Rise

Rise started as elite athlete sleep research, then moved to consumers. Working with Northwestern teams, Rise showed coaches that sleep tracked with sprint speed, three-point percentage and bench press gains, using lab-grade devices only elite sports could afford. Around 2020 it pivoted from white-glove team coaching to a phone app for anyone, with no device needed.

Co-founder and CTO Leon Sasson once thought sleep was a waste of time until his friend Jeff pulled him into the research. Word spread between teams before the pivot. Rise has since grown to over 10 million users with a 16-person team. Leon says many people still don't connect feeling tired with needing more sleep: "That is why we're in business."

Subversive

Rename the science into words users already understand. Rise built on two well-studied systems, the homeostatic sleep drive and circadian rhythm, and renamed them sleep debt and energy schedule. The labels emerged from research and customer interviews until people could make sense of them.

Sleep debt reflects weeks of past sleep, not only last night, so one bad night after weeks of good sleep leaves you fine, while one great night after two poor weeks helps only a little. Leon notes that "homeostatic sleep drive" would make a poor product name.

Subversive▶ 7:47

Use onboarding to teach one concept, then ask for the trial. Rise asks the right questions, then gives back each person's sleep need and sleep debt before offering a free trial. In user interviews, people say they are glad they learned a concept during onboarding.

Users arrive desperate after trying supplements, tea, coffee, energy drinks or another wearable, with anything from a year of Oura data to none. Leon: "once you get that clicks for people, then you almost like earn the right to be like all right, do you want to start a free trial?"

Subversive▶ 11:34

Keep every question easy; length matters less than pace. Rise avoids questions that make people think hard and avoids open-ended text, but treats efficiency as a non-goal. More screens and questions are fine if the person can keep a steady pace.

Health, finance and complex products justify longer flows than entertainment apps, with Noom as an extreme case. Answers add "good friction" because they personalize the experience for users ranging from elite athletes to people with sleep apnea.

Subversive▶ 12:29

Users winding down in bed give a good flow 5 to 10 minutes. Most Rise users download the app on the couch or in bed, scrolling, not solving a three-minute task. That leaves headroom for a longer onboarding as long as it is entertaining.

Rise treats this as a bet on its own users, not a rule for every product. Leon: "If something is good and it's entertaining, it's worth 5 10 minutes, like for sure."

Subversive▶ 15:21

Test the screens before the paywall, not just the paywall. Moving the paywall or changing the few screens before it has often produced wins for Rise without touching the paywall design. Rise calls this the pre-paywall onboarding phase and considers it crucial.

Those screens build anticipation that something worth having sits behind the paywall. Nobody needs a sleep app, so the product has to tease why they will want it. Leon says many teams fixate on the paywall and undervalue onboarding tests.

Subversive▶ 18:28

AI makes it worth personalizing for small user segments. Segments like night owls aged 25 to 35 or night-shift workers, each 10 to 20% of users, were not worth building for before; AI now makes it practical. A larger user base also lets Rise A/B test segments once too small to measure.

Sports coaching cost thousands per athlete per year and could not scale. Leon says users "start churning if the product feels too generic."

Subversive▶ 17:18

Judge tests on day 31 and day 60 revenue, not trial starts. Trial start rate, paywall shown rate, onboarding drop-off and day one retention are fast but can mislead. Rise built predictive metrics that tie fast readouts to day 31 monetization and day 60 net revenue, now a default part of its growth stack.

A higher trial start rate can bring the wrong users, while some changes lower the first purchase rate and still win because those users stay longer.

Subversive▶ 19:49

Revisit called tests at 3 and 6 months and change decisions. Rise automatically re-checks past experiments to see how winning and losing cohorts retained. Some tests judged losers or flat, and not released, had cohorts that retained a lot better three months later.

Rise does not need publication-level rigor, so a likely winner the team likes can ship without statistical significance.

Subversive▶ 35:15

Label each test "do no harm" or "expected to win" before launch. A do no harm test ships unless metrics get worse; it covers directions the team likes or removals of clutter. An expected winner that fails to move its metric does not earn its place by default.

Deciding up front avoids rehashing whether to keep something that was supposed to move the needle and didn't.

Subversive▶ 23:00

Keep A/A tests always running to check the machinery. Always-on A/A tests catch tracking and infrastructure problems, and often show more noise than expected because the user mix shifts by day of week and time of day.

Rise says it took too long to adopt this.

Subversive▶ 25:42

Run parallel tests on separate surfaces to raise velocity. Onboarding, pre-paywall, paywall, win-back paywall and day one app open can each carry their own test. About ten tests across the stack at once is normal, but not ten on one surface.

This depends on analytics set up so tests do not interfere. Leon says ten may be pushing it.

Subversive▶ 23:58

Use a standard stats engine and define metrics cleanly. Rise uses Segment for tracking, Amplitude for product analytics and GrowthBook on the warehouse for test statistics. GrowthBook was its biggest move toward standardizing experimentation and is cheap if data is organized.

Product managers should not judge winners in spreadsheets, given pitfalls like sample ratio mismatch. Metrics like trial start rate need clean definitions because real journeys include reinstalls and returns. Leon: "never write your own statistics engine."

Subversive▶ 28:28

Decide on one or two North Star metrics per test. With 20 metrics on a test, one or two will hit significance by chance, so Rise makes decisions mainly on one or two primary metrics.

