Onboarding is the only feature every user ships with
Why the least glamorous phase kept winning the prioritisation argument at Noise.
Every roadmap conversation I've ever been in has the same shape. There's a headline feature everyone can picture in the keynote: the social layer, the challenges, the thing that demos beautifully. And there's onboarding, which demos as a sequence of forms. Guess which one gets the sprint.
At Noise I made the opposite argument, out loud, more than once: onboarding first. Not because it was broken in an obvious way (it worked, in the sense that a determined user could get through it) but because of a property no other screen in the product has. Onboarding is the only feature with a 100% exposure rate. Every single person who will ever use your product meets it, and they meet it at the exact moment they care least, know least, and are most willing to leave.
Failure here is invisible by design
The trap is that onboarding failure doesn't look like failure. A crash produces a stack trace. A broken checkout produces support tickets. An onboarding that quietly leaks new users produces… a slightly disappointing acquisition report three weeks later, which everyone attributes to marketing.
In a demo, this is even worse. The person giving the demo has done the flow fifty times. They fly through it. The room nods. Nobody in that room has ever experienced the product the way a real user does, holding a new watch in one hand, the app in the other, wondering why pairing has a spinner that never resolves. Abandonment in onboarding is invisible in a demo and expensive in production, and that asymmetry is exactly why it loses prioritisation arguments it should win.
The argument at Noise
The visible ask from stakeholders was the engagement features: buddies, challenges, the social loop. Those existed, and they underperformed. My argument was that they underperformed partly because the users who would have loved them never got settled into the product at all.
So the rebuild put the foundation first, onboarding early, and the engagement features after. The trade-off was real and uncomfortable: a stretch where, from the outside, the roadmap looked like it wasn't moving. No new screens to show. That is a genuinely hard thing to defend in a review meeting, and I understand every PM who has folded under it.
What shipped: the long single-pass form became guided, smaller steps with visible progress. Device pairing (the first moment hardware meets software, where a dead end reads as a broken product) became one combined screen with a recovery path and a QR fallback instead of two disconnected screens and a prayer.
The takeaway
When you're fighting for priority, stop ranking features by how exciting they are in a demo and rank them by exposure × invisibility of failure. Onboarding sits at the top of that list in almost every product, almost every time. It is the only feature every user ships with. The rest of your roadmap is optional to them.