of the skills necessary to build a predictive model that uses regression analysis but they need to learn how to apply them differently. If you have a BI Competency Center or other concentration of analysts you may find suitable people there to work with you on projects but you may not. Either way, make sure you engage these folks early in a project so that they can see how upgrading their skills and changing their focus will help demonstrate a better ROI from the BI/DW investments already made.

The last option is that there is a center of modeling excellence in the company but it is very narrowly focused. For instance, on risk modeling or in an actuarial group. While these resources will be ideal for projects dealing with their specialty it may prove difficult to get access to them for projects in other areas. Even if they can participate it may be hard for them to apply these skills to operational systems and to new domains. On the plus side you should be able to use them as a proof point for the overall value of modeling and analytics and that should make it easier to hire/contract with the necessary resources.

In general, unless you have an existing group of modelers developing models for your domain area, you are likely to be looking for outside help initially. If the analysts you use, inside or out, are not used to the decision management approach and expect to be able to deliver a report describing a model rather than work with you to deploy that model into a decision service then you will have a change management task ahead of you. Fortunately the data mining/analytic modeling community is increasingly aware that a model is no good unless it gets deployed.

Some useful links for you:

Technorati Tags: analytics, BI, business intelligence, data mining, decision automation, EDM, Enterprise Decision Management, innovation, operational BI, predictive analytics