The onboarding mistake we fixed twice

The first version of BoardSnap's onboarding was a three-screen tour. Illustrations, feature explanations, a "Get Started" button. Standard stuff.

Activation rate (percentage of users who snapped their first real board within 24 hours): 34%.

I knew this was bad. I rebuilt it.

### Version 2: Interactive onboarding

The second version replaced the static tour with an interactive flow. Users set up their first Project during onboarding — give it a name, optionally add a URL. The flow ended with a prompt to snap their first board.

Activation rate: 41%.

Better. Still not good enough. More importantly, the delta between users who completed onboarding and users who snapped a real board was still large. People were finishing the setup flow and then... not snapping.

### What version 1 and version 2 had in common

Both versions were trying to explain BoardSnap before letting users experience it. They were optimized for understanding, not for experiencing.

The mental model was: if users understand the value, they'll be motivated to snap. This was wrong.

Understanding doesn't create motivation. Experience creates motivation. Nobody becomes a photographer by reading about photography. You hand them a camera.

### Version 3: Demo board first

The third version skips the explanation almost entirely. New users open BoardSnap and immediately see a pre-populated Project with a demo board — a real-looking whiteboard summary from a realistic sprint planning session, with action items in various states, a plausible summary, a set of pinned context notes.

The first thing you do in BoardSnap is interact with a board that already exists. You can read the summary, tap action items to change their state, chat with the demo board. The app is already doing its thing before you've read a word of copy.

Then, after you've touched the demo content, there's one prompt: "Snap your first real board."

Activation rate: 67%.

That's a 33-point improvement from version 2 and a 33-point improvement from version 1. The mechanism: when users have already experienced what a board looks like, the instruction to snap their own board is concrete rather than abstract. They know exactly what they're going to get.

### The side effect we didn't expect

Users who went through the demo-first onboarding also used the tri-state action item model correctly at higher rates than users from previous versions. They'd already tapped items in the demo — the interaction was physically familiar before they encountered their own data.

Pre-loading the interaction pattern before the real data arrived reduced the learning curve to near zero. There was nothing to learn — they'd already done it.

### The lesson

For any tool where the primary value is experiential — you have to do something to get the value — onboarding should deliver the experience before it delivers the explanation.

Explanation is for users who are already motivated. Demo is for users who are still deciding whether to be motivated. Every new user is still deciding. Give them the experience before they decide against you.