by Ryan Harrington

A few weeks ago we discussed different types of questions that could be answered through predictive analytics. We included examples that covered everything from determining which customers might churn to figuring out what profits might look like in the future to what products might be bought together. For any organization, there's almost a never-ending number of questions that predictive analytics can answer. 

To get started asking those types of questions, it's helpful to understand the broad buckets of problems that predictive analytics can solve. These broad buckets are:

  • Clustering
  • Prediction
  • Association

Each of the questions that we looked at previously fits neatly into one of these buckets. Here's how that looks:


  • What customer segment does a new customer fall into?


  • Which leads are most likely to become paying customers?
  • Which customers are likely to churn?
  • Which marketing channels are providing the most lift?
  • What will our profits look like over the next two years?
  • How does sentiment about my product effect revenue?


  • Which products are customers likely to buy together?

Understanding these broad buckets is the key to using predictive analytics. Over the next several weeks, we'll cover each of these buckets in order to give you and your team a better understanding of how to use data to your advantage. Each week we'll discuss

  • Big-picture definitions
  • Common techniques
  • Business case studies

With this information in hand, your organization will be one step closer to using data to make more effective decisions.


AuthorRyan Harrington