Built With Science：用Quiz替代落地页

Built With Science scaled Meta ads from zero to $100,000 a month in profitable spend this year. Almost none of that traffic goes anywhere near the App Store.

Watch on YouTube

"About 90% of our traffic is actually going to our website, and we have a really long personalized quiz that we send all leads to," says Ethan Ethier, the company's Head of Growth and Operations. The quiz isn't a lead-capture gimmick — it does three jobs before a user ever sees a price. "It educates the user on the actual product, it overcomes objections, and it increases the perceived value of the product," Ethier explains. And the answers do real product work: they build the user's personalized training plan inside the app, so the funnel doubles as onboarding.

They didn't take the web-first approach on faith, either. The team ran traffic head-to-head, App Store versus quiz, and the quiz won on conversion — decisively enough that direct-to-store now only makes sense for the highest-intent users who already know what they want. One wrinkle worth noting: when they rebuilt the same quiz on Android, it converted on par with the web. It was only App Store traffic that lagged.

The influencer app that waited a year to run ads

There's a persistent myth that if you have millions of subscribers, your app markets itself. Built With Science was founded with YouTuber Jeremy Ethier — Ethan's brother — whose channel has over 7 million subscribers. The app still had to be grown the hard way.

The app launched in early January 2025. The first Meta ad didn't run until a full year later. "We wanted to double down on improving the product," Ethier says. "As long as we're constantly improving our trial-to-paid and retention, that lifts the floor for everything else." The logic: organic traffic converts at your ceiling, paid traffic converts below it, so buying traffic before the floor is fixed just buys expensive proof of the same problem. Once trial-to-paid, retention, and churn looked healthy, they scaled — and today organic converts at 35–40% trial-to-paid, with colder Meta traffic around 25%, at a premium $189/year price point.

Just as counterintuitive: Jeremy barely appears in the ads. "We're not relying on Jeremy at all for ads," Ethan says. The team wanted to solve what he calls the bus problem — everything depending on one person's YouTube channel — so the Meta program runs on B-roll, testimonials, and static creative. A channel that scales without the founder's face is a channel that actually scales.

The pricing page change that moved annual from 60% to 80%

The team's biggest win of the year didn't touch the price. Built With Science charges $189 a year or $30 a month, and the old pricing page showed both plans side by side. Roughly 60% of subscribers picked annual.

The winning variant showed only the annual plan upfront, paired with a day-by-day timeline of the free trial: day 1, you get your personalized plan; day 12, a reminder that your trial is about to end and you can cancel anytime. Monthly still exists behind a view-all-plans option. "We didn't change the price, we didn't change the positioning. It was simply a UI shift which had a massive impact for us," says Ethan. Annual take jumped to 75–80% — at nearly $190 a year, a substantial LTV shift from a layout change that mostly worked by shrinking the perceived risk of starting the trial.

The same page also produced their biggest fail: an experiment that added more context to clarify the value proposition actually hurt conversion. Their read — at the moment a user is motivated enough to be looking at prices, more information is friction, not reassurance.

A gym buddy plan that 15% of trial starts take

One of the more unusual pricing mechanics in the app came straight from user requests for a family plan. During the web quiz, users can add a "gym buddy" — a partner who joins at 15% off for both people.

Around 15% of everyone who starts a trial adds one. That lifts average order value on its own, but the quieter win is retention: gym-buddy subscribers stay on the app longer, the same dynamic family-plan apps see with their highest-retaining cohorts. And because the question lives in the quiz, the pricing page stays clean for solo users — no paywall clutter for an option most people won't take.

70% of their experiments fail. That's the system working.

"Probably 70% of the experiments that we run fail in the first go," Ethan says. "And that doesn't mean it's a failure... often we're iterating on these failures to uncover a winner."

The process behind that: be obsessed with the problem, not the solution. The team starts with funnel data in Amplitude, finds the biggest drop-off, lists every plausible problem with that step before discussing a single fix, locks in the most likely problem as a team, and only then brainstorms solutions. Everything gets documented — which is what turns failures into iterations. A no-trial toggle test (pay upfront, get 20% off) initially lost; the written record pointed to copy that framed the free trial as the losing option. A rerun that made both choices feel good is now showing a strong lift.

In the full episode, Ethan and David also get into Jeremy AI — the in-house chat coach one engineer built in a weekend that became a customer-favorite feature — quiet feature launches that keep engagement metrics honest, and how running their own in-app training studies feeds both the science and the marketing.

Guest links:

• Ethan Ethier on LinkedIn
• Built With Science
• Built with Science Careers