We Onboarded Our First Beta Users — Here's What Broke

Last week I sent out the first 50 beta invites for the app. I've been building this thing live on stream for months — every architecture decision, every refactor, every late-night debugging session. And now, for the first time, real people were using it.

Here's what happened.

1. The onboarding flow.

I built the onboarding flow in about four hours during a stream. It worked perfectly in testing. With real users, three people got stuck on step two and never completed setup. The issue? I assumed everyone would tap "Allow" on the notification permission prompt. About half of them dismissed it, which silently broke the rest of the flow downstream.

Fix: make notification permissions optional at onboarding. Prompt again contextually when it actually matters.

2. Image uploads on Android.

I tested on iOS. I own an iPhone. Classic mistake. On Android, the image picker was returning a URI format I wasn't handling correctly, which caused the profile photo upload to fail silently. Users thought it worked. It didn't.

Fix: two lines of code once I knew what to look for. The debugging took three hours.

3. The feed algorithm showed new users an empty state.

New accounts have no connections, so the "For You" feed was empty. I knew this was a problem in theory. Seeing a real person stare at a blank screen made it visceral. Empty states kill apps.

Fix: seeded new accounts with a curated starter feed of builder content. Not perfect, but it's something to scroll.

The "Trending" tab. I spent a full stream building a trending algorithm. Not a single beta user tapped it. Not one. I'm leaving it in for now but deprioritizing any further work on it until I see organic demand.

Saved posts. Same story. Built it because I thought people would want it. Nobody used it in the first week. Might be a discovery problem, might be a "not valuable enough" problem. Watching.

The DMs.

I didn't build a DM feature. There's no messaging in the app. But three separate beta users asked for it within the first 48 hours — unprompted, in the feedback form. That's a signal I'm taking seriously.

I'm not building it yet. But it's moved up the list.

The next beta wave goes out in two weeks — another 100 users. Before then I'm fixing the onboarding flow, patching the Android image bug, and improving the empty state experience.

The app is getting better every week. That's the whole point of building in public. For the full picture of where things stand heading into launch — what's done, what's behind, what's getting cut — 30 Days to Launch is the current state of play.

If you want early access, join the waitlist on the app page.