Where does that latitude/longitude come from?
In our second post of this series, “Which Data Source is Best? SDK vs. Bid Stream vs. Beacon,” we discussed the various sources of location data and how you can use each of them. We also hope we made one point very clear: GPS data is the absolute gold standard in location.
So when it comes to bid stream data, everyone wants GPS, but a lot of that lat/long information actually comes from centroids. Far less precise than in-device GPS, centroids represent any central location, like a city or a country, and they are not the most reliable source for accurate location data.
Here’s why: GPS data can only be collected when a user gives express permission via an app or mobile website to do so. When a user refuses permission for an app to use their device’s location data, in order to serve an ad, a centroid is used to estimate a location instead. A third-party IP-to-location provide will, in real time, send the user’s IP address to a partner who will see who the carrier of that phone is, then fling back a location they associate with that carrier. So, if you’re a Verizon customer, your location may be assumed to be New Jersey, since that’s where Verizon is headquartered – even though you’re in Westchester, New York.
There are many bad actors in the digital advertising space who will try to pass centroid data off as GPS. Sometimes this is simply due to a lack of education, but sometimes it’s due to greed. Data with lat/long information is extremely valuable in today’s market. However, by trying to pass off less accurate centroid data as GPS, these publishers are doing more harm than good. Selling bad data puts all data into question; bad data yields poor results, and no one will pay a premium for data that may not deliver.
And yet, nearly half of the data that comes to Blis is centroid data that publishers are attempting to pass off as GPS. We have rigorous processes in place to catch it before it ever can reach our customers, but of course, not all advertisers work with us. Advertisers working with less responsible partners may never know how poor the quality of their location data is!
To illustrate how bad the problems with centroid data can be, there was a campaign we ran at Blis years ago that involved several geo-fences in metropolitan areas across England. However, shortly after the campaign began, we realized that 95 percent of the campaign’s volume was coming from a single geo-fence – one which wasn’t even in the most densely populated areas encompassed by the campaign. Further investigation revealed that millions of bid requests were coming from one specific set of lat/long coordinates – an area less than a single square meter. The only way this could have possibly occurred would have been if thousands of smart phones were stacked on top of each other in a pile over a thousand phones high, and all requesting ads at the exact same moment. It was physically impossible.
It’s because of fraudulent data like this that we invented our SmartPin technology. We decided that rather than battling fraud in court, educating marketers and providing reliable solutions was a far more productive path.
To ensure you’re getting the best location data possible, working with Blis is the obvious solution – but we understand you may have partners you trust in place already. If you’re not confident you’re getting the best quality data, be sure to ask the following questions:
- Where do you source your data?
- Do you use bid stream data? Is it strictly GPS, or do you use centroid data, as well?
- How do you validate that the data is GPS data?
- What’s your methodology for filtering bad data?
- If your provider can’t answer these questions, they’re not good enough for you. If they can answer, ask them to prove it.
At Blis, we strive relentlessly for transparency. Saying that data is high quality isn’t good enough; we need to keep asking questions and keep digging. We will continue to do that, and to share heat maps and visual representations of our data. And we’ll keep doing what we can to educate marketers, so that you don’t have to trust blindly in your data partners. You’ll be able to ask the right questions, and know what the answers must be.
We’ll take this conversation a step further in our next blog post, which will arm you with the right questions to ask about the cleanliness of your location data.