Recently, I tried some more location-aware stuff on my iPhone: life-logging apps, context sensitive notifications, automatic logging of when I get to work, that sort of thing. I love the idea in theory, but implementation is rough in practice. I think I know why. Part of it is the inexactitude of cell phone GPS technology. Another is that many of the companies and individuals building location aware apps live in places like Silicon Valley and other places with a lot less human density. Most cities in the world, let alone the United States, aren’t this dense.

I live in New York City. I work in Midtown Manhattan. Measured on Google Maps, my morning coffee shop is within a 700 foot radius from my office. That sounds like a lot, but even the smallest geofence I can set up with IFTTT’s iOS location channel is large enough that it triggered a reminder to log my coffee intake while I sat at my desk. Three times. So, I turned that off. Looking into it, the smallest geofence range I can create in the iOS Reminders app is a radius of 328 feet. (IFTTT’s smallest range appears to be larger.) Combine that with the quirks of GPS data, and no wonder I’m seeing my Starbucks loyalty card on my lock screen when I sit at my desk. The geofence is just not narrow enough.

How much of this is a hardware limitation, and how much of this my environment’s density being a special case, I’m not sure. For what it’s worth, I’ve been a geolocation edge case from the moment I moved to an apartment over a coffee shop in West Philadelphia. (Am I home? Am I getting coffee? Is there even much of a difference?) But, if we’re working towards a context-aware world, with our little GPS-laden phones at the center, it behooves the people making the technology to figure out how to pin someone’s location down better. Until then, over eight million people will be left out of the revolution. All because their phones don’t know if they’re at the office, or at the bar across the street.

App developers working in this space, or any space that relies on making generalizations about human behavior, need to think a little more about their potential audience. Not everyone drives a car, not everyone has a commute above the surface of the earth, and not everyone gets their coffee more than 1000 feet from their desk. For location-aware apps and services, being able to identify small differentiations means the difference between an app you can use, and an app that just frustrates. It’s early days, and these are early adopter blues. Better to point the issues out now, before the normals get on board.