Retention and expansion
Retention tactics
How apps kept users and subscribers: notifications, habits, win-backs and cancel flows. 94 tactics from 49 founder interviews.
Most founders worked on retention inside the product: they found the one action users repeat, such as a daily question, a timer or three workouts a week, and built streaks, widgets, reminders and progress records around it. Few measured the effect; most linked a feature to retention by correlation. The biggest reported gains came from rebuilding the core habit, while a price cut, a chat feature and some push wins failed to hold users.
- Flamme put its daily couples question at the center of the home screen, with a morning reminder and an evening nudge about the partner missing out, and reported day-30 retention rising from 4.2% to about 50% alongside other changes.Flamme (1)Flamme (2)
- Zumba placed users in goal-based programs and said joiners watched twice as many videos with month-2 retention about doubled, though users chose to join; Ladder said 70-75% of users log every workout, which it saw track with retention.ZumbaLadder
- Symmetry A/B tested a closing onboarding screen where users pledge to work out and reported about 3% better retention; Coconote's offer of 7 more trial days kept about 27% of people in its cancel flow.SymmetryCoconote
- Opal called its focus streak, shown as an animated flame with gemstone rewards, its biggest retention win of the year; Jungle said a tree that grows as students answer questions raised questions answered by about 70%.OpalJungle
- Shared plans came up often: Simply said multi-profile family accounts retain best in every cohort, and Built With Science reported 15% of trial starts adding a gym buddy at 15% off each, with buddy users retaining better.SimplyBuilt With Science
- Some efforts did not hold users: Zumba turned off chat threads nobody used, a B2B price cut for churning customers kept them maybe one more month, and Duolingo saw early push notification wins turn negative after about six weeks.ZumbaA B2B subscription startupDuolingo
Summary written 2026-09-26 from the tactics below. Figures are as founders stated them.
What worked, with the best evidence
What didn't work, or only partly
- Flamme: Relaunch the date planner as an AI relationship coach right after ChatGPT came outDay-30 retention rose from 0.8% to 4.2%, still too low; press but no traction
- WiseSam: Give new users one-tap task buttons so the AI assistant learns about them fastUsers get more engaged once the agent knows them; still needs more guidance
- Duolingo: Sell ads directly to brands and have app characters introduce themUsers like the format; slow to sell because it's custom
- Burner: Phase out one-off credits so everyone goes through a subscriptionFairly high churn from short-term users; founder now thinks credits might have fit them better
- Zumba: Add chat threads where users talk about classes and their health journeyUsers didn't engage; feature turned off
- Life360: Chase every possible battery or location bug, even if most are false alarms80-90% of chases are false alarms; old battery bugs still hurt trust
- Lower the price to keep customers who are churningFailed; maybe one extra month retained if they don't care about the product
- Reading.com: Add onboarding slides and parent guides explaining why parents must join inStill many cancellations from parents who won't take part
Companies in this topic
Benchmarks
- App retention and churn rates · 31 figures
More tactics (74)
- Life360: Give daily location sharing away free and charge for safety featuresGrew to 80M+ active users and ~$400M revenue; about 1 in 8 families pay
- Welltory: Keep basic and possibly life-saving health tracking free; charge for deeper analysisOver 60% of conversions happen after day one, over the next couple of months
- Lobby: Add a group photo countdown to video calls that people share to StoriesCredited with more sharing and pulling in left-out friends; #1 in UK, Italy, Germany 3 months later
- Fastlane: Put all call notes in one AI dashboard and score which customers love youShowed signups who said 'I need marketing now' stick around better than curious ones
- Flamme: Hide your partner's answer until you answer the question yourselfCreated a back-and-forth loop credited with lifting retention; no separate number
- Ladder: Let users log reps and weights, then suggest next weights and show their progress70-75% of users now log every workout; guest sees it correlate with retention
- Duolingo: Charge $1-3 to fix a broken streakA good share of users paid; showed how much users value their streak
- Duolingo: Run 3-6 month holdouts for changes whose effects show up lateNew push lifts can turn negative after about 6 weeks
- Subway Surfers: Swap characters with other hit games, with no money changing handsBrings back lapsed players and boosts both games; Crossy Road got the bigger bump
- LinkedIn: Give subscribers free months of partner apps like Notion, Spotify and HeadspacePerk users tend to stay longer; no numbers given
- Bloom: Give $2-4 of free stock at signup; let users earn up to $75 moreFounder says free stock beat cash for building a long-term mindset; no numbers
- Zumba: After each class, celebrate, ask how hard it was, and suggest the next classMore people come back for their next class; no numbers given
- Built With Science: Build an AI coach trained on the creator's videos that reads each user's own dataOne of customers' favorite features, founder says; no numbers
- Visible: Test new features with an early-access group, then explain changes with resultsKept a failed feature away from most users; eases resistance to change
- Cat on Chair: Turn a plain Pomodoro timer into a cat game with collectible furniture and catsOver $18K in the last 30 days; 127K downloads in a crowded category
- BlueThrone: Buy apps that get lots of downloads but lose users, then fix the productA couple of such deals brought outsized returns for founder and buyer, per the guest
- Simply: Offer a family plan with up to 5 profiles that covers every appMulti-app, multi-profile accounts retain best in every cohort
- Life360: Sell trackers at low margins to pull people into the app and keep them longerTile owners retain better than non-owners; Pet GPS requires a subscription
- Cal AI: Let users post progress photos, meals and streak badges to Instagram storiesA lot of users share progress and meals; no numbers
- Opal: Add focus streaks with an animated flame and gemstone rewards at milestonesBiggest win of the year for retention and growth; no numbers
- Opal: Let users tap to crack open a gemstone when they hit a focus milestoneThe app's most loved moment; users remember it and talk about it
