Mapping Physical Spaces for Ad Targeting: A Conversation with xAd
Does my mobile location say that I’m in a Target store? Or am I in the parking lot? Or in the Starbucks within the Target? Or am I driving by at 60 mph? These details matter for contextual relevance in ad targeting. But most mobile ad placements fail to differentiate.
xAd is looking to change that with its newest tool, Blueprints. It announced the product late last week, but we had previously gotten a sneak peek during an ad targeting session at BIA/Kelsey NATIONAL. I’ve embedded the highlight reel from that session below.
Blueprints builds from xAd’s “footprints” product, which has been gathering data in the marketplace for months (see our launch coverage). By tracking user spatial behavior, it’s been able to build a robust data set of user location and movement patterns.
Blueprints takes that a step further by overlaying the physical orientation of different businesses, using satellite imagery. For example, many stores (think strip malls) are set back hundreds of yards from the street, where place data usually centers.
This is the kind of thing that makes a lot of sense at, say, a car dealership. Pinpointing someone on a dealer lot has considerably different ad messaging implications than someone driving by, or waiting in the maintenance/service area of the dealership.
“How do I know someone is on a parking lot? How do I fence that?” said xAd platform head Dan Hight. “We created a way of making polygons on the fly, and we’ve hired blueprint editors as we’re calling them, to do this for the top 350 brands and the larger scale of the retail block.”
We previewed this latest announcement (including a visual demo) in an on-stage interview with Hight, along with some of the broader tactical matters of location targeting. The video is below, and stay tuned for lots more conference footage.
Location technology is difficult to do “correctly” and this guy makes two major broad generalizations/misrepresentations.
First, when asked what they do differently for accuracy of location, he responds with “derive lat/long from actual lat/long”. What does that mean? He thinks making fake coordinates from real ones are more accurate?
Second, he says it is fair to assume more people opt-in with their network versus any other major network – why would that be the case? There would be no way to measure that through SDKs of their own network let alone know that of competitive networks.
We should be asking more challenging questions and holding tech companies responsible for answering them truthfully.