JiWire Rolls Out Attribution Product As Mobile Ad Market Zeroes In On ROI

If 2012 was the “year of mobile,” 2013 is the year of proving it works.

The conversation in the mobile advertising industry has shifted focus in recent months, moving away from targeting toward attribution and measuring return on investment as brands begin to invest meaningful spend in mobile and, in turn, expect results. Following a string of product releases from competitors, JiWire, a San Francisco-based mobile-local ad startup, has rolled out a new attribution tool called Location Conversion Index this morning, which draws on a similar technique used by competitors to measure in-store visits but allows marketers to normalize those numbers against a wider sample.

Here’s how it works. Like the attribution products released by Sense Networks and PlaceIQ in the spring, JiWire measures in-store visits by indexing the device identifiers of the devices to which it serves ads. Then, listening to the billions of ad requests passed through its network every day, the service records when an ad request from a device which it has already served an ad originates from within its client’s store. However, JiWire’s product normalizes the in-store conversion rate (the percent of users who were shown an ad, and were seen on an ad network within the store later) against a control group, effectively separating the wheat from the chaff in a mobile campaign.

“The real challenge is understanding how do we separate a normal store visit, which would have already occurred, from an incremental store visit that happened as a result of seeing an ad,” said David Staas, the company’s president. “When you break it down to the building blocks, when you measure foot traffic you’re measuring a real-world behavior. And when you do that, you have to take into account a range of other factors.”

When a campaign is launched, and a marketer selects an audience (maybe women under 50 who have visited a target in the past year) the company creates a separate control group with a look-alike audience, as well as a third, more general, collection of users. At the end of the campaign, the company compares the performance of the users exposed to the ad with their historical behavior, as well as the behavior of the control group and the general users, and then provides the advertisers with information about the effective lift of the campaign.

For JiWire, the normalization technique undoubtedly improves the accuracy of the in-store metric, but it’s still unclear whether the metric itself will satiate marketers’ desire for transparency. The base technology still relies on a sort of happenstance action, where users who were shown an ad open an app (which uses an ad network used by the company serving the original ad) while they’re in the store days later. It’s a solid reporting metric, but lacks the consistency needed to price against in the way advertisers can buy TV ads based on Nielsen ratings.

One thing is clear: the product focus in the mobile-local ad sector has shifted from targeting to attribution, and the pressure from marketers develop clearer attribution metrics will only increase over the next year. In a sense, it’s a consequence of mobile’s success, said Staas: “As marketers start to shift a higher percentage of their budget in into mobile, and it becomes 10% plus of their marketing budget, it’s not just test-and-learn anymore,” he said. “They’re under a lot more scrutiny to show their own effectiveness and demonstrate real return on investment.”

Steven Jacobs is Street Fight’s deputy editor.