We’ve recently been doing a lot of “Uplift modeling” for direct marketing, which gives an interesting perspective on this question: we find that you should typically be marketing much less to the Best Customers (who’ll buy anyway), but instead focus on the “Persuadables” on the margins where your marketing can tip the balance.

Historically marketers tried to predict which customers were most likely to buy after receiving a direct mail piece (or email or phone call). Unsurprisingly, the Best Customers tend to score very highly in such models, and thus are heavily targeted in subsequent marketing. But this may not be effective in getting them to buy more (how many times a month can you eat at Burger King?!), and can even end up over-saturating and driving them away.

More recently, direct marketers are warming to the idea of Uplift modeling, which instead tries to predict how much the marketing effort *increases* (or decreases!) each customer’s likelihood to buy, as compared to no marketing. Best Customers often don’t score nearly as high: because they’re going to buy anyway, we don’t need to invest in extra incentives for them (in fact we often label them as Sure Things). It’s much better to go after customers on the margins (who we call Persuadables) where your targeting can have a real impact. You also want to avoid the “Lost Causes” (who will never buy) and especially the “Sleeping Dogs” who would have bought until your marketing pushed them away (by annoying them, or giving them an opportunity to make a considered decision and shop around).

More at our website.