JiWire Rolls Out 'Location Graph'

Location-based ad company JiWire Wednesday unveiled a new ad service designed to help marketers target specific audiences based on the billions of pieces of location data it has collected. The company's Location Graph harnesses that data to create anonymous user profiles based on the types of places people visit, from beauty parlors to drug stores to airports.

Instead of defining audiences by social connections, like Facebook's famed social graph, JiWire's new targeting capability relies on historical (and present) location data to improve campaign performance.

The Location Graph offers deep insights into user behaviors, some more expected than others,” said David Staas, interim CEO of JiWire. "Using location, we can see how similar audiences tend to exhibit the same patterns and compare them to other audience segments, like moms versus dads versus students, etcetera."

Staas explained that through its network of public Wi-Fi locations and third-party mobile apps in which it serves ads, JiWire has profiled over 500 million devices and created more than 3 billion location data tags. That database is expected to double to 6 billion in the next six months, with 16 million location data points added each day. “We take all those location tags, and through our algorithms we begin to understand which locations become indicative of different audience segments,” he said.

A mobile user, for example, that visits a children's park, a grocery store, zoo and maternity store within a given month, suggests that person is a parent. Such information can also help marketers to target look-alike audiences. Among other findings from use of Location Graph data:

-60% of women eat at the same three restaurants each month.

-23% of people who go to Peet's also go to Starbucks.

-43% of those who shopped at Best Buy also shopped at a competitor's location.

What does it all add up to? JiWire reports that in beta testing with ad partners using Location Graph to target campaigns among its audience of 55 million on mobile phones, tablets and laptops it has seen a 30% increase in click-through rates. The company also notes that brands such as Microsoft and Comcast are using the service to boost campaign results.

Staas added that JiWire will incorporate Location Graph data broadly across its ad offerings rather than selling it as a separate, premium service. “We want to bring it to bear on as many campaigns as possible,” he said, including both brand advertising efforts and those tied more directly to a user's immediate proximity. JiWire ad formats include Compass ads, for reaching nearby consumers, in-app ads, and display ads that run on the home page of its Wi-Fi network of hotspots in cafes, airports, hotels and other locations.

JiWire's Location Graph naturally raises questions about privacy. For example, its announcement highlights that moms who go to zoos also tend to frequent beauty salons, child care centers, counseling services, and restaurants. Who wants to be tracked everywhere they go, especially when it comes to more sensitive visits like going to a counseling center?

Staas reiterated that the Local Graph information is anonymous, aggregated data that doesn't include personally identifiable information (PII). The Location Graph is also certified by privacy services provider TRUSTe, whose Mobile Ads initiative allows people to opt out if they don't want to receive targeted ads on the mobile Web or within apps.

Still, mobile users are increasingly wary. According to a survey by Truste and Harris Interactive released last month, 85% of smartphone users won't download an app if they don't trust it. And across all platforms, consumers were 60% more concerned about privacy issues than last year.

Recommend (2)
1 comment about "JiWire Rolls Out 'Location Graph'".
  1. John Grono from GAP Research , August 15, 2012 at 9:33 a.m.
    Just because 60% of the women in their data base eat at the same three restaurants each month does NOT mean that 60% of women eat at the same three restaurants each month. You cannnot extrapolate to the entire population from a sub-set of that population unless it is representative of that population.