Commentary

For Marketers, Big Data Is Not Always Better Data

  • by , Op-Ed Contributor, September 7, 2017

For all the information that Google has about you — and it’s a lot — there is one data point that earns the company more revenue than any other: namely, your searches.

Your search activity is so valuable because it is a precise measurement of what you want at a specific time. And naturally, marketers are willing to pay to reach the right person at the right time. Indeed, for marketers, that information is the holy grail.

By contrast, the sum total of every place you’ve ever visited on the web is not nearly as useful to marketers. Even if some expert analyzed every financial transaction you ever executed online or every social media post you ever made, that data in aggregate has far less predictive power than information about what you want at this very moment.

And yet, in the world’s collective worship of all things Big Data, many have lost sight of what most accurately predicts buying behavior. It’s not just the size of the data -- it’s the precision of the data that really matters.

advertisement

advertisement

In other words, the truly effective way to know what a person wants at any given time is for the person to tell you precisely, as we all do at Google when, say, we’re shopping for new socks. (Argyle and 100% cotton, please.)

And yet some in the data business persist. Take BlueKai.com, a cloud-based Big Data platform that enables companies to target ads at you. Check out the BlueKai Registry and you can see all the cookie-based interest segments associated with your specific device that BlueKai has aggregated on you. This information is used by marketers to slot you into certain categories — and to target you with ads based on your browsing history.

I recently went to the “Hobbies & Interests” category of the BlueKai Registry to find out what marketers can see about my browsing behavior. I had to laugh. Marketers must think I’m the Most Interesting Man in the World.

For example, BlueKai reveals that I’m into Jewelry, Fine Dining, Vitamins, Stamps & Coins and 200 other diverse and fascinating areas of interest, while, in fact, these are simply topics that I’ve read about on the internet recently. They’re not really my most passionate personal enthusiasms.

This shows that the digital trail we leave has limits in terms of predictive power. Statistically, I would say that my digital trail indicates maybe 5% of what I’m truly interested in. And no matter how much ad networks would like to convince marketers otherwise, the other 95% of what interests me as a consumer isn’t going to be found by mining my web-surfing results for more data points.

To accurately discern that 95%, marketers would need extremely precise data on what’s in my fridge at that particular moment. Or what’s on my mind after I leave work. Or what my kids said to me during our last conversation.

What’s needed by marketers isn’t ever-bigger volumes of data. What’s needed is different, more diverse data. I’d argue that this is why Facebook does a decent job predicting what you want in your newsfeed. Because it’s able to see you from many diverse angles and understand you as a whole person.

Don’t get me wrong. Scale is indeed helpful when it comes to data. But what’s far more helpful to marketers is precise data that allows them to target individual consumers with better understanding and accuracy.

Even the smartest AI-powered neural network, leveraging mountains of data about you, will always miss something. It may know what music you stream, but to be truly effective, it also needs to know which songs you used to sing on road trips with your friends or what your wedding playlist included. By the same token, Netflix will never get your preferences exactly right until it realizes that you think Caillou is the most annoying TV character ever, no matter how much your 4-year-old loves him.

If we want to move into a world where companies are recommending things, not marketing them, we need to move beyond boasting about the size of our data sets and instead think about what data we need to truly understand consumers on an individual basis. That’s the real holy grail.

Next story loading loading..