Visible · B2C · wearable for pacing chronic illness
Launch an AI feature that explains your health data in plain language
Early-access users hated it; back to the drawing board
What they did
What happened
- Stage
- product
- Effort
- first version, early-access only
In their words
$10M ARR Without Ever Testing A Paywall — Luke Martin-Fuller, Visible
David BarnardThat's awesome. All right, the next question, what's the worst experiment or change that you made in the past year, your biggest fail?
Luke Martin-FullerSo our biggest fail would be, it's on the core product side, so actually a feature that we built. We tried to launch a product, a feature that would help you make sense of some of your data in natural language. We figured that this was a great way to get with the trend of AI and people experiencing things in natural language. And our first cut of that really did not work. We shipped, as we do with a lot of features, it went out in our early access hub and people were like, "We hate this." So that was a really good learning. And it was super naive. So we've gone back to the drawing board on that. We still got conviction that helping people make sense of their data in a kind of conversational or a natural language way is like, there's so much we can do, but our first cut was a fail.
David BarnardYeah. That's really cool though. So you only rolled it out in your early access community. And so that fail didn't actually make it to the broader community and you didn't get that kind of negative sentiment across your entire customer community.
Get tactics like this every Monday
New tactics from the week's founder interviews, each linked to where it was said.
More from this episode
- Put up a site with a 3D mock-up and a waitlist before building anything5,000 waitlist signups in about a week
- Start with a free app and 100 waitlist users in a Facebook groupFree app grew to ~50K users by word of mouth before any wearable
- Send 100 free wearables to users and iterate before charging anyoneCharged only after over a year; then $1M to $10M ARR in two years
- Let users opt in to share anonymized data with university researchersPapers in Nature and with Mount Sinai; builds trust when buyers research; no revenue number
- Buy Polar's heart-rate sensors retail, then wholesale, then co-design a custom bandReal-time data other wearables couldn't give; guest calls it a productive partnership
- Sell the band at cost (~$80) and charge $20 a month for the appLowers the barrier to buy; host thought $80 looked too cheap to be good
- Run only Meta ads with customer videos, and earn back each dollar the same day$1M to $10M ARR in two years, cash-flow positive; ~50% of customers from paid
- Recruit members to film ad videos; pay a flat cash fee per video usedMain source of Meta ad creative; run by a team of two
- Test each member video with a small ad budget; scale only the winnersOnly 1-2 videos a week ever get scaled spend
- Hire a chronic-illness creator you paid for a video as social media managerKickstarted outreach to the right creators; founder says it helped, no numbers
- Use a web quiz that tells poor-fit visitors the product isn't for themSame quiz since launch; guest says it protects retention; untested
- Sell on the web so buyers can pay with HSA or FSA moneyAdds payment options, better ad signals and no 30% app store fee
- Offer a free app without the wearable so people can try before buyingA big chunk of acquisition, all word of mouth; barely promoted
- Move the web funnel off a flaky no-code tool so you can test itGuest's biggest win of the year; test results still to come
- Test new features with an early-access group, then explain changes with resultsKept a failed feature away from most users; eases resistance to change
- Stay cash-flow positive to $7M revenue, then raise a small, quiet Series ARaised at ~$7M revenue from an aligned investor; funds data science and product
Tags: ai-feature, early-access, natural-language