The Failure Of Voyeur Marketing

Do you remember Facebook Beacon? When I think of “Voyeur Marketing” I think of Beacon. For whatever reason, Facebook thought people would be totally in love with random “partner” companies knowing everything they did on Facebook, but also that all of your friends would like to know you bought a dildo last night. A big one. Voyeur Marketing is the idea that being able to watch someone live their private lives gives corporations an ability to sell them things they don’t even know they need, or just whatever crap you want before some other corporation gets to them. In effect, Voyeur Marketing is an attempt to subvert the market by being the first one to offer a consumer some products so they don’t see your competitors’.

I predict that this kind of online advertising will fail horribly, and depending on who you talk to it’s already on the way out. It seems that marketers are finding out that ad effectiveness on voyeur platforms like Facebook and Twitter simply don’t work. The question though is why? To many people it seems intuitive that knowing how many children I have, their ages, whether I’m secretly gay, how big my porn collection is, who I’m having an affair with, and whether I have a drug problem are great ways of selling me diapers, leather suits, more porn, hotel discounts, and heroin. This is at least how the information is pitched to firms. “You will know the deepest darkest secrets of our users, and you can sell them anything.”

It’s almost a kind of first cousin, twice removed, blackmail scheme. “Heyyyyy, I saw that your daughter bought a pregnancy test. You should go to Target and buy some diapers.” It’s not direct blackmail, but it definitely hits all the key points of exploiting people’s private information to make money.

However, I believe that there’s a very specific quantifiable reason why this kind of marketing simply doesn’t work. It’ll be difficult to get any real numbers on this as the people in control of the data (Facebook, Google, Twitter, 23AndMe) have a vested interest in never letting any of their customers see real success information. Given that, I have a guess at why this kind of marketing doesn’t work which is summarized as:

The only statistically valid indicator of future purchases is past purchases.

I believe that knowing personal information does not predict future purchases because, even if they may eventually lead to a purchase, the distance between life event and purchase is separated by research or brand loyalty. This means that private life events end up being statistically invalid as predictors of future purchases because of confounding. The only consumer behavior that works as a predictor of what you will buy is what you bought in the past.

When I look at the emails Amazon sends me vs. the ads I see on Twitter this is fairly obvious. I buy art books, so Amazon shows me new art books, as well as new strings for my guitar, paint, and replacement vitamins. How can Amazon do this? Because I’ve bought art books, a guitar, strings, and paint from them before. Compare that to Twitter, where the ads I see in my tweet stream are completely ignored most of the time. Twitter is so desperate to make their ads effective they’re force following celebrities like William Shatner onto Mastercard. That’s a very good sign their ads are ineffective and they’re grabbing at straws.

Google is an odd case where half of what it does follows this, and half doesn’t. They are very effective at placing ads that match search queries, but that’s only because they use arbitrage to get marketers to figure out the placement for different keywords. The correlation is strong because matching key words of products I’m searching for to products is combined with information on past searches for those keywords. It’s again a case of knowing my past searching behavior will predict my future behavior. Then we have the ads on GMail which hilariously scan my email and offer me totally useless ads. My behavior with my friends is not an indicator of whether I want to buy more lube today. Sorry Google.

I also believe there’s another effect going on here that explains why Amazon and Google Adwords is more successful than other companies. If all anyone had was behavior information then behavior information would win out in the marketplace as the best information. The problem is Facebook is selling an inferior product in the market because voyeur marketing has to compete with information that’s based on real things I bought in the past. Voyeur marketing will be inferior as long as there’s companies who have access to the things I really bought.

This leads me to another prediction about social networks:

Most social networks will start offering marketplace options for direct sales to their users.

I’m actually baffled why this hasn’t happened on platforms like Instagram, Flickr, and Facebook. These platforms have many brands, many creative types, and many consumers. All they need is a “Buy This Art” button and handle all the sales details and they’d be minting money. Additionally it’d alleviate the creepiness of advertising if there was an actual marketplace someone could go to review products and buy. These companies would switch from the model of “We sell lots of users to advertisers who then…uh…” to the model of “We have the largest marketplace of users and the simplest path for them to buy your product.”

To be clear, since people can’t read, I am not saying merely putting a buy button on everything, although that’d be a great first step. I am saying these companies will create full on marketplaces to compete directly with Amazon. In a radical version of this I can see Amazon merging with a social network or inventing their own to gain this combination.

The idea in this is that the fallacy of building a large user base just to sell ads to marketers doesn’t work because the behavior of purchases is too far removed from the products on most social networks. Instead, the social networks need to close the gap and simply offer the products right there with the easiest possible path to purchase. Until then their information on who I’m having sex with can’t compete with Amazon’s information about how many condoms I bought last week.