Subversive▶ 34:34

Look at what users did, not only the target metrics. A/B tools can make teams lazy about judging on a few metrics like day 30 revenue. Rise also reviews behavior, talks to users and checks for confusion; a losing test can still show most users preferred the new version.

Leon: "AB testing tools do not replace going deeper into what you actually ship."

Subversive▶ 33:33

Stop predicting winners and test instead of debating. Leon no longer forecasts whether a test will lift 5%, 10% or go negative, because clear winners fail and tiny changes win big. Since building is easier now, Rise tests instead of arguing.

The habit comes from Rise's start as a science project: hypothesis, test, measure. Leon: "you don't need to hash it out in 20 conversations."

Subversive▶ 31:31

Rise serves people without wearables in a crowded, roomy market. Most Rise users have no wearable, and the app works with or without one, including alongside Oura. Rise treats Oura and Whoop as part of a big market with room for several players.

Leon: "there's a reason why there's no one else, cuz maybe users don't want it." Rise is also named as a rival that, like Bevel, gives guidance from data. It is unclear how its phone-only sleep debt estimate compares with wearable data now that Oura and Whoop track sleep debt too.

Subversive▶ 39:10

Rise uses AI to guide its 5 to 10% with clinical problems. About 5 to 10% of users have insomnia, sleep apnea or other problems an app cannot solve. Rise uses AI and a small team to help them navigate pathways like CBT for insomnia and CPAP, including insurance, testing and providers.

Leon: "we know where the people that are tired are and they are in the Rise app."

Subversive▶ 43:43

Make AI products context-aware without forcing a chat interface. Users will expect products to understand their situation without explaining it 27 times, but most do not want to chat when they want a quick fix. Rise builds context and closes the loop while keeping a UI.

Leon: "That doesn't mean products need to be chat-based."

Subversive▶ 46:16

Rise serves people without wearables in a crowded, roomy market. Most Rise users have no wearable, and the app works with or without one, including alongside Oura. Rise treats Oura and Whoop as part of a big market with room for several players.

Leon: "there's a reason why there's no one else, cuz maybe users don't want it." Rise is also named as a rival that, like Bevel, gives guidance from data. It is unclear how its phone-only sleep debt estimate compares with wearable data now that Oura and Whoop track sleep debt too.

Sub Club

Web only worked once built from scratch, unrelated to the app. After Apple's IDFA change in late 2021, Rise rebuilt its app onboarding on web twice and failed, shelved it, then fell short again around 2024 trying to mirror the app and avoid Apple fees. It worked after hundreds of iterations on a funnel "not at all related to the app."

Rise first blamed missing health data and wearable integrations. Leon calls the first approach "pretty naive" and says the Apple fee is one small piece of the equation. In the last year the pieces began "really clicking."

Superwall▶ 27:28

On web, help people see their problem before offering a plan. Web visitors are still weighing the problem, so Rise goes deeper on helping them recognize it, such as showing that exhaustion may come from sleep, and saves solutions for the app.

Rise had to unlearn the app's focus on time to value and an aha moment in the first 30 seconds. Telling web visitors their plan and features may fail because they are not sold yet.

Superwall▶ 23:11

Web reaches older, e-commerce minded buyers app ads missed. Rise reaches 40, 50 and 60 year olds at scale on web, people who would not normally download an app to fix a problem. They start with lower intent but often high willingness to pay.

The older audience tilts toward Facebook over Instagram, though Meta and Google perform about the same and creative decides results. Leon notes this may vary by company.

Superwall▶ 37:23

A half-price first month beat a free trial on web. Rise mostly skips free trials on web and offers roughly half off the first month, and skews toward monthly and quarterly plans because web users are earlier in consideration.

Free trials attract people who cancel a minute later, which hurts the ad algorithm's ability to find buyers. Leon says this is specific to Rise; some teams find free trials work just as well on web. No conversion numbers were given.

Superwall▶ 18:18

Polish billing and cancellation before trusting any web signal. Off Apple, payments, cancellation, plan changes, refunds, support and chargebacks become the company's problem. Without easy cancellation, trial conversion looks high while reviews suffer and card systems flag the business.

High chargebacks and refunds lower the payment acceptance rate. Leon: "if not, you're just making money on people that have no clue how to cancel a subscription."

Superwall▶ 21:00

Test each creative on web and app separately for two chances. Because the audiences differ, creative that flopped on app often starts working on web and the reverse. Rise did not do this at first but now tests most creative across both.

Creative is planned around personas, hooks and value props rather than channels; web just brings a different persona, such as more women over 50.

Superwall▶ 35:28

Web gives faster creative signal and steadier overall results. Ad platforms treat web as plain e-commerce with clean data, so creative results come in much sooner, while on app they can take a week or two and platforms may see only about 50% of traffic. Run separately, web and app performance rise and fall independently.

A bad month on web can coincide with a good one on app, which Leon says makes the acquisition engine more robust to flukes.

Superwall▶ 31:00

Matching each funnel to a creative persona is the next unlock. Web lets a funnel match one persona, like a tired person wanting productive mornings or a 50-year-old woman with menopause-related sleep problems. Rise sees interesting success but not yet a breakthrough.

App campaigns cannot do this because users must download mid-flow, and it needs much more coordination between marketing and product.

Superwall▶ 31:51