Skip to main content
Choosing the Right Analytics: Time-to-Event Models

We continue our series on choosing the right analytics for the job with an overview of time-to-event models. Knowing how a customer is likely to behave is useful, but knowing when a customer is likely to behave that way is even more powerful. A time-to-event model predicts a window of opportunity when the customer is most likely to act.

Many businesses make decisions on what to promote to whom anywhere from a few days to a few weeks out. The propensity of a customer to buy a specific product, however, varies over time. By timing campaigns and other promotions to these windows, businesses can improve relevancy for their customers, driving higher response rates.

Similarly, the likelihood that a customer will end membership in a loyalty club or premium service varies over time. Some customers fail to renew, and others just stop using the service. Businesses want to prevent this from happening, but if they act too soon to offer retention incentives, they may be spending unnecessarily. Knowing when attrition is most likely to occur enables the retailer to time actions for the highest likelihood of success at the lowest cost.

Here’s how some of our clients have used time-to-event models:

  • One client used this type of analytics to increase the return on investment from promotional mailings. When a popular new DVD or videogame came on the market, for example, the client sent offers only to those likely to buy the product within the offer redemption period. Response rates were two-to-three-times higher than when the same offer was sent to everyone.
  • Another client improved its ability to predict when customers were about to make a big purchase by incorporating customer clickstream data into time-to-event models. Many consumers do extensive online research before making an online purchase or walking into a store to inspect the item. By analyzing clickstream data from its site, along with customer purchasing histories and historical behavior patterns, the client could pinpoint the right moment to make an offer.

Next time, our series will continue with an overview of uplift models. If you can’t wait for the next post on this topic, check out our Insights white paper: "Which Retail Analytics Do You Need?" (registration required).

related posts