GroundTruth has developed the only patented location verification algorithm in the industry. In order to verify every location signal, we score and categorize every signal and visit based on questions such as:
- Was the location signal sourced from GPS?
- Is the signal derived or stale?
- Is it an outlier signal?
- Is it a fraudulent or suspect device?
- Does the signal show within one of our Blueprinted polygons
- Does a visit occur during open store hours
- Does a visit fit typical employee behavior
- Does the speed of the device indicate a drive by?
We filter for the following inaccuracies:
- Centroids: Centroids are lat/long coordinates that correspond to the exact center of a broad geographical location including countries, states, DMAs, cities and zips. Publishers may sometimes pass centroids when they are uncertain where a device is truly located. For example, if the publisher cannot define exactly where a user is but can say the device is somewhere within the US, the passed lat/long will default to the US centroid location which is in Kansas. Wherever GroundTruth recognizes a centroid we remove that ad request.
- Fraudulent signals/ block lists: Through an integration with a leading global intelligence platform and fraud protection provider, GroundTruth is able to monitor its entire network and identify both general and sophisticated invalid traffic. In addition, GroundTruth has automated and manual processes to measure and root out problematic supply to ensure clients have access to quality supply.
- Randomized lat/long: Sometimes a publisher may generate a randomized lat/long in order to pass any location information because they can receive more money for an ad request that has location data attached. However, randomized lat/longs may occur in patterns, which can be recognized by GroundTruth’s system and subsequently removed.
- Carrier IP signals: Carrier IP signals can be inaccurate because often the location coming from this signal is attached to the address of the carrier’s data server despite the consumer’s true location. GroundTruth does not accept location information coming from Carrier IPs.
- Outlier signals from the same device: Does the same device appear in two distinct locations within a short timeframe
These are just a handful of examples of the things GroundTruth’s location verification looks for when analyzing each piece of location data. This level of analysis of every location signal allows GroundTruth to be very selective in the signals we use for location-based activities because of the scale we’re able to accomplish. Without enough scale, vendors are forced to use less accurate and precise location signals to make up for their lack of scale.