How to keep a free tier that pays off

1 The bottom line

Keep the core experience free enough that users recommend it, and gate what heavy users value most. Free users are your marketing, so weakening the free tier can cost more growth than it earns.

2 Key takeaways

• Make the free tier good enough that free users recommend the app. Word of mouth only pays out when the free experience feels complete.
  • Kenneth Schlenker, Opal15:30 saw paid share of monthly users fall from about 20% to 9%. DAUs passed 1 million, and he says the payback is tenfold.
  • Cem Kansu, Duolingo36:26 moved Practice Hub and Explain My Answer from paid to free and accepted some revenue loss.
• Gate the feature your heaviest users value most. Paywalls convert best where intent is highest.
  • Kenneth Schlenker, Opal17:45 tested one to four free blocks. Three gave free users a great experience while power users still paid for more.
  • Busuu limited free users to one language and saw conversions among language switchers rise 83%.
• Gate what costs you money or adds friction, not what drives growth. The wrong cap hits users exactly when they would spread the product.
  • Claire Butler, Figma1:02:45 found that capping collaborators hurt, so they capped files instead.
  • Giordano Contestabile, Life36028:20 gates only costly features, such as roadside assistance, where each tow costs money.
• Show free users what is paid and ask again in context. People cannot want an upgrade they do not know exists.
  • Mobbin added Pro badges and free-to-paid conversions rose 35% with no other change.
  • Michal Parizek, Mojo13:47 shows existing free users a paywall at most once a week. It drove about 15% of new revenue.
• Do not weaken the free tier for a quick revenue win. Gains from squeezing free users fade while the growth damage stays.
  • Giordano Contestabile, Life36025:36 found that cutting free location history doubled subscribers from that hook. They did not ship it.
  • Cem Kansu, Duolingo43:32 says paywalling free features works for 6 months to a year, then stops.

You have to essentially have two distinct products that are great on their own.

Sean Ellis

2.1 Decide what the free tier is for

Free is a growth channel, so judge it by the users and word of mouth it brings, not only by conversion.

• Ask whether a non-paying user would recommend the app, and give away more if not Kenneth Schlenker, Opal16:01
  • A small or trial-like free tier does not create the word of mouth that makes freemium pay. Kenneth Schlenker, Opal16:01
  • Free users tell friends whether or not they pay. Word of mouth is the majority of Duolingo's growth engine. Cem Kansu, Duolingo36:26
  • Life360 is not useful solo, and the top discovery channel is parents telling other parents. Giordano Contestabile, Life36023:32
  • Free access brought in students, now about two thirds of Opal's DAUs. They pushed the app to peers and parents, then asked their schools to adopt it. Kenneth Schlenker, Opal13:50
• Accept a low conversion rate when volume makes up for it Tanuj Chatterjee, Super Unlimited14:59
  • Do not throttle the free tier. The top of the funnel is the superpower. Tanuj Chatterjee, Super Unlimited14:59
  • A substantial paid tier needs a very large audience. Users used to generous free chat elsewhere will leave a restrictive app. Ajay Mehta, Tolan47:04
  • Super Unlimited gets over a million downloads a day, mostly organic. Tanuj Chatterjee, Super Unlimited14:59
  • Over 90% of AllTrails users are free, and they feed the SEO and word-of-mouth funnel for a low-priced subscription. Ron Schneidermann, AllTrails39:00
• Book free usage as marketing spend and keep its cost low Elena Verna, Lovable51:47
  • Lovable sponsors free credits at every hackathon and event. It tracks LLM cost on giveaways as marketing, because event hosts do the activation work. Elena Verna, Lovable51:47
  • Where you can, cut the cost of a free user. Bolt runs code in the user's browser, so cost per free user stays near zero and the free tier cannot be used for mining or DDoS. Eric Simons, Bolt23:47
  • Lovable passed $200M ARR with about 100 people and very little paid marketing or sales spend. Elena Verna, Lovable51:47

2.2 Choose what stays free

Keep free anything that spreads the product or delivers the first win, and keep the core at full quality.

