Infographic: What is Predictive Modeling?

What do you get when you combine artificial intelligence and big data? In the world of online advertising, you get predictive modeling. It’s a highly sophisticated method for getting close to your customers – closer than ever before.

Targeting has been around for a while, but until now, it has mostly involved making some guesses about the kind of people who will buy your product, and then placing ads where you think they’ll be (aka demographic, contextual and geographic targeting). It can serve up a pretty narrow audience, which may seem like you’re hyper-targeting; but if they don’t convert, you’re back to square one.

Retargeting can take things up a notch, but predictive modeling is some next-level stuff. It starts bigger than traditional targeting, creating an algorithm with millions of data points in addition to the basics (like shopping habits, social activity, even weather) and casting a pretty wide net initially. That’s when the artificial intelligence part kicks in, tracking every impression and every conversion in real-time and then adjusting the algorithm to continually hone in on the audience that’s actually buying what you’re selling.

But predictive modeling isn’t just changing how we target – it’s also changing how we measure. No longer is the click-through rate the be-all, end-all metric. Because it tracks so much data, predictive modeling can also measure things like brand favorability and intent to purchase – and optimize your campaign accordingly.

Check out the infographic for more on predictive modeling.