Syndicated from b-eye network
Eric Siegel is blogging in the build up to Predictive Analytics World (at which I am speaking and teaching a workshop). He posted Predictive analytics FAQ #1: Prerequisites for success.I liked his list but personally I would have added another:
Know the decision(s) you want to influence
Unless you have a decision or decisions in mind when you start it is all to easy to generate insight without any understanding of how you will make it stick – how it will make a difference. Starting with the decision in mind helps focus the analytic effort on improving that decision. If it is a strategic decision, it focuses the project on helping an executive. If it is an operational one, it means thinking about how the results could be deployed into production.
I would also add something to his point about “General understanding and buy-in of a predictive analytics initiative by stakeholders across business functions”. For many predictive analytics projects this needs to include the IT department and the folks working on BI/DW projects and all too often predictive analytics projects include neither.
Comments on this entry are closed.
Great post, James. There still seems to be a lot of confusion out there about how the most value can be created from predictive analytics and BI in general. You have a talent for making the important issues clear. To me, “how insight will make a difference” is equivalent to “how predictive analytics will increase shareholder value.”
The only way to find out is to begin with the decisions I have (whether current or contingent on the future) since the only way I can hope to positively affect shareholder value is through my decisions. Once I’ve identified my decisions, I can prepare and gather all relevant information about the very uncertain future, whether that information originates from data, predictive analytics, or an employees’ knowledge. Then, I’ll use that information to help me quantify the shareholder value of my decisions, so I can choose wisely. Only by explicitly linking information to the shareholder value of decisions, can I discover the true value of my predictive analytics.