- Runna: Target serious runners with training plans that adapt after every runWon the race-training niche; later bought by Strava
- Wispr Flow: Put a small always-visible widget or calendar reminder where users already workGuest now uses voice nearly as much as typing; no company numbers
- Notion: Get early-adopter creators to post tutorials and templates for your new AI featureDrove organic adoption of Notion AI and better new-user retention; no numbers
- Flamme: Add a home-screen widget showing distance to your partner and days togetherWorked well, per the founder; copied from a competitor; no numbers
- Ladder: Read your five-star reviews to find the result users want, then target 3 workouts a weekBecame the goal the whole product is built around; no retention numbers given
- Ladder: Add a home-screen calendar widget that lights up workout days, and show how to install itAbout a third of users have it; guest sees it correlate with retention
- Ladder: Add nutrition tracking to a workout app after users asked for it in the surveyEarly read: users doing both convert and stay much more; no numbers yet
- Duolingo: Let people pick which day you remind them before their free trial endsLifted conversion; no numbers given
- Subway Surfers: Move the game to a new city every three weeksGuest credits it with keeping the game fresh for 13 years; no numbers
- LinkedIn: Keep adding paid features, with a team as big as the growth team20+ new benefits in two years; guest credits it for acquisition and retention growth
- LinkedIn: Offer a two-person plan at 20% off, plus a big annual-plan discountBoth offers retain better; no retention figures given
- NGL: Have every user post their link, so friends who quit see it and returnChurn offset by a very high redownload rate; still popular 3 years on
- Apple: Stop charging $129 for OS upgrades so everyone updates automaticallyEnabled automatic overnight updates; profits at record highs anyway
- Tolan: Improve conversations and add a shared planet until D7-D15 retention risesD5, D7 and D15 retention went up; growth then started paying off
- Sunflower: Build everything around the one simple thing users do every dayTimer is the most important feature; AI chat alone had poor retention
- Lose It!: Keep the free version good enough to reach the goal without payingRetention very stable after doubling the price, per guest
- Skylight: Don't mark which features are premium once someone has subscribedRenewal rates high even for light users of paid features; no numbers
- Tinder: Build engagement and retention first; push monetization only years laterReached 30-40M monthly users before revenue became a focus in years 4-5
- Burner: When someone deletes their number, offer them a new oneRetains some users into a second number; 20-30% of new subs are returning users
- FinalBoss Sour: Pack visible value into product photos and release limited flavor dropsGuest says it made organic storytelling much easier; no numbers
- Copy a rival's strength score and add game-style ranks for each musclePerformed 'brilliantly' per the guest; Stronger later added ranks too
- LinkedIn: Tell subscribers what their benefits are and how to use themMessaging changes alone improved retention; no numbers given
- Ladder: Let other users cheer you on live while you work outGetting cheered correlated with higher retention; exact numbers not recalled
- Uber Eats: Sell a monthly membership mainly to stop customers switching to rivalsGuest says members are way less likely to use a rival; no numbers
- Condé Nast: Bring readers together in person and use the app to connect event to eventGuest describes a positive cycle; no numbers
- Cat on Chair: Have the cat leave a gift after a finished session, and trash if you quit
- Life360: On day 2, nudge people to an unused feature using a family member's nameNo numbers; first-week activation strongly predicts users still around a year later
- Super Unlimited: Walk into elevators with a debug build to test dropped connections
- Duolingo: Give the free 3 days as a reward right after a finished lessonFelt more rewarding than on the learning path; no numbers
- Lobby: Invite friends to a 1-minute video catch-up at a random minute each night
- Jungle: Split onboarding into three short steps spread through the first session
- Zumba: Push new users to finish 3 classes: default weekly goal of 3 and a counterUsers who reach 3 classes are much more likely to stay
- Zumba: Send workout reminders at the day and hour each user usually works out
- Life360: Write down which free features can never be taken awayToo new to judge; founder says ask again in a year
- Add a $20/year tier using free Apple weather data next to the $40 plan
- Symmetry: Add ranks, streaks and a friends feed to workout trackingRanks are the most-liked feature
- Cat on Chair: Let users pet the cat: it meows when tapped or when you say its name
- Go Tall: Let the height prediction change weekly as people log sleep, meals and stretches
- Flame: Ask couples one question a day; show a partner's answer only after you answer
- Life360: Add pet and aging-parent features so families keep using the app after kids leave
- Cal AI: Let users join friend groups and topic groups with streak leaderboards
- Ladder: Give a check mark per workout and a weekly streak for 3 of themGuest says it reinforces the 3-a-week habit; no numbers
- Lose It!: Raise the price only for new users; keep current subscribers on their old priceAvoided an untestable churn risk; no numbers on the effect
- Split renewal rates by plan, discount, and web vs app before reading themWeb buyers renew much better, so blended renewals rise from mix shift alone
- Duolingo: Block gambling and sketchy ad types, even if it costs revenueAccepted lost ad revenue for better experience; no numbers
- Jungle: Show the paywall right after the first quick review, but let people skip
- Tractive: Ping owners when their dog hasn't been walked in a dayGuest, a user, calls it phenomenal; no business numbers
- Wired: Make editor Q&A sessions subscriber-only and add them to the subscription
- Clear 30: Sign users up by phone number and text them funny, personal messagesNo results given; founders cite ~95% of texts read within 5 minutes
- Harpagia: Focus on keeping players coming back before worrying about making moneyFounder credits it; no separate numbers
- Cal AI: Redesign the meal logging screens so people keep paying after the 3-day trial
- Cal AI: Add daily logging streaks, copied from Duolingo
- Reading.com: Extend lessons to third-grade level and plan more apps for later reading stages
New tactics every Monday
The week's founder interviews, boiled down to what they did and what happened.