On the one hand, it seems the world is getting less predictable. On the other hand, looking at the successes of Big Data and AI, haven’t we all at the same time become more predictable?
Isn’t that how those kids in Macedonia made thousands of dollars running fake articles on social media? Isn’t that how James Corden got famous enough to host the Grammys?
As I thought about this, I remembered that I’d thought about this before. About 11 years ago. Let’s see how 2006 sounds from the vantage point of 2017.
Fortune talked about recommender systems a few years back.
What’s a recommender system? Well, take Amazon’s “if you liked The Da Vinci Code, you’ll love Blink.” Now move from book-to-book relationships into book-to-other relationships: “If you liked the Da Vinci Code, you’ll like a Jura Capressa espresso maker.” That’s a recommender system.
Fortune’s example was www.whattorent.com, helping slackers save time at 10PM Friday night at what was the local Blockbuster by predicting what movie they’ll love. [Remember Blockbuster? Just eleven years ago…]
Fortune interviewed whattorent’s two founders at a coffee shop, and put them to the ultimate test: pick two strangers in this restaurant, and—just by observing them—guess their favorite movie.
They settled on a guy and a young woman. After much clever psycho-babbling, the founders guess: Starship Troopers for Joe, Roman Holiday for Renee.
And wouldn’t ya know it—they were dead right.
You can hear Fortune cuing up the PGA graphic—“these guys are good!” And indeed that’s our reaction—wow, how could anyone pull that off?
But wait. What if we’re mixing up cause and effect? Maybe it’s not that two twenty-somethings are great predictors. What if we’ve just all gotten way more predictable?
Everyone had their favorite Beatle. If you preferred John to Paul, it said something about you—to everyone. Because everyone had a common reference point. The Fab Four were global litmus tests.
Since then, culture got way more global. Africans wear Arizona t-shirts; Valley Girls know Tibetan monk choirs. The weapons of mass dispersion are well known—iPods, MySpace, YouTube, Hollywood [can you believe – this was only 11 years ago…the iPhone was still a year away…]
Everyone wants to be different—but we share referent points from which we diverge. Jeans, music, hair, slang… Take five variables with five values each: five to the fifth power is 3,125 combinations. Sounds like a lot, but it’s based on a small set that’s easy to reverse-engineer.
People don’t predict us: we self-identify, and the code is easy to read. Marketers love this stuff.
Ironically, it also makes it easier to trust others. When a British Stones fan meets a Jagger aficionado from Beijing—the world shrinks.
The question is: can we keep the diversity while enhancing the trust?
Well, that was my question then. My question now is similar, but updated: can we keep the authenticity while mechanizing the means of connection?
This is most evident in commerce. You still know, in your bones, when you receive a mechanized spam email, trying to pass itself off as personal. I suppose scams may be getting more sophisticated; but a ton of people aren’t even bothering to be sophisticated. They confuse the ability to target and segment with the desirability of doing so. Just because you can doesn’t mean you should.
We’re all pretty predictable. That’s OK. Go ahead, predict me – just let me know there’s someone behind the prediction machine, someone who cares enough to add the whipping cream topping by making it personal.
The difference between being sold to by a person and being sold to by an algorithm is the difference between talking to a person who used a robot to find me, and talking to the robot itself. I don’t mind being predicted – just don’t insult me.