Why are most “targeted” marketing offers so bad?

Offers, which are ostensibly based on data and analytics, are almost universally bad. Companies should change their approaches to offers, or stop bothering with them altogether.

If you keep track of the marketing offers you get and you are bored someday at work or home, do a count of what percentage of the offers actually appeal to you. For me it’s about 0.1 percent—I find one in every thousand or so offers to be something I might actually want to buy. Once I received a discount offer for a restaurant I really love, and the event was so rare I teared up a bit.

Perhaps your offer hit rate is higher, but I doubt it’s very much so. Offers, which are ostensibly based on data and analytics, are almost universally bad. They waste a massive amount of resources, including our time and attention. They also waste the marketing energies of the organizations that issue them. Companies should change their approaches to offers, or stop bothering with them altogether.

Just to illustrate the problem, here is a sampling of emailed offers that current sit in my inbox (fortunately in the “Promotions” category, where I can usually safely ignore them):

  • A promotion for a retailer offering me women’s leather accessories (no, I am not into those).
  • A “tickets go up $200 on Monday” pitch from a conference that I never attended, in a city far away from where I live.
  • A “flash sale” for a hotel chain where I have only ever stayed on business (and where I am therefore unlikely to take advantage of a “flash sale”).
  • A request for funds from a well-known politician whom I once met in an elevator—so I doubt she really remembers me—for a political campaign I don’t support.
  • A concert announcement for Peter Frampton, a singer whose music I last purchased in 1976.
  • A “teen acting workshop” from a university where I do have a minor affiliation; alas, however, I have no teens.

There is quite a bit of variation in the approaches and analytics companies use for targeted marketing. However, I would argue that targeted offers fall into five categories: retargeted offers, well-meaning but poorly-targeted offers, offers that benefit the offerer rather than the potential consumer, offers that are OK except for the context, and well-targeted offers that benefit you. Notice that only the fifth of these has much chance of succeeding. Let’s discuss each type in order to better understand how to improve the overall genre.

Retargeted offers: These are particularly popular today. Retargeted offers are online pitches for things you have already looked at or searched for on the web—ideally recently. This is perhaps the least creative form of offer. It does succeed sometimes—perhaps the doorbell rang when you were attempting to buy something, and you didn’t come back to the purchase transaction. But for the most part, if we abandon a search or purchase, we intended to do so. It’s annoying to be continually reminded of what we decided not to buy.

Well-meaning but poorly-targeted offers: Many companies create offers that are not objectionable except for the fact that there is no reason to believe that the recipient actually wants the goods or services on offer. Perhaps they don’t know, and have not bothered to find out, the potential customer’s age, gender, residential location, likely income, and online or offline shopping habits. All of this information is readily available if you look hard enough, and some customers will actually furnish it themselves. So the existence of poorly targeted offers is evidence of not trying hard enough. Not much ventured, even less gained.

Offers that benefit the offerer rather than the potential customer: This is one of the greatest problems in targeted offers. Companies somehow believe that, because they have excess inventory or because they have money from a manufacturer to promote a product, you will be interested in it. As the quantitative analyst Kaiser Fung notes in a blog post, most companies view offers as a way to get you to buy things that you don’t want. They would drastically increase their success rate with offers if they tried to interest you in items where you have already evinced an interest. At a major drugstore chain whose loyalty program produces large numbers of targeted offers, the managers of the program say they have found that people seldom go for offers for things they haven’t bought recently. Even if they have recently bought toothpaste, they don’t redeem offers for discounts on toothbrushes.

Offers that are OK except for the context: The context is critical to whether customers will respond to offers or not. If you’re sending out mobile offers, of course, you want to make sure that the shopper is near your store. That can be difficult to accomplish technically, however. More straightforward contextual issues arise when the customer contacts you. If the customer calls your call center to complain about one of your products or services, should you make an upsell or cross-sell offer during the call? Of course not. Another company I once worked with, a large bank, was presenting what it considered attractive mortgage offers on ATM screens. Who has time to consider offers while standing at an ATM—and if you did, what would others behind you in line think about it? Make sure your offer is contextually sound.

Well-targeted offers that benefit you: These represent the gold standard in offers. They are created to reward you for your loyalty as a customer. They are sincerely intended to save you time, attention, or money, or to give you a valuable opportunity to buy a scarce good. I have found that the only companies that have much luck with their offers are those that truly value their best customer relationships and know that they have to show appreciation with valuable offers. Analytics at these firms are used to determine what you really want, not to take a shot at offering you something you probably don’t want. A large UK retailer and a large US-based gaming firm have been fantastically successful with this approach.

Offers are generally bad because companies generally don’t follow the philosophies behind the fifth offer approach—well-targeted offers that benefit you. This seems obvious, but it apparently isn’t. No amount of analytics will make customers want something that doesn’t suit their needs. No CRM technology will do that either. We want what we want, and the job of targeted marketing analytics is to figure out what that is and make it available at an appealing price to our best customers.