I had an extended discussion yesterday about data analytics for websites and retailers. The subject was the idea that companies are collecting data about their customers, then selling/trading it and using it for things like targeted advertisements and shopping experiences tailored to a customer’s personal profile.

This is something that’s a hot button for quite a few people. The word “creepy” is thrown around a lot, as is “invasion of privacy”, which I think is understandable but I also feel can be a somewhat hypocritical concern. My immediate thought was that people have restaurants and shops where they’re a regular– where the staff knows them by name, knows their favorite meal/size/tastes, and gives them an extremely personal, tailored experience. This is considered GOOD CUSTOMER SERVICE, and it’s the kind of thing that gets a place repeat business and good reviews. The process by which this happens isn’t far removed from, say, Target collecting user data from customers who use credit cards to pay. Frequently, those establishments have a profile at a terminal in the back that they’re pulling up so they can ask you how you like your new shoes from last time, or if you want “the usual”.

It’s interesting to me, because it begs the question of what’s different about the restaurant where you’re a regular and Target, that you shop at a bunch. Why is it good customer service when your server knows your name but creepy and privacy-invading when it’s Target? When your favorite bartender moves to a new bar and you start frequenting that one, your information is shared with that new bar; how is that different from your profile being shared? There’s no opt-in there, you didn’t choose to have your information shared, but you’d (likely) be happy to get the same kind of personalized treatment at the new bar as you enjoyed at the old one, but many people are much less cheerful about seeing Facebook advertise their recent Amazon browsing.

My take is that the difference is tact. If your favorite restaurant were to loudly announce you as you walked in (“our FAVORED CUSTOMER!”), and make indelicate personal comments like “we see you’ve ordered the double-fried monte cristo with extra bacon and jelly the last nine times you’ve been here, would you like the number of a cardiologist?”, you’d probably stop going. Not because you object to the data being collected necessarily, but because it’s being used ineffectually and unsubtly. It’s too transparent that if you browse something on Amazon and don’t buy it, you see the exact same thing on Facebook later, heedless of whatever reason you may have had for not purchasing it.

I think the real problem is that the use of data analytics isn’t good enough. It’s a blunt instrument, lacking both the finesse and the tact of what we consider the more “personal touch”. It doesn’t help us to see the same Amazon listing that we already browsed, but it might to show us similar, cheaper products, or ones with higher ratings, or ones with more features. It’s not helpful for Target to hit you with advertisements for size XXXXL briefs, but it would be helpful for it to send you ads for laundry detergent or toilet at about the time you usually run out.

Data-driven marketing frustrates us because it doesn’t feel like an effective use of our data– it’s just that this is a few steps removed from an individually bothersome incident. It’s not that data analytics are necessarily creepy, it’s that when they’re used badly creepy situations can occur. Like your favorite diner recommending you a cardiologist, whether or not they’re accurate in their assessment it’s them using the information they have about you poorly.

A comment I’ve heard is about “control” and “opting in”. Putting aside that we already cheerfully hand massive amounts of data to social media platforms, it’s not terribly difficult to keep a big retailer from collecting data on you– simply don’t give out your e-mail address and pay for your purchases in cash. I suspect, however, that the convenience of a card is too overwhelming for a majority of people. Here’s the thing with a card– paying for something with a card is trading on your reputation to make purchases. It’s essentially a voucher that says “such and such financial institution thinks that I am a reliable enough person that I will pay my debt to you for these things I am taking from your store”. You can’t trade on your reputation, though, without some tie to yourself, without giving up some amount of personal data. Companies have used that personal data — data provided through lines of credit — for their own ends for centuries, this is nothing new. There’s a cost to the convenience of using a card to pay for things, and it’s not generally in the price you pay for goods.

A lot of this is also that I feel like Pandora’s Box is already open– even if we overwhelmingly decide that data analytics-driven marketing is creepy and unacceptable, it’s not like corporations worldwide are going to suddenly go “oh, yeah, I guess we shouldn’t do that”; they’ll just get better at being subtle. This is already happening. For one company, analyzing customer clickthrough data revealed that customers were less likely to look at one of a series of advertisements if all of them were accurate to their customer profile than if one or two of them were blatantly miscategorized. Having three or four home decor and sleepwear items and then something wildly incongruous like a motorcycle repair kit got more hits than if the repair kit were absent. It makes the marketing feel less personal, like it just HAPPENED to stumble on some things you liked.

I’d much rather see this sort of advertising become more subtle and tactful rather than simply try to fool people into thinking it’s not as good as it actually is.