Subscribe to DSC Newsletter

What is the additional information the predictive model brings?

I was challenged by a Senior management of the company with such a question and I was speechless.

I built models based on the data. Without much knowledge about the key variables, the model generated a very interesting result which are consistent with the knowledge of the Senior Managers. I've expressed that Predictive Modeling for marketing activities can bring in more clients, improve cross-selling, revenues per clients, loyalty of the clients, blah blah blah all those textbook benefits etc. Then the management said: "Okay, I could have a bankers/marketers to do all these and I know all of these knowledge. I do not need a model to do this."

I supposed the model is a mimic of human judgement, not something that outrun an expert and professional, right? What should be the best answer?

Views: 522


You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

Comment by April Smith on March 11, 2013 at 9:56am

An individual human being (banker, marketer, etc.) will always be imperfect at analysis. We'll never have personal experience with all the variables that affect the system we're trying to predict. We're influenced by a variety of cognitive errors such as the recency effect, where a conversation we had on the train that morning suddenly gets coded in our brains as the 'truth' about the relationship between this marketing channel and that consumer behavior. That is why the subject matter expert can always provide hypotheses on which variables should be considered for inclusion in the model, and should certainly be leveraged in fine-tuning the model if unusual relationships emerge, but an individual person/brain (or nearly just as bad, a team of folks who are subject to group think and other group-level cognitive biases) is not the right tool to produce reliable, predictive outcomes for a large number of cases like a predictive model will. It is too easy to influence us with saliency and emotionality, leading to ad-hoc opinions based primarily on individual experience that we are prepared to defend with vehement conviction (as your senior executive demonstrates).  

On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service