Geofencing Series: Part 2

BY Mike Lu

Last week’s post delved into the common factors that affect programmatic inventory reach and scale. Part Two of this series will outline an effective strategy for running a successful mobile geofencing campaign when inventory is scarce.

Mobile geofencing campaigns, especially those with tight radius requirements are fickle in terms of reach and scale but there are tactics to mitigate delivery issues in most cases. According to AppNexus, a leading DSP, about 70 to 85 percent of all mobile impressions do not contain the critical lat-long coordinates required for building a geofence. What to do in this case when there is low inventory availability in the target area?

Leveraging traditional geo-targeting on Mobile

While lat-long radius targeting provides the most precise targeting for mobile geofencing campaigns, scale oftentimes becomes an issue due to low population density or lack of location share-happy mobile users in the target area. Traditional geo-targeting on mobile can be a viable alternative when utilized properly to build a facsimile geofence. The key is the location data from the much more readily available IP address assigned to each mobile device. Once connected to a WIFI for example, a user’s general location information (none personally identifiable) such as the State, City or ZIP code can be retrieved for targeting.

Mimicking a Radius

The ZIP code will most likely be the second most granular level of geolocation information available for targeting behind lat-long coordinates. Leveraging this more readily available data point a radius can be built around a target area by utilizing neighboring ZIP codes. This effectively builds a geo-fence similar to what you would see lat-long coordinates and a radius.

To create this type of geo-fence, the first step is to identify the ZIP code of the target lat-long location. This can be done easily by inputting a set of coordinates into a mapping service such as Google Maps and triggering the on page ‘What’s Here?’ mechanism (e.g. in Chrome on a Windows machine, right click the pin icon to toggle the sub menu).

Once the ZIP code is determined, all ZIP codes within the target area can be retrieved using a tool such as ZIPCODEAPI.com. The radius can be controlled in either miles or kilometers and the tool will return a list of ZIP codes that fall within the defined radius.

One thing to keep in mind is that geographic areas that define a ZIP code are not entirely symmetrical and can vary pretty widely in terms of coverage and population density. One way to mitigate targeting beyond an advertiser’s boundary limits with this method is to start with a much more restrictive radius value when generating the surrounding ZIP code list. 30% of the original targeting radius would be a good starting point. For example, if the initial radius targeting requirement is 15 miles, retrieve ZIP codes within a 10.5-mile radius. If a campaign is not hitting budget, slowly increase the radius to include additional ZIP codes until the max boundary limit has been reached.

This method while not perfect, is the closest alternative to pure geo-fencing except with much better scale and reach. There are certainly exceptions where geofecing campaigns with lat-long coordinates have scaled perfectly fine without the need to implement a strategy such as this one.

As mobile adoption and user’s comfort level with sharing personal information such as location with devices and apps increase, there should be much greater inventory availability in the near future. Until that time comes, it is always wise to have an alternative strategy to deliver to advertiser budgets in your back pocket.