A web quiz, a one-plan pricing page and a gym buddy: how Built With Science grows subscriptions
Most subscription apps treat the App Store listing as the front door. Built With Science sends the overwhelming majority of its traffic to a long personalized quiz on its own website instead, then lets the answers shape the plan a new user sees inside the app. The approach is worth studying not because it is unusual, but because the team tested it against the obvious alternative before committing.
Why the quiz, not the store listing, is the front door
Ethan Ethier, who leads growth and operations at Built With Science, describes a funnel in which roughly nine out of ten visitors land on the company's website and are routed into a detailed quiz rather than being pushed toward a store download. The quiz is not a lead magnet that collects an email and disappears. It explains what the product does, answers the objections a skeptical buyer would raise, and raises the perceived value of the subscription before any price appears. The same answers are reused to assemble the user's training plan once they are inside the app, so the acquisition step and the onboarding step are the same conversation.
The team did not assume the web route would win. They ran store traffic and quiz traffic against each other, and the quiz converted better by a wide enough margin that direct-to-store traffic is now reserved for people who already know exactly what they want. One detail complicates the simple story: when the quiz was rebuilt for Android, it performed about as well as the web version. The gap was specific to App Store traffic, not to app stores in general.
A useful way to read that result is as a routing decision rather than a verdict on platforms. If your product needs explanation before purchase, a quiz can carry that explanation. If your buyer already knows the category and the price, a store listing may be the shorter path.
Fix the floor before you buy traffic
Built With Science launched in early January 2025 and did not run its first Meta ad until a full year later. The reasoning Ethier gives is about sequencing: organic traffic converts at the ceiling your product can currently reach, and paid traffic converts below that ceiling. Buying traffic before trial-to-paid and retention are healthy mostly buys a more expensive demonstration of the same weakness. Once those numbers looked solid, the team scaled paid acquisition from nothing to $100,000 a month in profitable spend.
The reported conversion figures at that point were 35–40% trial-to-paid from organic traffic and around 25% from colder Meta traffic, against a $189 annual price. Treat these as a dated reference from the company's own account rather than a benchmark you can assume for your app; pricing, audience and offer all move the number.
A second choice is just as instructive. The ads barely feature Jeremy Ethier, the YouTuber whose channel brought more than seven million subscribers to the brand. The team deliberately avoided building a paid channel that depended on one person's face, and instead ran B-roll, testimonials and static creative. A channel that can scale without the founder in frame is a channel that can keep scaling when the founder is unavailable.
The pricing page change that shifted annual share
The largest single win described did not involve changing the price. Built With Science charges $189 a year or $30 a month. The original pricing page displayed both options next to each other, and roughly 60% of subscribers chose annual.
The winning variant showed only the annual plan at first, alongside a day-by-day timeline of the free trial: day one delivers the personalized plan, and day twelve reminds the user that the trial is ending and can be cancelled at any time. Monthly remained available behind a view-all-plans control. Annual share rose to 75–80%. Ethier's summary is that neither the price nor the positioning changed; a layout shift did the work. The likely mechanism is that the page reduced the perceived risk of starting a trial rather than making the annual plan look cheaper.
The same page produced the team's biggest loss. An experiment that added more explanation of the value proposition reduced conversion. Their interpretation: once a user is motivated enough to be reading prices, additional information acts as friction rather than reassurance.
Use this as a checklist item rather than a template. If you test a single-plan-first layout, keep the alternative reachable, keep the trial terms visible, and measure annual share and trial starts together, because a layout that raises annual share while suppressing trial starts is not a win.
A paired plan that lives in the quiz, not the paywall
Users asked for a family plan. What shipped was a "gym buddy" option offered during the web quiz: a partner joins at 15% off for both people. Around 15% of everyone who starts a trial adds one. The immediate effect is a higher average order value. The quieter effect is retention, since subscribers who join with a partner stay longer, mirroring what family-style plans tend to do for the cohorts that retain best.
Placing the question inside the quiz also keeps the pricing page uncluttered for solo users. That is the design lesson worth borrowing: an option most people will not take can still be valuable if it is asked for at the moment it is relevant, rather than displayed to everyone at checkout.
Most experiments lose, and the record is what makes that useful
Ethier puts the first-attempt failure rate at about 70%. The team does not treat that as a verdict on the idea. Their process starts with funnel data in Amplitude, identifies the largest drop-off, lists every plausible problem with that step before discussing any fix, agrees on the most likely problem as a team, and only then brainstorms solutions. Every experiment is written down.
The written record is what separates a dead end from a rerun. A test that offered a no-trial option with 20% off initially lost. The documentation pointed to copy that framed the free trial as the worse choice. A rerun in which both paths felt attractive is now showing a strong lift.
That distinction matters when you read a failure rate. A test that loses and leaves no notes is a cost. A test that loses and leaves a specific hypothesis about why is an input to the next test.
Frequently asked questions
Does a web quiz replace the App Store listing entirely?
No. In this account the quiz handles the majority of traffic, while direct-to-store links remain for users with high intent who already know what they want. The quiz won a head-to-head conversion test against store traffic; it did not eliminate the store.
Why did annual share rise without a price change?
The variant showed the annual plan first and paired it with a day-by-day trial timeline, while monthly stayed available behind a view-all-plans control. The team attributes the shift to reduced perceived risk at the moment of starting a trial, not to a cheaper annual price.
What is a gym buddy plan?
It is a paired subscription offered during the web quiz. A partner joins at 15% off for both people, and around 15% of trial starts add one. The option sits in the quiz so the pricing page stays simple for solo subscribers.
How should I read the 35–40% and 25% trial-to-paid figures?
They are the company's reported figures from a specific period, with organic traffic at 35–40% and colder Meta traffic around 25% against a $189 annual price. They are a reference point, not a forecast for another app with a different offer and audience.