After model Kate Upton sashayed through the Super Bowl ad for Game of War this year, installs of that mobile game soared.

But can advertisers measure the exact impact of that ad — and others — on app installs?

Oren Kaniel, chief executive and cofounder of mobile ad analytics platform AppsFlyer, told me that advertisers of mobile apps have been flying blind when it comes to TV ads. Unlike clicks, display logs, and other digital trails for online ads, TV ads just appear — and are gone, with no tracking codes, cookies, or referrers.

But his Tel Aviv-based company is intent on changing that, with today’s launch of what it says is the first capability to attribute app installs to TV ads.

AI Weekly

The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.

Included with VentureBeat Insider and VentureBeat VIP memberships.

The days when app companies advertise on air, “without knowing if it works or not, just because the agency [says so] — those days are over,” he said.

AppsFlyer’s solution: attribute an app install to the “last touch,” or the last ad that likely had an impact on the installation.

It does that by setting different windows of time during which it assumes various ads are effective: a few minutes for a TV ad, 24 hours for an online ad without clickthroughs, or 7 days for an online ad with clickthroughs, even if they don’t immediately lead to an install.

When an install happens, AppsFlyer attributes the last touch based on which of those windows are closest in time to the install time.

For example, let’s say an ad for the Clash of Clans mobile game airs on a Boston TV station at noon on a Thursday. AppsFlyer’s platform has access to the airtime data, and it starts a 20- to 60-minute window of attribution when the ad airs. The actual window is configurable by the advertiser and can be longer, but let’s assume it’s 30 minutes.

Let’s also suppose that earlier in the day, at 10 a.m., advertising networks showed online ads for Clash of Clans in the Boston area, but nobody clicked on them. If there is no click trail, the ads are assigned a 24-hour window of attribution in the AppsFlyer platform.

If an installation happens at 12:20 p.m. Thursday, AppsFlyer will attribute it to the TV ad, because it is the closest “touch” of the overlapping windows.

The actual attribution environment, of course, is much more complex, with more than a thousand media sources’ ads being tracked. This “last touch” approach is a purely circumstantial way of giving the TV ad credit for the install, but, given the current media ecosystem, it may be the closest way without actually interviewing the user.

An app install attribution screen in the updated AppsFlyer.

Above: An app install attribution screen in the updated AppsFlyer.

Image Credit: AppsFlyer

Kaniel sees this approach as “the start of a TV ad revolution.” His platform does Lifetime Value calculations of the TV ad, based on the fact that an install means the app maker has just landed a user — at least until she uninstalls the app.

The platform also measures return-on-investment (ROI) calculations, using post-install actions inside the app like booking a hotel room in the Hotels.com app or buying a boost in a game.

And attribution can have big consequences. If the TV ad instead of a mobile ad led to an install, the mobile ad network might get paid less.

Kaniel said that no other company — including such competitors as Tune, Apsalar, Adjust, and Kochava — currently has this capability.

With that capability, Game of War developer Machine Zone might have been able to tell whether spending $4.5 million for 30 seconds of airtime was really worth it.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More