• by Chuck Martin , Staff Writer, June 4, 2014

Getting the right message to the right person at the right time has been a long-time goal in mobile commerce.

The idea is that a highly relevant marketing message at the time and place most relevant to a mobile shopper could provide that consumer with useful information to aid them in a purchase decision.

All of which, of course, is much easier said than done.

One of the missing links in the process is the precision of the location data used to target the messages, as I wrote about here some six months ago (Missing the Target with Targeted Ads).

I caught up with Eli Portnoy, founder and general manager of Thinknear, this week to check on the status of location data and gauge the market reaction of his company’s research that shows that some location data is quite far off the mark.

“The reaction has been awe, shock and surprise in general,” said Portnoy.

His most recent research, based on reviewing 3.5 billon ad auctions on exchanges measured against 53 million impressions, found that just over a third (35%) of impressions were accurate within 325 feet.

The results were based on asking users to pull their GPS location data, which was then compared to the location where the ads supposedly appeared.

The study showed that about one in five (19%) of ads received were up to six to 60 miles off.

As might be expected, Portnoy said there is higher conversion when ads are more accurately targeted.

The reality is that an mobile shopper who receives a totally irrelevant ad just before they enter a particular store might be somewhat annoyed if not bewildered, though there’s not much a large number of them are likely to do but move on.

“Consumers don’t have an outlet to respond, even though they may constantly find ads that are so off,” said Portnoy.

Companies like Foursquare have an advantage in location accuracy since people explicitly check in at a certain location, providing a rich trove of location data, especially over time with millions of check-ins. Many others, not so much.

“This is one of the hardest problems out there,” said Portnoy.

At least it’s identified.