Blog Post

JiWire builds a location graph to make mobile ads relevant

As Internet expert Mary Meeker of Kleiner Perkins has helpfully pointed out, mobile still has huge monetization potential that is still largely untapped. Location-based advertising may prove to be one of the key ways in which mobile Internet companies can boost mobile advertising revenue and close the gap with traditional Internet ads.

Location-based ad provider JiWire is trying to capitalize on this opportunity with the launch of a Location Graph that it says will help it deliver relevant ads based on what it can learn from the way people move around from place to place. By understanding the relationship between locations and what it says about the people who frequent them, the Location Graph will help JiWire anonymously identify distinct audience groups, from soccer moms and businessmen to luxury shoppers and tech enthusiasts.

JiWire has a database of more than 3 billion location tags pulled from latitude and longitude information and other data that connects to real places. As the Location Graph pulls in anonymized location data from mobile devices using UDID and other persistant identifiers, it can start to understand where people are visiting with their devices. Then it can map that movement to what it knows about the various locations and the way certain people favor certain places.

This database allows JiWire to go beyond simple targeting based on what app an ad appears in or geo-fencing, which only looks at the general proximity of a user to a location. With the Location Graph, JiWire can better identify a user for an advertiser, who can choose to try to walk them into the store with an incentive if they’re nearby or push out general brand advertising to build awareness. Having a history of where a user has gone can also allow JiWire to push ads from advertisers that are not just nearby at that moment, but are also close by the places where a person has often visited in the past.

JiWire’s interim CEO David Staas said early beta testers, including Microsoft (s msft) and Comcast (s cmcsk), have seen a 30 to 45 percent lift in click-through rates when using the Location Graph. He said the Location Graph shows how mobile advertising can become lucrative for publishers and effective for advertisers.

This reminds me of Sense Networks, another mobile advertising startup that is targeting ads not only based on place but also by the unique behavioral profile of users. The approach is a more sophisticated take on location-based advertising and it means that different users in the same place can get more personalized ads based on who they are and where they like to go. Users, however, have to be comfortable with the fact that companies are gathering anonymous information about their whereabouts. JiWire is trying to allay fears by giving users an opt-out option and also using TRUSTe to handle the user location data.

Here’s a look at some of the connections the Location Graph is able to find between places and people.

5 Responses to “JiWire builds a location graph to make mobile ads relevant”

  1. The problem with location based ads is that the cost to filter down a targeted ad (server look ups and profile data storage) is usually higher than the premium an advertiser will pay for the ads. Also the reach of the ad campaign diminishes as more geo targeting is applied… And advertisers love reach!

    I’m sure these hurdles will be overcome one day but it’s certainly not a quick win… Good luck JiWire

  2. U.S. Media = Hacks

    Lol women eat out more than men, no wonder none of them know how to cook these days! What happened to all the real women. Now days we just have Snookie wannabes. Fake, stupid, lazy and shallow.

  3. Sense Networks

    Thank you for the mention & congrats on the great new data! Yes, we’ve been building unique behavioral profiles from historical location for very sophisticated, individualized messaging for a few years. Prioritizing privacy & not storing raw data is critical, as the future holds predictive re-targeting at scale… [email protected]

  4. Locomizer Team

    It’s great to see that the idea we’re working on is being validated by this announcement! While our implementation is different the direction is similar to JiWire’s. We’re building Geo-Behavioral User Interest Profiles based on user local footprint but in the much deeper context. Plus, we found a solution to privacy concerns in regard to sharing personal location data by not storing it at all. alex @locomizer

  5. Dev Bhatia

    As I’ve posted elsewhere, this is good stuff. All they have to do is overlay the location data with some olf-fashioned offline segmentation analysis, and they are onto something. Location is certainly an important variable in mobile ad targeting, maybe THE most important variable.