September 16th, 2010
A very interesting study on the rise of “Apps Culture” has just been published by the Pew Internet & American Life Project. My day job doesn’t afford me the time to read more than the synopsis and a few choice bits further down the page, but there are some thought-provoking data points here. For instance:
- 23% of Americans now live in a household that has a cell phone and no landline (that goes for 100% of my own daughters);
- 35% of Americans have cellphones with apps (although only two thirds of those use the apps);
- One in ten Americans has downloaded an app in the last week, and one in five Americans under 30 download an app every week.
It’s not that it really comes as a surprise — the long-term decline of subscriber access line counts in the wireline business, the unloading by Verizon of their sparser market territories to Frontier and Fairpoint, the insane valuations that get applied to mobile social media startups all are indicators of the shifting center of gravity.
The implications that are interesting to me are, of course, how this transformation — which is still in its early stages — impacts the way networks are operated, and the software infrastructure needed to support that. Two things come to mind:
- backhaul capacity, driven both by the overall volume of downloads and by the certain emergence of very bursty, event-driven download storms, is going to continue to grow at extraordinary rates, and the tools for commissioning, upgrading, and measuring results will need to get much better, much faster.
- if the richest mine of emerging value is in the app, and if app population in end devices (both traditional cellphones and new target devices) is growing exponentially, then tools for managing apps (in all the ways they must be managed) need to be invented, improved and scaled. Your smartphone has apps. It’s going to have hundreds more. But so is your car. Your thermostat. Your blender. Can someone help me keep track of all of that? Upgrade it all? Know when it’s broken or compromised or dead-ended?