Less than 10% of Mobile Ad Impressions Are 'Hyper-Local' with Lat-Long

Last week I moderated an evening workshop about mobile ad exchanges and mobile advertising more broadly. The event was sponsored by DataXu and intended to introduce agencies to the concept and mechanics of mobile ad exchanges. It featured a mini-ecosystem of company representatives:

  • Rob Kramer, EVP Sales, Celtra
  • Jim Payne, CEO, MoPub
  • Brent Gaskamp, President, Collider Media
  • Iryna Newman, Mobile Marketing Manager, Groupon Mobile
  • Lara Mehanna, GM Mobile, DataXu

There were lots of interesting questions and issues discussed. It was a great event.

However I was struck by a comment made by Groupon's Iryna Newman during the session. I'm paraphrasing but she essentially said that she would pay a premium for as many lat-long mobile impressions as she could get her hands on -- but there simply aren't enough of them. 

This seems a strange comment given the much-touted location targeting capabilities of mobile apps and ad networks, and the frenzy around LBS and "hyper-local" advertising.

There are still numerous barriers to delivering lat-long information to advertisers. Privacy is one, especially on iOS. But many mobile ad networks are offering location only at the country, state or DMA level, without any precision beyond that. 

Some networks and publishers represent they can offer a lat-long but may in fact be "faking" it. 

On the mobile Web you're typically only getting IP-based targeting; and that faces the same accuracy challenges in mobile that it does on the PC. There's also a perceived lack of demand from advertisers for "hyper local impressions." However, the Groupon remark contradicts that very clearly. 

I was told by someone in a position to know that only about 5% to 10% of mobile ad impressions currently carry a lat-long. If accurate, and I assume it is, it's somewhat shocking given the rhetoric of mobile advertising and its targeting capabilities. 

There are various ad forecasts now in the market that argue that a substantial minority or a majority of mobile ads will be geo-targeted in the very near future. The analyst firms that developed these forecasts may be largely unaware of these fundamental "plumbing" and infrastructure challenges (mostly on the display side). In search it's a completely different situation and the same is true for individual apps.

Google, with the advent of Chrome for mobile, is seeking to remedy this for Android-based handsets. It will follow users from PC to mobile and also have much more data about them when they're mobile. 

As a general matter, there are display workarounds involving landing pages that can generate more location precision. But the industry currently faces a gap regarding what it says it can do and what it can actually deliver at scale.

Eventually there will be alignment. But I was quite surprised to learn about all these limitations.