Building your own geofence solution seems easy enough: find the point of interest you want to geofence, find the latitude and longitude point it corresponds to, identify the precise boundaries in the real world the fence corresponds to, make sure the location sensors in your app are tuned just right.. Okay, maybe not so easy. Despite the complexities involved in a project that bridges the digital and physical worlds, some may insist on taking the geofence building process in-house.
A brilliant team of developers and product visionaries may be able to make it happen, but many companies come to Skyhook to power this process. While building a geofence system yourself may seem to save you time and money, the benefits of using location experts instead have been proven time & time again.
Whether you’re going it solo or choosing a location provider, this post begins a series of articles that give a few key considerations to make around:
- Sourcing the content that makes up your geofencing campaign
- Development experience & location expertise
- Time & money
Geofence Content - 3 Things to Consider
Before you start building your database of geofences, it’s important to know how much professional handling you’ll need to make your geofencing campaign a reality.
Understand that the onus is on your team to supply all of the content needed for geofencing. What kind of content are we talking about? You need to start with a latitude and longitude point for each area of interest you’d like to geofence. You also need a radius that defines the size of the area you’re monitoring. If your fence isn’t captured by a circular area, some finagling of the code and advanced algorithms needs to be in place in order to setup polygonal geofences.
Where can you get all of this venue and geographic information? Hopefully there’s an internal database of current chains or store locations that you can pull from easily. The alternative would involve licensing store locations from 3rd party databases who specialize in keeping business addresses up to date.
How many geofences do you need? For example, are you interested in triggering an experience at a particular venue in Austin TX, just one chain of stores, or multiple different store chains across different locations? This could be a handful, a few hundred or thousands.
If you’re doing a small-scale campaign - something that requires up to 100 geofences in one contained area - an in-house build may get you by. If you’re a larger chain or want to monitor multiple stores across the country, you’re going to face scaling problems. Keep in mind OS limitations for the number of geofences you can create.
You need a professional provider if you’re monitoring over 100 geofences, which will require your app to dynamically toggle nearby venues on and off for each individual user as they move.
For retailers of well-known brands, it can seem an obvious choice to bring builds in-house - you have all of your store addresses, right? But keeping all of that data updated in real time can be a challenge.
For example, one brand chain looking to run a national campaign came to Skyhook with a complete list of their store addresses. When we checked this list against our nationwide database, we found that over 50% of the lat/longs listed were incorrectly positioned.
Why does this occur? Keeping up the exact location coordinates for venues is strenuous work and previously needed only for things like mailing addresses or general navigation. Companies may have the correct addresses listed, but the lat/long points are imprecise and can be within a 100 meter radius away from where the actual venue is. Oftentimes, this inaccurate dot on the map can seem to be “good enough.”Our Context product pulls from the Wi-Fi signals of store MAC addresses, and uses algorithms tuned with machine learning to identify the proper venue those signals belong to. The result is a very precise location reading of your store boundaries. This means that your stores can speak the language of the device it’s trying to connect with, and therefore, your mobile users.