How Expedia ads now trail you from desktop to mobile
In a travel industry first, Expedia has begun “re-targeting” advertisements at its customers who switch between desktop and mobile devices.
Between November 2012 and March 2013, more than 55 million people who visited Expedia’s website via their desktop computers and later switched to using mobile devices saw advertisements on their mobile browsers encouraging them to download or use the Expedia mobile app.
No other online travel agency has ever re-targeted and tracked ads across device platforms at the same users on a similar scale.
Cracking the code of cross-screen marketing
Advertisers such as Expedia have long known their customers’ desktop browsing patterns, thanks to cookies or “pixels”, which are bits of code that let advertisers track a user’s Web browsing activity and serve relevant ads based on that information–in Expedia’s case, via a third-party French company called Criteo.
Yet tracking the whereabouts of customers on mobile devices has been out of reach. That’s mainly because mobile industry players have yet to agree on standards for third-party cookies and pixels.
Expedia hired mobile ad technology provider Drawbridge to pinpoint when a mobile user is probably the same person as a user who had recently done searches at expedia.com via a desktop browser. Drawbridge was eventually able to re-target 38% of the desktop visitors on mobile devices, or about 54 million users.
Drawbridge does its work by noticing common patterns in cross-screen metadata by analyzing records of when ads have been served to those devices.
It uses this anonymized metadata, such as desktop and mobile devices that use the same Wi-Fi signal at similar times of day, to make educated guesses about when a mobile device user may be the same anonymous user who previously visited a website on a desktop browser.
Last week, Drawbridge’s desktop-to-mobile retargeting campaign for Expedia was honored in the SEEK Awards as the best digital marketing initiative of the year. On June 27, the startup’s Expedia campaign won the travel industry marketing award in the 2013 Digitals.
Higher conversion rates
The cost per booking — Expedia’s favorite metric — was much lower for re-targeted ad campaigns than for the control group. Re-targeting boosted downloads of Expedia’s app by people whose search behavior suggested they were primed to book. That led to more reservations.
But just as importantly, re-targeting also helped Expedia’s marketing team correctly attribute which of its campaigns (across multiple channels and with varied messages) was generating the most click-throughs and, ultimately, sales.
In a series of A/B tests, Expedia’s retargeting campaign lines had conversion rates 336% higher than non-retargeted lines, according to the company. A person familiar with the campaign tells Tnooz that click-through rates jumped about a hundred basis points (one basis point is one-hundredth of a percent), though neither company’s representatives confirmed that.
Drawbridge saw how many of these users came back to the desktop from mobile to take action. That was a fresh data point for Expedia. Until now, Expedia’s mobile ad campaigns might have been converting customers, but the company had no way to attribute the proportional value of a sale to any given one of its mobile ad campaigns.
By adding a virtual breadcrumb trail, Expedia can now track a consumer’s path and see if it leads to a booking later on a desktop browser. The company can then hone its message and marketing techniques to reach and convert similar prospects more effectively.
Expedia also now knows how many of its customers are true multi-platform consumers and it can tailor its marketing spend accordingly.
Rosenblum explained that the startup, which has six ex-Googlers out of its staff of 40, uses software and proprietary statistical data mining techniques to filter through cross-screen metadata.
Drawbridge observes ad requests on desktop ad exchanges, such as Google’s DoubleClick Ad Exchange, used by all advertisers placing digital ads on desktop browsers. It observes similar ad requests at mobile ad exchanges, such as MoPub and Nexage. (It gains access to these databases when it places bids on behalf of Expedia and other companies on the ad exchanges.)
The startup logs that information and correlates it.
When a user who visited expedia.com on their desktop brower and looked at search results, rate details or an information site page, Drawbridge received cookies (or pixels), which help it study cross-platform behavior.
Case in point
Say you live in Palo Alto and the first thing you do when you wake up is turn on your smartphone (as one does in Palo Alto).
By checking your e-mail and the weather, your actions typically send ad requests to these servers. Your wi-fi signal is associated with an Internet Protocol address. Your browser, your device’s ID, the timestamp, and other associated information is also shared.
After showering and breakfast, say you fire up your laptop to check more e-mails and the news. Again, your actions send out ad requests, with similar metadata, but this time the metadata is logged with the desktop ad exchanges.
In both cases the metadata is, in effect, anonymized in the sense that your name and mailing address and credit card number aren’t revealed. But many parts of the identifying information are similar.
Logging 4 billion ad requests a day, Drawbridge uses statistical probability to make guesses with high levels of confidence that the devices are likely to be owned by the same household.
It looks for desktops, laptops, smartphone, and tablets that access the Internet in the same geographical locations during the same time periods each day and whose users have similar Internet usage behaviors across those correlated devices.
On a database of a trillion observations of ad requests, the company claims to have matched 600 million unique desktops (via cookies) with mobile devices (via educated guesses), including millions of Android, Blackberry, iOS, and Windows Phone devices.
After making its conclusions, Drawbridge’s self-service APIs flow directly into the reporting systems of Expedia.
It’s not surprising that Expedia is the first travel company to experiment with cross-platform re-targeting, says Rosenblum:
“If you took all of the companies globally and ranked them in sophistication and skill for performance advertising, Expedia would be the top or at the top.”
Sean O’Neill is the Editor-in-Chief of Tnooz and is based in southern New Jersey, in the US.
Before joining us, Sean was a regular contributor to BBC Travel, a senior editor of BudgetTravel.com, and an associate editor at Kiplinger’s magazine.
Follow him on Twitter (@sean_oneill).