How to structure subscription tiers

1 The bottom line

Design tiers around distinct customer segments, and put the paywall behind what extends the product, never the core reason people come or the growth loop. Tiers that fit real differences in value and cost capture more revenue without confusing buyers.

2 Key takeaways

• Build tiers for segments that differ in the value they get and what they will pay. One tier underprices your avid users and overprices your casual ones.
  • Avid cross-device Evernote users said they got hundreds of dollars of value while paying $45 a year Naomi Ionita, Evernote17:36.
  • A portfolio company at $50 million ARR with one tier and no pricing work could probably double revenue Jeff Morris, Tinder19:51.
• Draw the paywall line by what the customer came to buy and by how new users arrive. Gating the loop kills growth, and gating the core feels like a bait and switch.
  • A Photomath rival capped scans and killed its core loop. Photomath monetized something else and saw roughly a 100x outcome, while the rival sold for perhaps 10x less or lower Francesco Polizzi, Discord29:19.
  • Give away too much and there is nothing left to expand into. Disruptors often land but cannot expand Madhavan Ramanujam, Simon-Kucher / 49 Palms57:34.
• Give each tier one clearly different promise. Buyers who cannot tell tiers apart pick wrong or buy nothing.
  • Duolingo's unclear tier differentiation held Max back until the conversation positioning and Video Call arrived Natalia Castillejo, Duolingo37:17.
• Check unit economics before you add a tier. If heavier use costs you more than it earns, your incentives clash with your customers.
  • Two years before DashPass, DoorDash lost 10 or 12 dollars per order in San Francisco, and a subscription would not have worked Jimmy Liu, DoorDash14:55.
  • Gamma requires the average cost per plan to be profitable and accepts losing money on its 90th percentile of heavy users Jon Noronha, Gamma38:54.
• Track every paywall entry point and test pricing page changes. Only a few features drive most subscriptions, and a clearer page lifts the whole funnel.
  • At Tinder, two to three power features drove paid conversion and the rest were incremental Jeff Morris, Tinder30:55.
  • GitHub's pricing page changes raised free and paid signups by 20+% Github, Github.

a single premium tier is often a mistake

Naomi Ionita, Evernote

2.1 Find your segments

Tiers only work when each one fits a group that needs and pays for something different.

• Segment by need and willingness to pay, not by persona. Madhavan Ramanujam, Simon-Kucher41:40
  • Test whether a segment lets you act differently in product, messaging and sales.
  • Know the segments from day one, but launch for the most attractive one first and add others later.
  • Only about 10 percent of companies claiming a segmentation strategy have a meaningful one. Prince Charles and Ozzy Osbourne fit the same persona but need and pay for different things.
  • Apple's iPhone ladder, Uber's Black, X, Comfort and Pool, and Eventbrite's three plans are examples of segment-based ladders.
• Research which users get the most value and what they would pay. Naomi Ionita, Evernote17:36
  • Build a plan per segment instead of one premium tier.
  • New Evernote users saw it as a pre-installed notepad and balked at $45.
• Think of tiers as packages for very different customers. Jeff Morris, Tinder19:51
  • Give each group its own bundle, features and experience.
  • Tinder's range ran from a younger, lower-income, non-US subscriber to a top spender who might pay $50,000 a year.
• Get one tier working before you add more. Ravi Mehta, Tinder30:00
  • If a single tier cannot work, that is a product market fit problem, not a monetization one.
  • Add tiers and consumables once the business is thriving and you can serve both low-engagement and high-engagement users.
  • Phil Carter notes there are over 100,000 consumer subscription apps in the app stores and fewer than 50 have reached billion dollar valuations.

2.2 Draw the paywall line

Keep free what brings people in and what they came to buy, and charge for what extends it.

• Keep the heart of why someone buys free. Michael Segal and Mark Ungerer, Skylight11:09
  • For the photo frame, photos to grandma stay free. For the calendar, events, tasks and chores stay free.
  • Put extras like video, recipes, meals and an AI engine in the subscription.
  • Do not advertise paid features so loudly that love turns into surprise or frustration after purchase.
  • Do not gate features users stumble into mid-configuration Michael Segal and Mark Ungerer, Skylight16:01.
• Never paywall the feature that drives your growth loop. Francesco Polizzi, Discord29:19
  • Discord's loop was friends inviting friends, so it charged for file upload size and custom emojis, which do not stop anyone joining a server.
  • What sits behind the paywall must still be genuinely valuable to someone.
• Split features into day one and day 100. Naomi Ionita, Evernote13:58
  • Put features that deliver value the first time someone uses the product in the first paid plan.
  • Push advanced features, or ones that pay off only at scale of data, into a Pro tier sold later as an upsell.
  • At invoice2go this rebalancing doubled the upgrade rate from Starter to Pro while the Pro price rose about 30 percent.
• Find the one or two features that actually convert. Naomi Ionita, Evernote20:55
  • Have users mark each current and planned feature as must-have, nice-to-have or not necessary, or give them 100 points to spend across features.
  • Or show a subset of about six features and ask for the single must-have and single least important, then rotate subsets several times Madhavan Ramanujam, Simon-Kucher29:30.
  • It is very rare that people convert equally across all features.

