How Tinder grows

In 2018 Tinder A/B tested a toggle in the app's top navigation bar that let users enter a new subscription tier directly. Treated as a small-to-medium 'throwaway' test on the experiment roadmap; the design team opposed adding the top nav. Result was read off the daily revenue dashboard. Guest's broader advice: audit where users can reach premium features and add more prominent paywall entry points; don't bury them in the nav.

Problem: many tiers (Plus/Gold/Platinum) x plans (weekly/monthly) plus a la carte items caused decision overload; some users bought nothing, others bought Platinum but only used Gold features. The ML team built models predicting willingness to pay; at paywall time the paywall asks the model which paywall/product to recommend for that user, instead of showing everyone the same static paywall. Rolled out small first (a couple of features/paywalls, not the whole scope) to limit cost and risk. A/B test: control = historical static paywall, treatment = model-chosen paywall; measured conversion and total revenue. Counter-metrics: repurchase, renewal and cancellation of users nudged to higher tiers; several model variants compared against each other and against no-model paywall.

Unlike paid-only Match and eHarmony, Tinder is free-to-play: no friction to join, so the network is large. Free matters most for in-demand users (mostly women, who get many right-swipes) so they have no reason not to be there; paid products target users who get fewer swipes (mostly men) with Boosts and Super Likes. Ads monetize free users a little.

Microtransactions designed around the core outcome (more matches): Boost shows your profile to more people (like LinkedIn's boost); Super Like sends a differentiated like. Guest's rule: good a la carte products scale roughly linearly in value (10 is 10x better than 1). Other microtransactions were tried; some stuck, some did not.

After the standalone Passport cannibalized Plus, tested a higher a la carte price and measured conversion, subscription cannibalization and total revenue. Price amounts not given.

Raised the a la carte Passport price again so the 7-day Passport costs the same as a 7-day Plus subscription (which includes Passport and more), and added an upsell on the paywall showing both: 'you can buy Passport, standalone a la carte, but also you can buy Plus subscription'. The subscription becomes the obvious better deal.

Every subscription feature had its own paywall entry point. A dashboard reviewed every morning showed clicks per entry point, conversion rate and the mapped incremental subscription revenue. Features that did not convert were either given more prominent placement or made free. Never moved a free feature behind the paywall.

Instead of one price, Tinder sells Tinder Plus (cheap on-ramp that gets people over the 'penny gap'), Tinder Gold (see who liked you) and Tinder Platinum (more Super Likes/Boosts, message before matching), and layers a la carte microtransactions (Super Likes, Boosts, bundles) on top to catch willingness to pay between and above tiers. Subscriptions give predictable revenue; consumables flex with need. Keep to 2-3 tiers to avoid complexity; the app avoids showing all tiers on one paywall. Few free users buy consumables: subscriptions act as a force multiplier because consumables (get more right-swipes) pair with Gold (see who swiped right). Built organically: one tier launched, learned what users valued, added the next.

Where Match/eHarmony required 20-100 profile questions (~20 minutes), Tinder used (1) Facebook login instead of account creation, (2) photo picking from Facebook, (3) a short Twitter-like bio. Users could go from curious to swiping in 1.5-2 minutes, getting to the aha moment (a conversation with someone) fast.

Passport (travel mode: swipe anywhere in the world) was previously only in the subscription. Chosen for unbundling because only some users need it, occasionally (when travelling or meeting people outside their city) and they don't want a subscription commitment. Offered as a standalone a la carte purchase for 1 day, 3 days or a week to capture non-subscribers. Criteria used: features whose use is situational (like Boost, Super Like) suit a la carte.

Put a paywall on the number of swipes a free user can make per day, after seeing people swipe through their entire city in a few sessions and use auto-swipers. The goal was a healthier marketplace; monetization was a side benefit. General principle: charge for one or two things early, then follow power users to decide which 'superpowers' to gate, considering what share of users should have each one.

Moved from a single tier (Tinder Plus) to three subscription tiers with consumables on top, designed as packages for different customers: from a younger, lower-income, often non-US entry subscriber up to whales spending as much as $50,000 a year. Required building new bundles and new features for each segment. Recommended the same to a portfolio company stuck on a single tier.

In 2016 built a dedicated revenue team inside product: the guest plus three iOS and three Android engineers, running a roadmap of small/medium/large experiments with revenue features treated as their own product line. Grew to 40+ engineers. Downside at public-company scale: finance-driven forecasting pushed the team into 'Excel culture' instead of testing ideas.

Tinder Select: a higher-end tier above Platinum targeting 'whales' who already spend heavily on top subscriptions and a la carte, offering better service and product. Guest's diagnosis: identity/brand fit; users of a mass-market app don't feel the environment is exclusive (like an expensive table in a regular club vs an exclusive club), so the luxury benefits don't meet expectations.

Tinder focused on core engagement loops and retention first, and only turned on serious monetization in years four to five, at 30-40M monthly actives. The guest contrasts this with today's apps that monetize in the first session and can mistake hard-paywall revenue from young cohorts for product-market fit.

Free Tinder users see ads; an ad impression is orders of magnitude less valuable than a subscriber, so it fills gaps rather than driving the business. Eiting adds that across apps he talks to, ads are usually 10-20% of revenue and help subsidize free users.

Spotted Tinder Gold subscribers spending $100-250/month on microtransactions. Researched their motivation: status signalling or utility? It was utility (more matches); users framed it against ~$400/month spent on dates. Built Tinder Platinum, bundling extra Super Likes and Boosts and letting people break a core rule: message someone before matching. Insight: they were hitting the limit of what Tinder could do for them, not of what they would spend.

Tested multiple price points for every product: Tinder Plus at $7.99, $8.99 and $9.99, other points for Gold, for Boosts and for bundles of Boosts and Super Likes, each multiplied by location (UK vs South America vs US). Result: hundreds of App Store SKUs, reportedly more than any other app.

Using the paywall ML model, users are shown the subscription first; those who don't buy are then shown the a la carte Passport as a 'second chance'.

Replaced a set-and-forget single price from early basic tests with a complex pricing grid driven by many factors, geography first. Pricing work started when the US market saturated and growth shifted to India and LATAM, which had much lower willingness to pay than North America/UK. Advice: focus pricing on your core market first (North America for most US developers).

Paywalls and paid interactions got as much design love as any core feature. Example: in 2017 added a gold shimmer animation to a paywall, about a couple of days of engineering time. Warns against PM-designed paywalls and against too many paywalls making the product feel spammy.

At Tinder, users were asked to subscribe only after the magic moment: getting a match and having a conversation. Contrast: an AI website builder pushed freemium users to pay before they could get a working (debugged) build.