Machine learning plus more machine learning for travel marketers

A few months ago several of us attended the Phocuswright Conference in Florida. All this time later, we’re still talking about a presentation there by Oliver Heckmann, vice president of engineering for Google Travel.

Heckmann shares our belief in the power of machine learning for travel marketers. He even argues that the science of machine learning has advanced so quickly that it’s become clear the world is in the early stages of a “second machine age” that will revolutionize life itself. 

Say what you will about Google (and we’ve been known to urge caution around the “datagopoly”), you have to give credit where it’s due. Google is right: machine learning is changing everything.

Toward the end of his presentation, Heckmann focused on the use of machine learning in online marketing for travel companies, noting how Google has learned to predict click-through rates (and thus help determine optimal placement) and to help customers more effectively automate bidding. 

Machine learning, Heckmann said, makes all of Google Travel’s products “more natural and more useful” to users. And again, giving credit where it’s due, Heckmann is right. Google has built tools that are pretty useful to e-commerce marketers and consumers.

Machine learning for travel marketers

But here’s the thing: those tools could be better.  

Google can tell you plenty about a visitor to your site … particularly if that someone is “valuable” or not to e-commerce in general. But that valuation isn’t customized. It’s based on things that are of interest to any e-commerce site (page views, time on site, etc.) It’s not based upon the unique patterns demonstrated on your site by your customers looking at your product mix. It doesn’t account for how your brand and your customers interact.

So at <intent> we’ve used machine learning to make Google’s tools better.

Our <intent> Target product layers intelligence into your Google Search and Display Campaigns … thus allowing us to make real-time predictions to set acquisition bids. We use machine learning on your site, using your first-party data, to understand your users.

That optimizes your acquisition strategy by 20% or more.

If you want to know more, drop us a line at

And if you’d like to hear Heckmann’s presentation, we’ve pasted it below.