2.3 Build the tiers

Pick a small ladder, give each step its own promise, and tie the steps to what drives your costs.

• If you have no ML team, start with three tiers. Shawn Gong, Tinder11:13
  • Some users buy the most expensive, some the cheapest, and the undecided buy the middle.
  • With one product the question is buy or not. With three it becomes which one to get.
  • Life360 used a model with about 900 data points to show Platinum instead of the default Gold to likely buyers. It doubled the share of new users subscribing to Platinum without losing a single Gold subscriber Giordano Contestabile, Life36047:48.
• Give each tier a promise that differs from the one below. Ravi Mehta, Tinder22:05
  • Tinder Plus is about swiping more, Gold about getting more matches, and Platinum about getting into better conversations.
  • Put your most emotionally resonant feature at the center of the premium tier Natalia Castillejo, Duolingo24:25.
  • Duolingo's Max resonates more with English learners, who often learn for jobs and careers, and with serious learners. Casual users may find Super enough Natalia Castillejo, Duolingo28:20.
• Tie tiers to what drives your costs. Ajay Mehta, Tolan18:33
  • When one subscriber uses the app one hour a week and another seven, a single tier leaves heavy users underpaying and light users priced out.
  • For AI products, put non-AI features and a taste of AI in a basic plan and full AI usage in a pro or max plan Phil Carter, Elemental Growth53:10.
  • Or differentiate mainly on credit amount, so users who run out simply upgrade Coco Mao, OpenArt28:16.
  • Hundreds, maybe thousands, of Tolan subscribers pay $200 a month for Unlimited, such as long-haul truck drivers who chat 8 hours a day.
  • OpenArt tried feature gating and did not find it worked well.
• Add a premium tier when your best capabilities cannot scale to everyone. Nick Turley, OpenAI41:25
  • ChatGPT Plus lost its uptime and capability edge once so many users joined. The $200 tier gave OpenAI a way to ship very powerful, expensive models to the people who care most.
  • The $200 tier was a 10X jump over the $20 plan, which breaks the standard SaaS pricing page look but fits the product.
  • Duolingo Max launched in 2023 with few features in few countries. The bigger impact came about a year later, from better features, wider scale and Video Call Natalia Castillejo, Duolingo36:15.

2.4 Sell occasional and shared use outside the ladder

Standalone purchases and shared plans reach buyers the main ladder misses, if you price them so the subscription still wins.

• Unbundle features that people need only at certain times. Shawn Gong, Tinder13:53
  • Pick features tied to timing or circumstance, like Boost during peak hours, Super Likes, or Passport for travel.
  • Sell them to users who do not want a subscription commitment.
  • Passport was sold for one day, three days or a week to non-subscribers, and its conversion "went crazy" at launch.
• Price standalone purchases so the subscription looks like the better deal. Shawn Gong, Tinder16:21
  • Tinder priced 7-day Passport the same as the 7-day Plus subscription, so Plus gives more for the same price.
  • If buyers skip the subscription, raise the a la carte price and keep testing higher Shawn Gong, Tinder15:55.
  • After the a la carte price rose, cannibalization dropped, conversion fell a little, and total revenue went up.
• Add cheap entry purchases and several ways to buy. Chaim Sajnovsky and Steve P. Young, App Masters17:08
  • Offer choices like remove ads, unlock certain features and unlock all, because people feel in control with options.
  • Add a one-time remove-ads purchase at $2 or $5. Once someone pays for anything they are more likely to pay again Steve P. Young, App Masters11:34.
  • A geometry solver app went from $1 a day to $17 a day by locking certain features and adding multiple purchase options, with no extra downloads.
  • Spotify and Netflix run ad-supported plans, and many people who won't pay full price take a lower price with ads Steve P. Young, App Masters9:38.
• Offer shared plans when the product is multiplayer or shareable. Jonathan Parra, Superwall49:45
  • A friends and family tier, for example through iCloud family, raises ARPU and LTV.
  • Offer a discounted two-person add-on inside the onboarding quiz, and keep only a small toggle on the pricing page for those who decline Ethan Ethier, Built With Science30:00.
  • Around 15% of people who start a trial add a gym buddy, and both get 15% off. A partner also keeps users subscribed longer.
  • Shared plans take extra implementation and Apple SDKs, which is why few apps offer them.

2.5 Show the tiers on the pricing page

The same tiers convert differently depending on how the page frames the value gap.

