The industry has been focused on cross-device targeting for the past couple of years, debating the value of deterministic versus probabilistic. In a cross-device world, deterministic relates to matching user IDs across devices with personally identifiable information (PII) like physical address, email address, or phone; and probabilistic associates user IDs with non-PII signals like IP address, wifi, lat/long, and time.
With the rise of mobile location for data targeting, terms like deterministic and probabilistic should be viewed in a more critical light, and the presence of first-party location dataset should be a primary focus for marketers. In other words, you should think about whether you really want to leave your location-based targeting to chance by relying on data with limited accuracy. Odds are you don't.
To put this in an understandable context, think about heading out on a road trip. You pick a driving playlist, and plug your destination into the GPS. However, if GPS relied solely on probabilistic data, then it would say, “you will probably arrive somewhere near your destination in approximately 4 to 10 hours." In that case, let's hope you picked a good playlist.
The cause of this level of uncertainty lies in the quality of probabilistic data. Without precise, deterministic data, you'll only reach your destination — or in the case of media, your target audience — on the off chance you drive by it following your probabilistic directions. If it's not accurate, and it's not effective.
A misstep with probabilistic data while driving may only cost you the price of a tank of gas, but the cost of poor data with audience targeting is significantly higher. You could miss your target audience completely.
Ad-tech vendors and marketers alike, need to be aware of the pitfalls of many location data sources. Why? Because not all data solutions are created equal.
Mobile is regarded as the media channel of the future, and media and data companies are vying to provide and monetize mobile data. With over 90% of adults owning smartphones in the U.S. alone and double-digit growth in mobile ad spend, there's a glut of mobile data which can be analyzed and segmented into signal versus noise. Solutions relying solely on aggregated ad call data (“third-party") for probabilistic segmentation and attribution have a woefully limited view into the behaviors of the vast array of targetable smartphone users.
If your brand's focus is reaching its target audience based on actual real world behavior, then you'll need a more precise, deterministic dataset. Without it, you'll only get somewhere near the target.
So, how do you home in on this target?
Ask providers where data comes from
As a buyer, navigating the crowded ad tech landscape requires due diligence and insight into what you're buying. Until the industry holds itself to a higher standard for data quality, savvy advertisers must ask where data comes from, in order to be confident that budget isn't being wasted. Identifying the source of your partner's data can help clarify whether the location data is accurate and precise. If the source is solely third-party data from ad calls, its depth and quality will be limited.
Focus on quality of data
According to the MMA, “for location data to be meaningful, it needs to be tied to specific places…The quality of a location-based audience profile is also dependent on the quality of the underlying place's data."
A scalable first-party userbase that explicitly shares its location will not only give you an accurate set of segments to target, but will also help qualify the veracity of any third-party data layers. So, ask your current mobile partner how they measure the quality of their location data. Verifying location accuracy to a very precise level is exceptionally difficult, and modeling and analysis without a deterministic dataset is not a comprehensive solution.
Figure out how they are mapping the data to places
The key to targeting based on real world behavior is tied to the ability to verify whether a user was in an actual place of interest. Placing someone in a Walmart is one thing, but discerning between unique places on a dense urban block or in a mall is much more difficult. The map-based view can only get you a tile, radius, or outline view of a venue; however, seeing the world through the lens of a smartphone is the only way to deterministically know which venue a user has visited.
Targeting based on location is not easy to do. With spend increasingly allocated to mobile, many companies are trying to get a piece of the location pie. Asking the right questions and identifying whether or not data is deterministic and reliable will mean the difference between success and failure.