Evan Schuman's StorefrontBacktalk

Gap last week ended a 2-week trial on a geofencing mobile ad effort, one that reinforced traditional billboard ads with mobile messages displayed to people standing right beside those ads. In some cases, those ads were right in front of Gap stores, and therein lies untapped mobile potential.

The initial test used the shopper’s physical location, but no other personal data (such as purchase history, other apps on the phone, Web search logs, personal demographics). However, personalization will likely be the subject of upcoming trials, said Dave Etherington, the SVP for marketing and mobile at Titan, the advertising firm that executed the Gap trial.

Using geolocation in front of stores is a retail marketing technique that goes back more than 100 years. The idea behind mobile geolocation marketing is that the store knows for certain that the customer is standing right in front of the store and can tailor messages to someone standing right there. Hasn’t that always been the rationale behind department store front window displays?

The twist with mobile is the potential to match customer-specific info with customized ads, such as “This beautiful white, textural drop-shoulder sweater would go perfectly with the Prussian blue, Rayon ruffle tank that you purchased two weeks ago. 35 percent off today only. And we have just two in stock in your size.” (If you really want to freak customers out, you could add, “But according to the security camera atop the window display, you seem to have lost weight. All alterations free, if you buy today.”)

Beyond customer customization, the mobile ads could also be customized by the time of day and other factors, Etherington said.

A Bloomberg news story on Wednesday (March 14) about Gap pointed out a frightening theory: that the chain’s key audience—18 to 34 year olds—is is being so battered by high unemployment and economy fears that it is cutting back discretionary shopping.

Assuming that’s true—and, to a certain extent, it almost certainly is—it may not be a matter of just cutting prices but making purchases so logical and timely that it’s hard to resist. An app that knows your current wardrobe and size and uses that information to make legitimate recommendations could be powerful. The same would be true for an electronics chain that knows about the life expectancy of devices you own and pitches you when they are likely breaking down.

Or—as Wal-Mart is experimenting—taking social media databases and calendar data and recommending a new business suit for an important upcoming speech or a new dress for a wedding.

The initial Gap trial, though, was far less adventurous. It used a series of ads on buses, train station platforms, stores and shopping areas in New York City, San Francisco and Chicago, and then coupled them with mobile ads that were served within “a ton of apps,” Etherington said, including The Huffington Post, Pandora, Words with Friends and NBC Sports.

The ads were only served to shoppers within the limits of various geofences, typically “within specific ZIP codes or blocks around specific media,” Etherington said. Although a ZIP code might seem like a fairly large geographic area for such a targeted campaign, it made sense given the nature of the ads. An ad on the side of a bus driving around midtown Manhattan, for example, could easily be seen by people throughout one ZIP code.

The advertising point of the initial Gap geofence trial was that reinforcing a physical ad with a mobile one would sharply increase shopper response (click-through). The typical mobile ad response is about 0.2, but the results of this trial hit almost six-and-a-half times that rate, Etherington said. “It helps (shoppers) connect the dots,” he said, connections that apparently happened 25,000 times during the 2-week trial.

Like so many ads, though, these were tracked for click-through, not sales. If purchases happened throughout the Gap site, for example, that could be easily matched to the ads, especially if the purchase happened right away. A purchase that happens three weeks later could still be matched, but it would require more flexible—and complex—scripts.

The real bounty would be if the store was able to close the loop between that mobile customer waiting for that train—the one who saw that billboard and then clicked on that ad—with that customer making a purchase in-store a month later. I know: Single-view-of-the-customer is so 20th Century. Then again, so are window displays that know your location.