Location intelligence company Gravy Analytics is introducing what it’s calling “location data forensics.” Its purpose is to filter out unreliable, inaccurate or fraudulent location data from the bidstream.

Location data often unreliable. Mobile ad exchange bid requests including location are often significantly more valuable than those without. That results in a lot of questionable location data being passed in the system. A publisher or app may pass questionable or fake location just to satisfy the bid request. According to Gravy CEO Jeff White, between 40 and 80 percent of bidstream location data it sees is either “fraudulent or suspicious.”

White added that the scale of the problem is largely unrecognized by advertisers and even the exchanges themselves. Location data is used for a widening array of marketing and analytics purposes: audience segmentation, offline attribution, proximity marketing, internal benchmarking and competitive intelligence, among still others.

Location accuracy matters more for some scenarios than others. For example, store-visitation attribution requires more precision than proximity marketing. And if you’re buying an audience of “auto intenders,” for example, you want the people who’ve actually been on car dealer lots in the past 30 days, and so on. The reliability of the data is critical — garbage in, garbage out.