• Put a feature in free if it feeds your growth model Elena Verna, Amplitude50:40
  • Free candidates are features that drive virality or network effects, support the aha moment, or create habit loops such as notifications. Gate anything that adds friction to growth. Elena Verna, Amplitude50:40
  • Use a recurring use case as the free anchor. Flo made period tracking the free retention driver and built deeper health features around it. Dmitry Gurski, Flo10:52
  • Margins plans to gate streak freezes, streak rescue and profile cosmetics, not core features, because core features drive social growth. Paul Warren, Margins49:00
• Never gate the core of the experience Sam Chaudhary, ClassDojo31:06
  • ClassDojo keeps core use free forever and charges for convenience extras such as the Memories yearbook album. Sam Chaudhary, ClassDojo31:06
  • ElevenReader keeps its best AI voices on the free plan and sells unlimited listening time instead. Jack McDermott, ElevenLabs36:50
  • About 5% of ClassDojo's base pays for something, which Sam says builds a great business. Sam Chaudhary, ClassDojo31:06
  • Penman got bad App Store reviews from stingy free limits. It made notes and notebooks unlimited on the basic plan, because the rating was worth more. Sobhan Eskandari, Penman4:37
• Leave one real reason to upgrade Ajay Mehta, Tolan30:15
  • Tolan's first free tier gave too much chat time and nothing else worth paying for. It then tested rationing the smartest models, chat time and memory strength. Ajay Mehta, Tolan30:15
  • About 20% of what you build drives 80% of willingness to pay, and it is often the easiest to build. Do not give that part away. Madhavan Ramanujam, Simon-Kucher55:04
  • Geometry Solver had everything free, so there was nothing to buy. Locking some popular shapes and labeling free versus paid gave the paywall something to sell. Steve P. Young, App Masters6:57

2.3 Gate what heavy users value most

Put the limit on the thing committed users want more of, so casual users never feel blocked.

• Cap the feature your power users care about most Kenneth Schlenker, Opal18:30
  • At Opal the capped feature is the number of blocks. Free users can still try several things, and committed users upgrade for more. Kenneth Schlenker, Opal18:30
  • GameChanger moved costly live streaming into its first paid tier. The first five live games a year stay free. Sameer Ahuja, GameChanger40:50
  • Many GameChanger users watch 15 to 25 games a year, and some reach 30, so heavy users see the logic of paying. Sameer Ahuja, GameChanger40:50
  • Chess.com's loop is free play, then game review, then another game. Users upgrade for unlimited reviews, and about 1.5 million of 200 million registered users subscribe. Albert Cheng, Chess.com46:25
• Pick a paid trigger that tracks committed behavior Busuu
  • Busuu's lock hit free users who already switch languages, so the upgrade felt like a natural next step. Busuu
  • Write the hypothesis and the segment before you test, so the result is easy to attribute to the lock. Busuu
• Sell paid features that break a free-tier rule Ravi Mehta, Tinder20:06
  • Tinder Gold shows who already swiped right. It speeds engagement up, so free users are not harmed. Ravi Mehta, Tinder20:06
  • For network products, think about the second-order effect of each free feature. Tinder capped swipes because some users swiped through a whole city and auto swipers broke gameplay. Jeff Morris, Tinder24:02
  • Tinder added north of $100 million in year-over-year revenue in its fastest growth years. Ravi Mehta, Tinder20:06
  • The swipe cap improved the ecosystem and also proved a great monetization lever. Jeff Morris, Tinder24:02
• Charge for extras once users already love the product Steve P. Young, Mematic10:43
  • Mematic gives meme making free and locks premium fonts, layouts and watermark removal behind Pro. Steve P. Young, Mematic10:43
  • Check which formats users pick most and test locking a couple of them. Steve is unsure it will convert, so treat it as a test. Steve P. Young, Mematic9:45
  • The app earns over $60,000 a month without pushing users hard to pay. Steve P. Young, Mematic10:43
  • Rootd is exploring convenience subscriptions such as dark mode and daily motivation, to avoid paywalling essential anxiety help. Ania Wysocka, Rootd24:35

2.4 Make the line visible

Free users should see what is paid before they tap, and get asked again when they show interest.

• Label paid features in the free experience Mobbin
  • Pro badges spell out the benefit of upgrading without a paywall redesign. Mobbin
  • Clear labels also stop free users from hitting unexpected paywalls. Mobbin
  • Before reworking pricing, check whether free users even know what the paid plan offers. Mobbin
  • Review features in every customer state: free, lower paid and target paid. Otherwise free users never see paid functionality exists. Elena Verna1:05:45
• Trigger a contextual upsell when users tap a locked item Steve P. Young, App Masters8:32
  • After a user declines the first upsell, give premium big home-screen space and show a contextual upgrade screen on locked taps. Do not repeat the screen that already failed. Steve P. Young, App Masters8:32
  • In Productive, tapping Join Challenge on a locked tab brings up a one-hour discount. Steve P. Young, Productive9:49
  • Chess.com converts users months or years after sign-up with contextual upsells and lifecycle messaging. Rik Haandrikman, RevenueCat35:33
• Re-ask existing free users on a schedule Michal Parizek, Mojo13:47
  • Show the paywall on app open or after a key event such as sharing, capped at once a week. Michal Parizek, Mojo13:47
  • A generous free tier is why users tolerate repeat paywalls. Michal Parizek, Mojo13:43
  • Only about 50% of Mojo's revenue comes on day one, versus roughly 80% typical across apps. Support and review analysis found no complaints. Michal Parizek, Mojo13:43
• Use free credits to get users trying, then run them out just after the first result Steve P. Young, Photo Revive AI6:52
  • Photo Revive gives 50 credits on login, and one restoration costs 35. Users see the value, then need more. Steve P. Young, Photo Revive AI6:52
  • Reward sharing, notifications and follows with credits. You cannot verify a follow, but the link still sends users to your page. Steve P. Young, Photo Revive AI5:31
  • After a dismissed hard paywall, one app told users about monthly free credits and asked for a review. Rating rose from 4.5 to 4.7 and reviews doubled to about 600. Steve P. Young, App Masters13:17