• Make the higher tier look like clearly more value. Github, Github
  • Show more features in the higher tiers, and give the features customers value most prominence.
  • To test packaging cheaply, Paddle gave every buyer all features and changed only the presentation, making the higher plan look like much more for about $10 more Steve Young and Lucas, Paddle48:12.
  • GitHub also dropped an outdated tier and used more vibrant colors for a cleaner page.
• Put a decoy or middle option beside your target plan. Steve Young, App Masters10:55
  • Split the monthly plan into a stripped-down and a premium version, and keep the core features in the plan you want people to pick.
  • Conversion rose 233% in the A/B test, and 86% of buyers chose the yearly plan versus 14% monthly.
  • In the Economist study from Dan Ariely, 84% bought print plus web when a print-only option was shown. Without it, 68% bought web only Steve Young, App Masters8:21.
• Name plan differences as benefits, not features. Madhavan Ramanujam, Simon-Kucher1:13:10
  • If you are passionate about every bell and whistle, you are probably pitching features.
  • Describe the difference as a benefit, such as inventory locations at Shopify.
  • SmugMug moved from a multi-page feature list to benefit-based communication and saw a double digit revenue improvement with no product changes.

2.6 Track, test and adjust

Treat the first tier design as a hypothesis, and let paywall data and user conversations move the line.

• Give every paid feature its own paywall entry point and watch it daily. Jeff Morris, Tinder29:56
  • Use a dashboard with clicks and conversion per entry point, mapped to incremental subscription revenue.
  • If a feature's paywall does not move the needle, make the feature free.
  • Low conversion can also mean the placement is too weak, so check placement before deciding the feature lacks value.
• Test simpler pages and talk to users before trusting a change. Shawn Gong, Tinder2:45
  • Tinder user conversations showed some people bought Platinum only because it was the most expensive while using Gold features, and others bought nothing because of too many options.
  • The first calls for better education or recommendations. The second hurts conversion.
  • Instapage's simplified pricing page converted at 3.44%, slightly below the original at 3.67%, so treat a cleaner layout as a hypothesis Instapage, Instapage.
• Roll expensive features down the ladder as costs fall. Ajay Mehta, Tolan23:00
  • Launch frontier capabilities in the top tier or a limited beta, then move them to Plus and then Light. Some can eventually land in free.
  • Photoroom kept most AI features in Pro for about two years, then by mid 2025 gave more away free and upsold specific features Olivier Lemarié, Photoroom16:40.
  • Suno gives a new model only to the paid tier and upgrades the free model later, though Mikey Shulman admits the heuristic is not deeply thought out Mikey Shulman, Suno23:09.
• Launch new features inside the paywall, not after. Michael Segal and Mark Ungerer, Skylight16:01
  • Do it when customers can quickly grasp the feature and the core product they paid for stays intact.
  • Moving a feature out of a paywall to test it is easy. Moving a free feature into a paywall is painful.
  • When moving from a $5 paid app to a subscription, make what people paid for the free tier and gate only new features Josh and Steve P Young, Miracle Morning13:01.

2.7 Where they disagree

Tier ladders versus one paid plan: it depends

	Tier ladder	One paid plan
Structure	Light, Plus, Unlimited	Free + ultra plan
Works when	Usage and willingness to pay vary widely	Priced against Spotify, Netflix
Buyer's question	Which tier fits me?	Free or paid? Cheap or expensive?
Example	Truckers on $200 Unlimited	Unlimited text-to-speech

How many tiers, and what separates them, depends on the product and who buys it.

• One paid plan or several.
  • ElevenReader collapsed its paid tiers into one unlimited plan and kept a larger free plan, because consumers judge price as free or paid, cheap or expensive Jack McDermott, ElevenLabs39:40.
  • On mobile, ElevenLabs keeps one core plan plus pay-as-you-go, because app store downgrades, discounts and syncing across Google, Apple and Stripe are hard Jack McDermott, ElevenLabs42:20.
  • Tinder and Tolan run three tiers, which suits novelty-driven, high-churn products better than habit products like Netflix or Headspace Ravi Mehta, Tinder30:00.
• Simple pricing or many packages.
  • Intercom gave away about $50 million in ARR to simplify, because simpler and fairer pricing makes customers stay longer Eoghan McCabe, Intercom18:32.
  • Basecamp caps its price at $299 for unlimited users, with no enterprise tiers and no custom work Jason Fried, 37signals11:08.
  • Tinder's biggest unlock was packages for very different customer types, up to the top spender Jeff Morris, Tinder19:51.
• Gate by feature or by credits.
  • OpenArt found credits worked better than feature gating Coco Mao, OpenArt28:16.
  • Tinder and Duolingo win with a distinct feature story per tier Ravi Mehta, Tinder22:05.
• Limit the lower plan or load it with value.
  • Gamma launched its $10 plan deliberately limited, for fear of cannibalizing the $20 plan. Over time it moved more value into Plus and found the lower plan retains as well or better. The split went from roughly 1/3 Plus and 2/3 Pro to about 2/3 Plus and 1/3 Pro Jon Noronha, Gamma36:00.
  • Super Unlimited would rather add value to the paid tier before asking for more money Tanuj Chatterjee, Super Unlimited13:34.
• A luxury top tier.
  • OpenAI's $200 tier works because it carries models that cannot scale to everyone Nick Turley, OpenAI41:25.
  • Tinder Select at $499 a month clashed with a mass brand, and Tinder is gradually scaling it down Shawn Gong, Tinder19:10.