Facebook's Happiness Obsession, and Why It Will Break the "Like" System

Apple (AAPL) may be building Facebook's new Like system into the frameworks of its next iPhone OS, allowing users to "Like" everything from songs to videos to tweets. But adding the "Like" button to smartphones could seriously mess up Facebook's already-dubious advertising plans.

Let's start with the impetus behind the Like button. Look at the recent entries on the Facebook Blog (screen capture at right; click for a larger version) and you see a welter of recent posts from PhD interns on Facebook's data team. Many of the posts are larded up with fatuous little observations like this one:

We discovered that the more people use Facebook, the better they feel and that those who share and communicate the most with their friends feel even better.

Also detracting from Facebook's research cred is the fact that its stated inspiration is this sociology book, which basically posits that we Americans are all becoming more disconnected and lonely. Never mind that this theory has had been battered by a bevy of worthy counter-arguments (see p.3 of this MIT lecture document, or this article for more depth) and never mind that independent studies like one from Danish consultancy Red Associates suggest Facebook actually leaves people feeling largely unfulfilled. Facebook thinks it's saving us.

But if that's Facebook's dream, why tear it down? Because it suggests that Facebook's principals are placing too much faith in the Like button when its actual utility -- from a revenue standpoint -- is questionable.

The Like button is supposed to be Facebook's way to collect super-valuable data about what people are buying, or at least, what they want to buy. Amazon (AMZN) has built an empire on the same data; Mint.com's gunning for it, and some well-backed startups are too. But as AdAge (and others) have pointed out, the Like button is actually an incredibly sloppy way of gathering that data. Worse, it's easily gamed. Says a writer in AdAge:

I can't think of a more regressive and easily manipulated system than simply measuring the number of "likes" on a page as a means of not only evaluating advertising effectiveness but whether advertising dollars will be accepted. If you don't have enough likes or your product isn't the right fit, your ad dollars will be turned away.

The LikeRank idea is tough to choke down even for a data junkie at his weakest. If you show me a cheeseburger and ask me if I "like" it, at any given time of day, you are likely to get a different answer each time. Even when I like something, I may not feel compelled to click "like" unless, of course, there was some benefit to me. Then again, if you give me candy every time I like something, isn't that bucking the system?

Let's say you're in line at the bank, waiting in traffic, or being a wallflower at a party -- any scenario in which you, the smartphone user, instinctively reach for your phone, whether out of boredom, nervousness or just rote habit. No new emails? No texts? Now imagine you can spend that time hitting "like" on iPod songs, tweets in Tweetie, or bar pages in Foursquare. That kind of idle "liking" could result not from any real enthusiasm for a song or product -- maybe that bar or song is just okay, but not awesome -- but just for that little dopamine rush that we all seek when we're playing with our phones during downtime. Given the opportunity, I would probably "like" 90% of the stuff on my phone, because hey -- I put it there.

As AdAge says, this could make for an awfully diluted dataset. And if Facebook is going to try to make its millions on advertising and e-commerce partnerships, it's going to be basing its relationships on a cheap medley of semi-useful data and misguided goodwill towards users. Not the best foundation for a revenue engine, is it?