In a couple of alarmingly short weeks, my startup client Memoir Tree will launch a free app that will make history, one story at a time. But first, a few words about beta-testing - most of them unkind.
Mobile app beta-testing is simple, in theory. All you have to do is recruit everyone who ever Facebook-friended or LinkedIn to you to download your app, which actually means first downloading that UDID discovery app and then TestFlight and then your app. No wait: not that version, use this version that introduces known bugs after the launch screen...
There is a better way. But it's the opposite of how most techies will cop to doing it.
Don't believe your beta buddies
Mining social connections for beta-testers may help startup entrepreneurs vet true friends from acquaintances, but it doesn't actually do much to test app functionality. Any kind soul who nobly endures downloading your beta is also likely to lie to you gently: no, those 17 buttons don't make your interface look mysteriously Masonic, and yes, those last 23 crashes are surely attributable to a freak connection glitch. Compliments from friends are nice, but critiques from interested strangers are gold.
Do ignore decades of software wisdom
Seasoned software engineers scoff at mobile developers for not adhering to time-honored beta convention: 100 dedicated (read: paid) testers must soldier through 4 beta versions over 3 months, until the release version is rock-solid and/or you reach the promised land with Elvis. Never mind that by the time testing is complete, your mobile app's first-round funding is spent while some quick-thinking developer has released versions 1 through 3.2 of a competitive app. This is no time for mathematical methodologies. Test, adapt, release, repeat.
Don't hope for the best
Instead of beta-testing in real use-case scenarios, many mobile ventures just do a timid, ticklish round of debugging before lapsing into the "see what sticks" model. This involves app developers closing their eyes very tightly, picturing a world of pug puppies and rainbows and understanding users, and pressing "submit to Apple Store." But when you open your eyes and peer into iTunes, there it is: a dim galaxy of one-star reviews, collapsing into a black hole from which your app may never, ever emerge.
Do crash-test in public
To avoid being lulled into false launch-ready complacency, glitch aversion therapy helps. So after beta-testing amid the domestic comforts of friends-of-Facebook-friends and favorite office toys, we introduced Memoir Tree into the wifi-less wild. We road-tested the app among total strangers, capturing oral histories at the Asian Art Museum's Year of the Dragon celebration, the US Slow Food Festival and the San Francisco Bluegrass and Old-Time Festival. Suddenly, against external event deadlines, the debugging process became stunningly efficient. Knowing we would be scrutinized as we fumbled with our smartphones during live interviews, we also made interface refinements and processing speed improvements.
Don't act like Google
Once you put a face to an end users, they're not so easy to shrug off with the usual "it's not perfect, but at least it's free" Google gambit. "Is this thing on? How can you tell?" asked one skeptical banjo player interviewed with the Memoir Tree app. He was right: our recording icon needs tweaking. I was momentarily mortified, then recruited him as a beta-tester.
So if you need me this week, I'll be under the hood, pushing launch features I thought could wait for version 1.1. If motivation flags, I'll turn on some banjo music.