Building in isolation is dangerous. You start to believe the thing works because you’ve been staring at it for a month and it seems fine. You forget that nobody else has the context you have. Nobody else knows where to tap, what the options mean, or why the results are good.
So on a Saturday night at the end of January, we sent five friends the link. No instructions. No context. Just "try this and tell us what happens."
Here’s what happened:
Friend 1 finished the quiz in 22 seconds. Got matched to a cocktail bar in the West Village. Went there. Texted us at 11pm: "This place is perfect. How did it know?" We rode that high for approximately 45 minutes before the next text came in.
Friend 2 got confused on question 3. The group size options weren’t clear enough — she didn’t know if "small group" meant 3 or 6. She guessed, got a mediocre match, and told us the whole thing felt random. She was right. The question was bad.
Friend 3 finished the quiz and got matched to a bar that had closed two weeks ago. He showed up, the lights were off, and he texted us a photo of a locked door. That one hurt.
Friend 4 loved the quiz experience but said the results page was confusing. Too much information. She couldn’t figure out which venue was the "top" recommendation vs. alternatives. The hierarchy wasn’t clear.
Friend 5 tried it on his older Android phone and the animations were janky. Transitions stuttered. The quiz felt slow. He said it felt "cheap." Performance on lower-end devices was clearly not tested enough.
So: one perfect match. One confusion. One closed venue. One UX failure. One performance problem.
Honestly? We were devastated for about an hour. Then we realized: this is exactly why you test. Every one of these problems is fixable. And we wouldn’t have found any of them on our own.
The fixes started that night:
Rewrote the group size question with specific numbers. Built a closed-venue detection system that checks status weekly. Redesigned the results page with clear hierarchy — one featured match, alternatives below. Optimized animations for low-powered devices, reduced JavaScript bundle by 30%.
By Monday, every issue was fixed. And the next Saturday, we sent it to 10 more people. The results were dramatically better. Not perfect — never perfect — but better.
Building a product is a series of embarrassments. You show someone your baby and they tell you it’s ugly. And then you make it less ugly. And then you show someone else and they find a different ugly part. Repeat forever.
But Friend 1’s text — "this place is perfect, how did it know?" — that’s what keeps you going. The engine works. It just needs to work more often.
— The Moodap™ Team