When to Use Predictive Modeling

Casey BartoIn our previous predictive modeling post, we explored how it works and what the difference is between predictive modeling and analytics.
Now that you’re enlightened on how predictive modeling works, here are some more tips from our resident data pros, Jeff and Sophia on when it’s best to use it. If you have any data questions you’d like to see answered, please leave them below!

What are Some Scenarios Where I Would use Predictive Modeling?

Jeff: “Generally speaking predictive modeling scales really well. What that means is if you build a model and apply that against larger universes that’s when it really makes an effective ROI point. Look at it this way, if you have 300 emails building a predictive model to figure out who’s most likely to open a certain email is not a good use of your time or your money, but if you have big data sets where you’re trying to refine a specific goal or you’re trying to cross reference against a different goal that’s where predictive modeling can help you out a lot because you’re trying to redo your segmentation strategy where you have this big audience of folks. With predictive modeling, you can really refine down your segmentation. Let’s say you have product line A and you’re launching a new product, it can help you find potential targets for that new product line.”

Sophia: “Another common application is for prospecting. Say a retail client has 3 million customers, but they want to reach out to other potential customers in the United States. There are 200 million households out there. How do they select the most likely household that’s going to become their customers? It’s basically selecting maybe 1 million out of those 200 million. Predictive modeling can help identify the top 1% out of the whole population. With traditional ways I don’t think it’s easy to do that.”

One Trackback/Pingback

  1. […] « When to Use Predictive Modeling […]

%d bloggers like this: