Founder Playbook · Sub Club by RevenueCat
10 tactics from Ora Levit
Value-Driven Growth: LinkedIn's Billion-Dollar Subscription Strategy
Watch the full episode“Within our premium subscriptions we run over a thousand AB tests a year we essentially test anything that we would launch into market uh whether it's a new feature a new layout a new way to talk about kind of the value that we provide it would all go through an AB test process.”
Run 1,000+ A/B tests per year — maximize learning speed by minimizing internal debate
LinkedIn Premium runs 3+ tests per working day across features, layouts, paywall copy, and value messaging. Levit's philosophy: time spent debating internally is worse than time spent getting signal from actual members. The bottoms-up structure means each PM owns their domain and drives tests independently — no central committee approving experiments. Speed of iteration beats quality of internal debate.
“We have a full team that's dedicated to the value that our subscriptions are providing and they're building new features and enhancing adoption of features... and that team size is similar to the team that drives growth from better pricing discovery packaging promotions etc.”
Value creation team must be same size as growth team — one without the other creates extraction without substance
LinkedIn deliberately equalizes headcount between value creation (new features, adoption, functionality depth) and value capture (pricing, discovery, packaging, promotions). Most companies over-index on one side. Levit's observation: pure growth teams extracting value without continuous value investment will eventually exhaust subscriber goodwill and cause churn. This balance is structural, not aspirational.
“We ask if you're looking for a job or if you're looking to grow your business or to hire or something else based on your answer uh we then decide what type of content to show you often times we also know because we saw that you just searched for a job.”
Personalize which plan you show based on stated intent and inferred behavioral signals
LinkedIn combines explicit intent signals (onboarding survey: job seeker vs business builder vs recruiter) with implicit behavioral data (recent search history, profile completeness) to determine which premium plan and benefit messaging to show each user. Personalization removes irrelevant offers, increases conversion, and improves retention. Every downside Levit found came from under-personalization, never over-personalization.
“Even if you were a P subscriber I might offer a free trial again usually we would do it with not within the first 12 months so if you tried LinkedIn for free in the last 12 months it will be paid but after that we do experiment and test right now with different scenarios of when and how uh we would offer uh that free trial.”
Re-offer free trials to lapsed subscribers — your product today is better than it was 2 years ago
LinkedIn experimentally re-offers free trials to past subscribers after 12+ months, on the grounds that the product has been substantially improved (20+ new benefits added). App store mechanics prevent this without workarounds but web-based products have full flexibility. The tactic compounds when combined with 'here are five things that weren't here when you last tried it' messaging — making the re-trial feel like a genuinely new product, not a second chance at an old one.
“We optimize as a true north the business for long-term revenue and I say long-term revenue not short-term revenue because there's a lot of revenue we forgo and don't want to take because we believe it will squeeze you know the funnel in a way that we're not interested in.”
Optimize for long-term revenue even when it means forgoing short-term gains
LinkedIn explicitly declines revenue that would optimize a short-term metric at the cost of member trust or long-term retention. Pricing experiments are evaluated on a 5-year revenue model, not just the first year. This long-term revenue orientation is enforced structurally: there's allocated 'revenue we won't take' and the team has permission to decline short-term extraction tactics. The result is a subscription business running at several billion dollars in annual run rate.
“As your subscription grows you see more of your members who tried you at some point come back uh that's a good sign uh it means that you know they got the value at some point in time for whatever reason they decided they're not going to use premium uh for the short term but then they came back later.”
Boomerang subscribers are proof of a value-first brand — design for dignified exits
Levit frames 'boomerang' (churned-then-returned) subscribers as a leading indicator of product health. If subscribers leave satisfied, they return when life circumstances change — new job search, new business venture. LinkedIn tracks boomerangs as a meaningful cohort and designs the cancellation experience to preserve goodwill rather than trap users with friction. A dignified cancellation is an investment in future resubscription.
“Members who generally use our features more and get to know premium better stay for the value they want to see what additional perks we have we refresh them every 6 to 12 months they want to see the monthly rotating perk.”
Partner bundles increase retention — members who engage with perks stay longer
LinkedIn Premium includes rotating partner perks (Notion with AI, Headspace, Duolingo, Spotify) that drive engagement beyond core features. Levit sees a causal link: members who engage with perks learn the full product better and retain at higher rates. The mechanism is variety — refreshing perks every 6-12 months gives subscribers a reason to re-evaluate and recommit. This pattern scales to any subscription that can negotiate third-party perks.
“You don't come to LinkedIn for AI and the magic of AI is a bit ubiquitous right like what what is AI really... the goal is not to sprinkle AI everywhere and say we're here ai is us like what does that really mean for a human right.”
Don't brand your AI features as 'AI' — members come for outcomes, not for the technology
LinkedIn positions AI tools around member outcomes (find a job faster, write a better profile summary, surface the right leads) rather than labeling them as AI features. Levit's principle: AI is infrastructure, not identity. When a feature works, members credit the outcome — not the model. Sprinkling 'AI' everywhere dilutes brand and sets expectations that are hard to consistently meet. Build for outcomes; let the technology fade into the background.
“What we call a premium session which is a session in which a premium member interacts with the premium feature and we want more essentially premium members to come more frequently interact more and for us it's a sign of value that they're arriving for the product which is a sign that they will continue staying with us for the long term.”
Track premium sessions, not just payments — feature engagement predicts retention before churn is visible
LinkedIn's leading retention indicator is 'premium sessions' — sessions where paid members actively engage with premium-exclusive features. Payments are lagging indicators; premium session frequency is predictive. A subscriber paying but not using premium features is a high churn risk even before cancellation. This metric bridges activity data and revenue data, giving an earlier intervention window before the subscriber has already made the mental decision to leave.
“LinkedIn's mission is to provide an economic opportunity for every member of the global workforce and that opportunity is not paid... we wouldn't add anything that would ever make something that's related to uh finding a job or applying to a job a paid functionality.”
Free ecosystem health drives paid subscription growth — never paywall the core value prop
LinkedIn protects free job-seeking functionality as a structural commitment — it drives the supply side of the marketplace (candidates) which sustains the demand side (recruiters paying for Talent Solutions). Making job applications paywalled would kill recruiter value. Levit's principle for any freemium product: find the free experience that sustains the ecosystem, protect it vigorously, and build premium only on top of features that help users achieve more — not features that restrict baseline access.