The hotel I stayed at in San Diego last month has been following me around for weeks. Seriously, it’s getting annoying – a one-night-stand that refuses to recede gracefully into memory. Everywhere I go on the web, it’s there waiting for me, promising me a “great location in the center of town” even though my stay has long passed.
That hi-tech workout clothing I ordered online last week? (very cool – it has embedded sensors that measure exertion and performance) There it is today on my favorite blog exhorting me to be among the first to reserve before the waiting list fills up.
Ah, Big Data. Making lives so much better, no? Advertising legend David Ogilvy once famously said that 50 percent of all advertising is wasted. The problem is you don’t know which 50 percent.
Now that every website we visit is generous with the parting gifts – one estimate reports that the average website deposits a dozen or more tracking cookies in our browsers every time we visit – advertisers, publishers, and anyone else who has an inclination can track where we go and what we do online. Now there are “super cookies” that can’t be deleted. And just wait as real-world sensors embedded in the everyday objects around us increasingly measure what we’re doing and where we go in real life.
This isn’t necessarily evil. I appreciate that Google Now mines my behavioral data to try to give me the information I’m most looking for when I need it. And Netflix and Amazon and Facebook and Expedia are getting better and better at learning enough about me to help sort out the choices I’m overwhelmed by. Driverless cars will be safer and home monitoring will help make us more energy efficient.
Measuring behavior is a much better predictor of future behavior than is traditional demographic research. Given that we are getting better at measuring everything around us, those who study how individuals and communities make choices are getting more sophisticated at understanding and offering us more of what we demonstrate we want (as opposed to the sometimes wildly inaccurate things we say we want).
Or are they? In the early days of the web, we were all impressed by how an explosion of choice fragmented the audience. The long tail hinted at the viability of unprecedented diversity. No longer did we have to settle for mediocre one-size-dumbs-all generic TV. Or books. Or music, movies or even spaghetti sauce. It was the end of the mass audience.
Except not. In the latest generation of the web, today’s mass audience dwarfs yesterday’s. It’s possible to reach and influence literally billions rather than mere millions. People and things you’ve never heard of now thrive with millions of followers. Some of the oceans of new content is inarguably better than it was. But the numbers also confirm that enormous numbers of people evidently prefer generic or mediocre (whatever that means these days). And “organic”, “artisanal” and “local” have been corrupted to be mass-produced.
And we’ve moved on. Sophisticated companies are no longer impressed by mere volume of clicks or page views (hear that little #thedress?). Is there anyone who seriously believes that Buzzfeed with its tens of millions of page views is more influential than The New York Times?
So the question is how to measure that influence. And what constitutes influence anyway? And on whom? The measurers are trying to determine attention and impact (though anyone who has tried to interest a large company in corporate sponsorship will find their heads spinning by the generic CPM ROI formulas they routinely use to measure the value of marketing deals).
For however generic and ham-handed these calculations are, the arts are even less sophisticated in trying to understand their audiences and potential communities. The amount of promotional content currently being cranked out by arts organizations is astonishing. Mostly so for the waste of effort and resources. Yes the production values are better, but most of it is catchup to 1990s-era messaging. And it is A: getting lost in a sea of other content, and B: while possibly entertaining (the best of it) it is ineffective in driving ticket sales or building community.
Meanwhile, the biggest advertisers, having learned that even the most beautifully-produced banner ads, video spots and blog posts are losing their effectiveness, have moved on to other things. Those things being finding out more about how behavior predicts choice and crafting tests and refining models that help them understand and shape behavior. Hollywood is going gaga over this kind of data research. And there are signs of success. This year data scientists used behavioral data to accurately predict Oscar winners.
Heady stuff. And no wonder everyone is excited about the promises of Big Data. But increasingly the problem isn’t getting enough data, but understanding what it is you want to measure and why. That means better understanding what you want to find out and asking better questions. If they’re the wrong questions the answers aren’t of much use. If I could only teach my hotel that maybe it would leave me alone.