2.5 Test the line and protect the free tier

Measure free-tier changes against both revenue and retention, and walk away from wins that hurt the base.

• Test limits on installs' LTV and total retention together Kenneth Schlenker, Opal17:45
  • Look for the sweet spot where both rise. Kenneth Schlenker, Opal17:45
  • Photoroom removed its free watermark only after testing the effect on free activation and retention against lost revenue. Olivier Lemarié, Photoroom22:00
  • If the product changes monthly, free-tier tuning goes stale. Suno has kept five creates a day for about 3 years and grew to $300M ARR by improving the product. Mikey Shulman, Suno24:40
• Refuse changes that win short term and weaken the free tier Giordano Contestabile, Life36025:36
  • Life360 declined the cut because it would make the feature less useful and sometimes less safe. It accepted the short-term cost to protect trust and the network. Giordano Contestabile, Life36025:36
  • The number one reason people give for not subscribing is that the free tier is good enough. Life360 still chooses not to degrade it. Giordano Contestabile, Life36023:32
  • Paywalling free features also lets a competitor offer the same thing free. Duolingo grew by being free and fun against paid, expensive Rosetta Stone. Cem Kansu, Duolingo43:32
• Survey upgraders to find out if the free plan is too good Naomi Ionita, Evernote17:28
  • If a top reason for paying is guilt over heavy use, the paywall is in the wrong place. Do not optimize for guilt-driven payments. Naomi Ionita, Evernote17:28
  • At Evernote, feeling guilty was one of the most popular reasons for converting to the $45 a year plan. Naomi Ionita, Evernote17:28
• Put fraud defenses in place before a generous offer launches Lenny Rachitsky, Lenny's Newsletter41:27
  • Lenny's Product Pass launched with a free year of tools from Cursor, Lovable, Bolt, Replit and v0. Fraud rings found API exploits and spread them through student networks. Lenny Rachitsky, Lenny's Newsletter41:27
  • He worked with Stripe and Substack to shut them down, and his engineer barely slept for a week patching holes. Lenny Rachitsky, Lenny's Newsletter41:27

2.6 Where they disagree

Some operators say lock more, others say give more, and the right answer depends on your growth model.

• How much to lock Steve P. Young, App Masters14:26
  • Steve argues that users take everything if it is free, so lock more content. Steve P. Young, App Masters14:26
  • Headway locks nearly everything except one free daily read. Steve P. Young and Chaim, App Masters9:52
  • An indie app cut its free features from three to one, as part of changes that lifted in-app purchase revenue 211%. Josh, App Masters client app5:50
  • Opal, Life360 and Super Unlimited say the opposite. Generosity feeds the organic growth that revenue depends on. Tanuj Chatterjee, Super Unlimited14:59
• Whether to move a free feature behind a paywall Jeff Morris, Tinder26:56
  • Jeff says teams overthink backlash. Users stay if the core product still delivers value, and you can reverse the decision. Jeff Morris, Tinder26:56
  • Duolingo warns that the metrics go green for a while, then the engine stalls. Cem Kansu, Duolingo43:32
• Freemium versus a hard paywall Tomas Hoyos, Apt41:30
  • Apt's freemium experiments lowered conversion versus a hard paywall. Tomas still backs freemium for reach and for a future employer marketplace. Tomas Hoyos, Apt41:30
  • Tolan does not want to stay hard-paywalled forever, but unlimited free use would break its economics, so some features must be limited. Ajay Mehta, Tolan40:58
• Free for individuals only works with a path up market Naomi Ionita, Evernote33:53
  • Figma traded off monetizing individuals because it knew the product would spread through companies. If you lack that bridge to team or enterprise plans, offer early users a year-one discount instead. Naomi Ionita, Evernote33:53
  • Photoroom takes the long view. A casual free user may start a business later and come back to pay. Olivier Lemarié, Photoroom24:45