When it comes to location-based based data, targeting is the low-hanging fruit. The ripe opportunities center on using location to enhance behavioral profiles and connect online ad exposure with real-world activities.
Easier said than done, however. Advertisers and agencies are interested, but the technology is still developing and there’s a fair amount of BS flying around the location data ecosystem.
Applying Location Data To Marketing
Location data is about knowing where someone goes and in what order they travel from place to place, said Gil Larsen, VP for the Americas at location data company Blis.
While marketers have long known about, say, triggering a push notification when an opted-in user walks past a specific location, more advanced use cases include creating real-world audiences on the fly or using location data for online/offline attribution.
“Getting a notification that says, ‘Welcome to Macy’s’ when you walk into a Macy’s is not all that interesting,” said Duncan McCall, CEO and co-founder of PlaceIQ.
Marketers can get a whole lot more creative than that.
Attribution: If a consumer looks at an ad on a device, that event can be used to tie store visits to ad exposure. When the device interacts with a beacon, sensor or Wi-Fi signal at a particular location, it’s possible to link the visit with a particular device ID.
“Being able to tell a brand that these or those people were seen in their various stores will impact where a brand allocates its marketing dollars,” said Neil Sweeney, CEO and founder of beacon network Freckle IoT.
Insights: Marketers can make inferences about consumer behavior, preferences and attitudes by analyzing how frequently and when people show up at a location over time.
If the same device is seen at the same grocery store three times a week over an extended period of time, the owner of that device is likely a loyal shopper. If that same device suddenly stops showing up, a marketer is going to want to know that.
And linking the spots a person visits over the course of a day or multiple weeks provides a more rounded view of their interests and priorities.
“It’s about seeing what a day in the life looks like,” said former Factual CMO Vikas Gupta. “It’s location as a data foundation to inform the overall strategy.”
Segmentation: Repeated visitation can then be used to build audience segments.
A movie theater once asked mobile location data company Verve to develop an audience of Sunday churchgoers so they could collect the device IDs of people who demonstrate a particular mindset based on the places they visit.
“The future is about visualizing the shopper as a complete individual,” said Verve CEO and founder Tom Kenney. “Where is that device? Where does that device live and work? All of these location events add up to powerful real-world audiences.”
Geofencing: A geofence is a virtual polygon that can be placed around the perimeter of a location based on latitude and longitude coordinates.
In theory, advertisers can target users within a geofence and track whether they’ve entered one. But the reality is a little more complicated.
As Placecast CEO Alistair Goodman noted, other types of data, like time of day, day of the week and weather, need to be taken into consideration before sending out messages to users within a geofence.
Is the store open? Is it peak traffic time and likely crowded? Is it raining? The attributes of a physical location and the contextual triggers around it are constantly changing and should be taken into consideration.
Push And Placed-Based Targeting: Used judiciously, notifications based on real-time location can provide value and drive foot traffic.
GasBuddy, an app that crowdsources info about cheap fuel prices, partnered with mobile offer network Koupon Media to target users with snack and drink coupons and entice them into gas station convenience stores while waiting for their tanks to fill. When an offer is redeemed, Koupon creates an anonymous unique identifier for that user in order to track what was bought and when. That information is shared back with both the brands and the retailers to help them refine the offers.
Gaming: “Pokemon Go” is a model example of using location to drum up user engagement – at least during its brief spurt of popularity. The free-to-play game combined augmented reality with the real world and required users to enable location services in order to play.
Unlike some apps that ask users to opt in to share their location without providing any value back, location is inherent to the “Pokemon Go” experience. Starbucks, Sprint, McDonald’s, 7-Eleven and others are all signed up as “Pokemon Go” brand partners.
Types And Sources Of Location Data
Location data isn’t derived based on any one signal alone, but through an amalgamation of several, said Matt Kamen, VP of engineering at Foursquare.
The three most common signals are GPS, Wi-Fi access points and tower cell triangulation, which can pinpoint a user’s approximate location by neighborhood or ZIP. Location can also be derived from the geolocation of an IP address used to connect to the internet.
GPS is considered to be the most accurate within 10 to 100 meters, provided no trees or buildings block the satellite signals. But GPS is a battery drain, which means that most apps use location services selectively.
“By and large, the OS on the phone is providing you with location heuristics, a combination of infrequent GPS data, Wi-Fi data and cell tower triangulation,” Kamen said. “But all that gives you is a deterministic point provided by the phone – a lat/long on a map. It’s at that point where things start to get differentiated between providers.”
Beacons, despite the hype, are just dumb pieces of hardware that broadcast location signals via Bluetooth over short distances.
If an app has a particular SDK integrated into it, the beacon signal allows it to register that the device was present at a location, which can be handy for attribution. It’s possible to use that same signal to trigger notifications and share content with consumers.
As a data source, beacons are extremely granular, said NinthDecimal President David Staas, noting that “beacon data can put a device in a specific shopping aisle within a grocery store.”
But because beacons are hardware, they can’t scale, said David Shim, CEO and founder of Placed. And because there is no overarching beacon map to say which beacon matches which particular location or product, there’s no easy way know which aisle or section a consumer goes to, Shim said.
“There are certain hot spots in a store that are valuable to a brand,” Shim said. “If someone walks into a Walmart because they saw an ad for an electronics product, the electronics brand doesn’t care if that person walked into Walmart and then went off to buy groceries. When we’re actually able to make this distinction, that will be something to talk about.”
Weather combined with location is starting to be used to help marketers make smarter media buys. Think targeting a consumer or optimizing campaign creative on the fly based on the temperature in a particular place or determining how rain affects campaign effectiveness.
