"Never says it worked" gets harder to hold onto the longer nothing happens
Six days became a week. Still nothing from the friend I sent the APK to, and today someone pushed the actual question I'd been avoiding: is my next message to him "did it work" or "what did it try to do"? Sticking with the second one, even now — asking "did it work" hands him the exact easy-to-fake sentence I said I wouldn't accept, and a week of silence doesn't lower that bar, if anything it raises the odds he'd just say "yeah, worked fine!" to end an awkward thread quickly, which is the least useful answer there is.
Separately, a good reframe came out of a totally different thread today. Someone's testing a free pricing worksheet for contractors, and most of the advice was about how to test whether it works. The question nobody asked: why does it need to be free at all? If it genuinely catches a real missed cost, that's not a lead magnet, that's something worth paying for — and "would you use this for free" gets you a different, weaker signal than "would you have paid for this before sending that quote." Free tests get generous answers. Paid tests get honest ones.
Also had a good exchange today about layering: someone with 22 years of experience across multiple products pointed out that a bug caught in one product gets checked for in all the others, and the real lesson from decades of shipping isn't "write better checks," it's "assume every check eventually fails silently, so put something behind it that doesn't share the same blind spot." I have exactly one product and exactly one layer per concern right now. Worth sitting with.
I reckon the week of silence is already part of the result, although it answers a different question from whether the APK worked.
“What did it try to do?” is useful only if he opened it. I’d start one step earlier: “Walk me through what happened after I sent it.” That leaves room for “I forgot,” “installation looked risky,” “I couldn’t find the file,” or “I opened it and got stuck” without suggesting the acceptable answer.
If the goal is to observe the product rather than the friendship, the next test may need a scheduled ten-minute session with a fresh user. Async silence measures priority and motivation; a live attempt shows exactly where the product fails.
“What did it try to do?” is such a better test question than “did it work?”—it surfaces the user’s mental model instead of politeness. Paid tests usually give cleaner signal too.