"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Jonathan Lenaghan, head of data science at PlaceIQ.
With mobile devices generating tens of billions of location-enriched ad requests every month in the United States, marketers have long experimented with the right ways to reach these in-market, on-the-go consumers with relevant advertising.
Location targeting, using DMAs and then geofences, was the first generation of geotargeting for mobile ads. Many companies simply took the vicinity of a device to a retailer as a very rough proxy for intent.
This sort of targeting still remains common in the industry, but improvements in the location technologies on smartphones have made possible the targeting of devices to within footsteps of large venues, such as big-box retailers and auto dealerships. This tighter targeting is a tremendous step beyond simple geofences and represents the higher end of static location targeting. The ability to locate a shopper inside a retail store, rather than simply nearby, comes much closer to inferring consumer behavior.
Tight location targeting is undoubtedly an extremely powerful proxy for consumer intent. But if it is used statically, it can suffer from both scale and richness. Here’s why.
1. The number of targeting opportunities is hampered because campaigns are limited to devices within a geographic boundary.
2. The inferred targeting persona, while highly targeted, is one-dimensional. Being in a big-box retailer does imply a likely big-box shopping behavior, but no other information about other behaviors or venue sightings is used.
Location retargeting solves both of these problems.
The Promise of Location Targeting
What exactly do we mean by location retargeting? Publishers’ ad requests will in many cases send obfuscated but persistent and privacy-friendly device identifiers, as well as latitude and longitude coordinates. These location histories can then become building blocks for very rich audiences, and yield very high-fidelity audience segments. With the right tools and strategies, location histories can enable a deep, intelligent analysis of the consumer journey, leading to smarter and more relevant and meaningful marketing throughout the path to purchase.
For instance, being able to distinguish between the retail loyalist who visits several times a month, the casual customer who visits occasionally, and the loyal customer of a competitor is extremely valuable. Layering this information with the visitation habits in other retail verticals creates very detailed consumer personas.
Retargeting is more difficult in mobile compared to the cookie-dominated world of desktop retargeting. Location retargeting offers retailers the chance to identify loyal customers, and later retarget them with more intelligently selected advertising. Even small amounts of historical location data are valuable for the personalization of advertising. Layering in a handful of location data points across multiple retailers and other venues paints a clearer picture for marketers of loyalties and preferences.
Utilizing location-retargeted audiences also significantly increases the number of opportunities to reach the consumer. Rather than waiting to send a message only when a consumer is inside or near a retail venue, marketers can now reach consumers at more appropriate times in their lives: at home, at work or even on the go.
Finding The Regularities
The ultimate boon for a marketer is to reach someone at just the right time in his or her consumer journey with the right targeting. Using location histories generates a much clearer understanding of this purchase funnel or consumer journey. For instance, knowing that a consumer tends to frequent venues associated with a healthy lifestyle, such as golf courses, gyms and larger grocery stores with lots of organic produce, enables marketers to target much more intelligently to that consumer throughout their week.
Location histories also allow the use of observed store visits to close the loop between receiving branded messaging and actually showing up in a store. In the absence of point-of-sale data, the best evidence of a successful conversion is a store visit. It's not perfect, but it's much better than counting clicks. It is intuitive, and data shows that recurring store visits are the best indication that a consumer will be visiting again.
Using location retargeting gives marketers the capability to segment consumers in several ways, such as regular visitors to a retailer, those who shop at both a venue and a competitor and those who shop only at a competitor.
Finding The Anomalies
Location retargeting can be used not only to find regularities in consumer behavior but also to find valuable anomalies. A device that is seen for the first time on a golf course or auto dealership lot may represent a consumer who just entered the consideration funnel and will soon be in market for a set of golf clubs or an automobile. Identifying these high-value users in these first moments of possible intent is possible from an analysis of location histories, and it goes beyond simply knowing they are on-site or nearby.
Addressing The Challenges
This type of audience targeting, of course, has complications far beyond static geotargeting. From the perspective of audience creation, there is significant technical and quantitative infrastructure required.
Crunching through billions of ad requests and merging them with other geospatial data sources is a tremendous computational task. The audience segmentation of more than 100 million devices and the act of serving ads to these audiences is a huge technological challenge. The value of creating rich audiences and segmentations, however, is worth finding the right technologies to solve these challenges.
There are also various privacy concerns. On the publisher and app side, it needs to be very clear that a user is opting into using location services. The computational infrastructure itself needs to respect users’ privacy as well by aggregating and storing only the context of a location, rather than a device’s actual coordinates. This means a device’s location history is stored as a series of scored behavioral indices, such as golfer 6.3 or business traveler 7.2, not as a series of location breadcrumbs.
Consumer information usefulness also has a finite shelf life so there should be policies for purging old profile data. Additionally, consumers should have an opt-out mechanism to ensure they are not retargeted if they do not want to be.
Finally, location retargeting does not generate a perfect consumer representation. Mobile devices do not leave location breadcrumbs evenly spaced in time. The ad request traffic can be uneven in spatial distribution and sporadic.
Location Retargeting: A Strong Supplement To Location Targeting
To be sure, there is great value in traditional geotargeting, especially when that targeting is used to build footprints rather than geofences. Targeting against location histories, however, significantly increases reach and allows for complicated, layered audiences to be built by combining many more facets of a consumer’s behavior.