A few weeks ago, one of our associates received a direct mail piece at her parent’s address. It read: “Congratulations! We heard you said Yes!”
Her parents were curious – was there something their daughter wasn’t telling them about her love life? Was their little girl getting married? No. It was a classic marketing blunder that created discomfort and dissatisfaction.
How did this very creepy mistake occur?
A large chain store has high-powered analysts mining purchase history data for clues and cues on life events like who is pregnant, who has moved, or who is getting married. Why? Because life events often lead to spending on goods and services. The company that can accurately predict a life event can beat its competition and capture market share.
Theoretically, their analysts did all the right things:
- They identified a life event: People who get married spend a lot of money to get their households set up. Their friends spend, too.
- They could connect an existing service to that life event: a bridal registry.
- They could determine the key combinations of purchases most people make right before they sign up for a bridal registry and deduce that those combinations are good predictors of an engagement.
- They built a predictive model and handed the marketing department a list of people who were likely to have gotten engaged.
The analysts probably warned the marketers that a great model and good predictors only get you close. These people are only more likely to be engaged and probability isn’t a certainty.
The Marketing Department got excited — they were so happy to have a list of people to target that they ignored the “likely” part of the message from their data analysts. Instead, they put together a mail piece that assumed 100 percent of the people on the list were recently engaged to be married and that 100 percent were likely to sign up for a registry.
They were wrong. Even worse, they failed to consider how their message might be received. Clearly, if they’re wrong, the misunderstanding could damage the customer relationship. If they are right, they cross the “Big Brother” boundary – how does this store know we got engaged? A tiny tweak of their ad copy could prevent people from feeling as if their privacy has been invaded.
So, what’s the answer? We could stop companies from using their data for marketing purposes, but it means that the 99 percent of us who are not getting married are going to get a piece of junk mail advertising for the bridal registry. Using data to give people timely offers for discounts on products they want is a good thing. The mistake here is in the message, not the method.