Factual’s Observation Graph Grapples With Connecting Location To Behavior

Location needs context to be understood. WPP's Xaxis is putting Factual's real-world activity analytics to the test.

Knowing where something is far from knowing why it’s there and what’s happening at a given point on a map.

Factual is promising greater contextual information about locations and the activity at those places with its Observation Graph, which is designed to buttress the company’s Geopulse Audience offering. That product is part of Factual’s umbrella product, Global Places, which gathers data covering over 100 million local businesses and points of interest in 50 countries.

The new tool also provided Factual with the occasion to extend its WPP Group data partnership with Xaxis, which has been testing the Observation Graph on behalf of agency clients.

Observation Graph is currently arranging geo-data on more than 200 billion observations on over 300 million mobile devices per month. As Factual has struck agreements to provide location data for brands such as Apple Maps, Microsoft Bing, Facebook Places, and Uber, the demand for deeper insights has grown as well.

“Whatever data we see about a user, we try to turn those into observations,”  Vikas Gupta, Factual’s director of marketing, told GeoMarketing, following a panel discussion at the Local Search Association’s Place conference in Chicago this past week. “An observation can be that a user was at a place. And we can observe that the user was at a Starbucks at a certain date and time. An observation can be an activity. We can detect with the right set of data if they’re walking, if they’re running, if they’re in an elevator.”

And that behavioral topography, if you will, gives brands more color about what to do with the wider consumer data they’re given.

The creation of the Observation Graph was always a function of what Factual could do for a client. The issue was that the attributes — people who go to Starbucks in downtown Manhattan on weekend mornings, for example —  always needed to be assembled from scratch.

Aside from saving time for clients and Factual engineers, the Observation Graph is a model that allows clients to mix and match location attributes and data on the fly. In addition, as a model, it keeps getting better with each use.

In tests that Factual has run on Observation Graph, the tool has been “33 percent more accurate at determining the specific place that maps to a device generated lat/long” when compared to the traditional method of simple geofencing / point-in-polygon calculations, notes James Kung, Factual’s director of product, in a blog post.

“The Observation Graph lets you create a very specific set of criteria that illustrates an audience, and it might end up being very small scale, but because we can predict what set of behaviors infers about other sets of observations, we can then predict who may be like them, and then add people to that so you can increase your scale and find people who have behaviors and observations that correlate to the ones that you want,” Gupta said. “So those are the two main benefits from a Geopulse Audience perspective.”

About The Author
David Kaplan David Kaplan @davidakaplan

A New York City-based journalist for over 20 years, David Kaplan is managing editor of A former editor and reporter at AdExchanger, paidContent, Adweek and MediaPost.