Faves
Faves is an app for saving and organizing the stuff you find on social media. I joined to work on the app, and today I also help the team figure out and build new products.
- Role
- Builder & designer
- Period
- Jun 2025 — now
- Status
- Ongoing
- Built with
- Figma, Claude Code, Notion
The app
I started by auditing the whole app, then redesigned the Pro plan, the onboarding and the paywalls, and gave it a more premium look.
On the paywall, people get a free week, we tell them clearly they won’t be charged today, and there’s an easy way to go with Plus or Free instead. The audit also flagged the forced trial, and it was removed. I designed it in Figma, and the team’s engineer built it.
How we test ideas before building them
Faves also works as a small venture studio now. With AI, building got fast and cheap, so the hard part is knowing what’s worth building. I put together the process we use to check that with real people first. Every week we take one idea through it.
- MonWrite the idea downHalf a page, six questions. If it doesn’t fit, the idea isn’t clear enough yet.
- Tue — ThuTalk to people who have the problemCalls, messages, fake door tests, and a prototype when talking isn’t enough.
- FriDecide: go, wait or killWe write it down with the evidence, so everyone can see why.
- Who
- One specific kind of customer
- Hair-on-fire pain
- What hurts badly enough to act on this week
- Current way
- How they cope today, and what it costs
- New way
- What we’d put in their hands instead
- Why 10×
- Why it’s not just a little better
- Money
- Who pays, how much, and why now
Then we try hard to break it: the market, the angle, the risks, every reason it could die. It gets a score from 0 to 100, and under 55 we don’t build it.
That test also found a rule in a partner platform’s terms that would have broken the product as planned.
Every test gets a kill date on the calendar. If we don’t run it by then, the idea is dead.
Don’t ask people to confirm a problem that makes them look bad at their job. They won’t.
We’ve written down and stress-tested more than 20 ideas. In the first 46 days we killed 8 of them, and it was the first time the team killed an idea because of what customers told us instead of after building it. Before we had this process, our quickest build went from prototype to people paying in 7 weeks, and then had a hard time keeping them. That’s a big part of why we now test before we build.