I almost called this post “Nobody Knows Anything” but that might have been overkill. I’ll say what I have to say and let you be the judge. Let’s say that you buy a friend’s newborn a gift. You have it shipped to your house. The data says, correctly, that you bought an infant gift. That might also lead to an inferred piece of data that places your household into the “presence of infant” bin, leading to you seeing lots of ads for diapers. If you’re the one placing the ads for those diapers, you’re wasting money.
Lots of the data marketers routinely use is of that sort. It’s inferred. You can see that some thinking at work if you’re a Netflix user: the recommendation engine infers what you might like based on your past viewing. Of course, if your kids or someone else in the house watch something in which you have no interest, the accuracy of those recommendations is diminished (which is part of why there are separate profiles available when you log in). Inaccurate data is, sadly, more the norm than an aberration. Since this data is really what’s behind personalization and targeting, that inaccuracy is a big problem. Any business that buys data from third parties – and an awful lot do so – may be putting garbage into their system. Unfortunately, most don’t know that because there is little transparency in the data business and it’s impossible to verify what’s good and what’s not.
What should you do? Invest in collecting your own, first-person data. You can also demand transparency in any other data you use (good luck with that) with respect to how it was gathered and what it really represents. Is it inferred or does it come directly from consumers (did someone tell you they had a baby in the house or did you guess they did because they bought one infant item?). Who owns the data and was it gathered with the consumer’s permission?
When Facebook tells its customers (marketers) that they have data on 41 million adults aged 18-49 in the US and there are only 31 million of those adults living in the US, you know much of the data is inferred and also that we have a problem. A recent study that found that 70% of marketers believe that the customer data their organizations are using for marketing is low quality or inconsistent. Why bother to market at all when you’re just flying blind?