Here are a few key use cases.
An apparel retailer has spent years investing in paid search, but only recently began investing in social media advertising. The cost per acquisition in this new channel is higher than for paid search - and if they relied just on their historical customer lifetime value (CLV), they would pull out of the channel altogether for not delivering a strong return. But using predictive analytics, they are able to determine the lifetime value of the new “social” customers within just a couple of weeks, and they may realize that these new customers were much more valuable than their paid search shoppers. Ultimately, this convinces them that social media is a profitable new acquisition channel for them despite the higher upfront cost.
It's January, and a retailer wants to forecast the revenue it will realize from all the new shoppers it acquired over the most recent holiday season. An approach based on historical analytics would look at the spend of holiday shoppers from last year. But the company’s marketing tactics and promotions are constantly changing from one year to the next. Is last year really a reliable benchmark? Instead, using predictive analytics, the retailer is able to extrapolate from just a few short weeks of observation of these new shoppers, and leverage everything else it knows about them – what channels they were acquired from, their demographic makeup, and what categories they started off buying – to arrive at a much more accurate forecast of their spend.
A retailer is looking to invest in a lifecycle marketing campaign to deepen its relationship with its customers. What products should it recommend to a given shopper? A historical approach would dictate recommending products that a customer has already purchased: After all, somebody who bought shoes in the past is probably receptive to an email featuring shoes. A predictive approach, in contrast, would attempt to identify what product to recommend based on what a shopper is likely to buy next. A retailer using this approach might recommend a belt or jewelry to somebody who previously bought shoes.