Knowing who lives along a hurricane’s path, for example, is an interesting data point to Lowe’s or The Home Depot, which can target those people with messaging about storm windows, while an over-the-counter pharma brand could target people based on air quality or high pollen counts.
“It all ties back to the idea of location data intelligence,” said Shashi Seth, chief product officer at location-based mobile ad network xAd, which acquired WeatherBug in November. “It’s about adding other layers of data in order to understand and impact user behavior.”
Where To Actually Get Location Data
To access location data, marketers can partner with location data vendors, license data from third-party providers, connect through APIs that hook into apps with access to location data, partner directly with location-enabled apps or glean location appended to bid requests on the open exchange.
But marketers need to be smart and educate themselves before diving in and signing any contracts, said Placed’s Shim, who advises asking prospective partners a few key questions, including:
- How is your data sourced? Can you name some of the apps you work with?
- How is your location data validated? How is your visit rate validated?
- Do you have a public list of partners where your services are available?
- Do you sell media? What percentage of your revenue comes from media?
- What percentage of your employees are research and engineering vs. sales?
The Quality Problem
Location data quality varies depending on the source and is determined by the interplay between accuracy, precision and scale.
Beacons are precise and accurate, but they lack scale. Location data derived from the open exchange has scale, but is notoriously specious when it comes to accuracy.
But not everything that is accurate is also precise. Accuracy is whether the measured value is close to the true value. For example, does the latitude and longitude listed actually represent where a particular mobile device is or was? Precision, on the other hand, is a measure of the exactness produced by a particular methodology and the reproducibility of the results.
In other words, it’s completely possible to be highly precise and have very low accuracy, like if an advertiser is able to target a particular spot with great exactitude … just not the right spot.
According to Foursquare, nearly 80% of location data in the bid stream is inaccurate. According to Skyhook Wireless’ numbers, somewhere around 90% of location data appended to ad inventory is incorrect. And ThinkNear found that more than half – 54% – of all location-targeted mobile ads are off by more than half a mile.
But only a small percentage of that is fraud-related. The data is simply a mess: Coordinates are truncated or transposed, decimal points are in the wrong spot, data is mislabeled or doesn’t come with enough detail and data requests can get interrupted by bad cell signals.
There are cases, however, where publishers purposely misrepresent the accuracy of the data they’re peddling because including location data with inventory fetches a higher price at auction.
“There’s a ton of data on the open exchanges – understanding the quality of it can be very difficult for marketers,” said Foursquare President Steven Rosenblatt. “The big question is whether the collection of that data is tied to an actual GPS location at a precise time with accuracy or whether it’s a signal that’s tied to an IP address, cache location, random number strings or other methods that are unreliable.”
The industry is working to get handle on location data hygiene.
Although some consumers will share their location data if there’s a value exchange in the form of a coupon of a more functional app experience – imagine using Waze or Google Maps without location services turned on – they’re also quite wary.
“Location data seems especially precious in the age of the smartphone,” noted Pew in a study on privacy and information sharing released in January 2016.
Location data is a minefield for all parties – be they first or third – that aren’t careful about respecting consent and opt-outs.
In June 2015, for example, mobile ad network InMobi was fined by the Federal Trade Commission for tracking geolocation without permission (including children’s data, a violation of the Children’s Online Privacy Protection Act) even if a user hadn’t opted in to share data.
Although location data is not considered to be personally identifiable information for people over the age of 13, it is treated as PII in the case of children.
Unfortunately, it’s not always clear to consumers how an ad network operates or why their location data is out in the world, which makes it difficult for them to manage their preferences.
“If I’m a consumer and I install some ‘Candy Whatever’ game, I might think I have some sort of direct relationship with the game, and maybe the game is tracking me, maybe it’s not, maybe I don’t care,” privacy advocate and Abine CTO Andrew Sudbury observed to AdExchanger when the inMobi news broke. “But without realizing it, I’m giving a company like inMobi the right to track me and serve me ads that correspond to my location.”
But while it’s still a bit of a mess, the complex web of ad tech middleman relationships is “starting to get better defined,” and that’s helping keep data permissions sacred, xAd’s Seth said.
The location space is well-populated with vendors, purveyors of data and service providers, including Adsquare, Cuebiq, Freckle IoT, inMarket, Mobiquity Networks, Near (formerly Adnear), Reveal Mobile, Skyhook Wireless, The Weather Company, ThinkNear, Ubimo and Unacast.
Here’s a brief rundown of some of the better-known players in the space.
Blis uses predictive analytics to examine anonymized location data and identify patterns that form the basis of audience segments. The platform includes real-time bidding and buying functionality.
Factual is a location-based data platform that builds segments for mobile publishers, ad exchanges and ad networks.
Foursquare calls itself a location intelligence platform and maintains two consumer-facing products (the Foursquare City Guide and a check-in app called Foursquare Swarm). Foursquare’s suite also includes a media business called Pinpoint that targets audiences based on where they go, analytics and attribution products and the Pilgrim SDK, which allows third-party developers incorporate Foursquare data into their apps.
NinthDecimal is a marketing platform that provides audience targeting measurement and insights to brand marketers pegged off of mobile audience intel collected from its publisher partners. NinthDecimal also sells data and media (separately) and provides a measurement and analytics solution.
Placed is an in-store attribution company that uses a panel-based approach to measure where people go in the physical world intersected with ad exposure.
PlaceIQ is a location-based audience and insights platform that combines data about physical locations with consumer activity data, using what it knows to reach targeted audiences and derive insights based on a user’s movements.
Verve is a location-based mobile SaaS platform that creates customer segments based off of a mix of real-world location-derived attributes, offline demographics and transactional data.
XAd gathers location data from ad requests via mobile apps which it uses to build audience segments and profiles for targeting.