4 guests 6 episodes 1,787 words

Retention Is Not Your Destiny: The Heretical Case Against the Most Sacred Metric in Growth

Is retention the single most important metric for a product?

Duolingo has over 9 million users with a year-plus streak. Their streak feature is the most impactful growth lever in the company's history, driving a $14B valuation that doubled in six months. By every conventional measure, retention is the metric that matters most.

And then there is Shopify, which deliberately optimizes for churn. Most businesses fail, so Shopify lowers barriers to get as many people trying entrepreneurship as possible, knowing that the winners in each cohort will generate enough revenue to make the whole thing work.

If retention is destiny, why is one of the most successful commerce platforms on earth actively encouraging churn? The answer reveals that "retention" is not one thing -- it is a family of related concepts, and conflating them leads to strategies that are precisely wrong for your business model.

Every growth expert says retention is the metric that matters most. Acquire all the users you want -- if they do not stick, you are filling a leaky bucket. But is retention truly the master metric? Or are there important cases where obsessing over retention leads you astray?

The 4 Positions

1 / 4

Retention is the single most important lever -- and gamification mechanics can supercharge it

Jackson Shuttleworth (Group PM · Retention Team at Duolingo)

Advocates: Jackson Shuttleworth (Group PM, Retention Team at Duolingo)

Duolingo is the strongest case study for retention as destiny. Their streak feature -- a simple daily usage counter -- is the most impactful feature in company history.

"Streaks is the most impactful feature. We have, right now, over 9 million users with a year plus streak. If you look at the numbers, I think it's been our biggest growth lever."
Jackson Shuttleworth
▶ 00:00:00

But Shuttleworth offers a critical caveat that retention maximalists often ignore: gamification mechanics only work when they are layered on top of genuine product value.

"Streaks are a great engagement hack. I'm of the opinion that any team, any app out there can introduce a streak, and if you figure it out, it probably works to retain users, but at the core, you have to have an app that people want to use."
Jackson Shuttleworth
▶ 00:07:15

This distinction matters. Duolingo's streak works because learning a language is genuinely valuable and the app is genuinely good at teaching. The streak amplifies existing value -- it does not create it. Teams that slap gamification onto a mediocre product are cargo-culting the mechanic without understanding the foundation.

When This Applies

Consumer product with daily use case, subscription revenue model, and a core action that benefits from habit formation

Retention is the foundation, but accruing benefits create real defensibility

Sarah Tavel (General Partner at Benchmark · former Pinterest)

Advocates: Sarah Tavel (General Partner at Benchmark, former Pinterest)

Sarah Tavel's hierarchy of engagement framework goes beyond raw retention to explain why some retained users create defensible businesses and others do not.

"What you realize when you look at social products is that there's almost this action which I call the core action of that product that forms the foundation of the product."
Sarah Tavel
▶ 00:05:45

Level 1 is getting users to complete the core action (retention). Level 2 is accruing benefits -- the product gets better with use, creating switching costs. Level 3 is creating virtuous loops that defend against competitors. Pinterest becomes more valuable as you save more pins. Your Spotify recommendations improve with every listen. Your LinkedIn network makes the platform harder to leave.

"The MAU thing doesn't really mean anything. It's really looking at users completing the core action."
Sarah Tavel
▶ 00:05:45

The practical implication: measuring MAU or DAU as your retention metric is almost meaningless. A user who logs in but does not complete the core action is not truly retained. The metric must capture whether users are doing the thing that creates value, not just whether they showed up.

When This Applies

Product where usage makes the product better over time (social networks, recommendation engines, content libraries), and you need defensibility against competitors

Optimize for churn, not retention -- some business models thrive on high turnover

Archie Abrams (VP Product · Head of Growth at Shopify)

Advocates: Archie Abrams (VP Product, Head of Growth at Shopify)

This is the most heretical position in the retention debate, and it comes from one of the most successful growth leaders in tech.

"Most businesses do ultimately fail. And so the way we look at it is can we lower the barriers to getting started and get as many people in the door trying their hand at entrepreneurship?"
Archie Abrams
▶ 00:07:06

Shopify's business model is fundamentally different from typical SaaS. Their revenue increasingly comes from payments (a percentage of GMV), not from subscriptions. This means a cohort of 1,000 merchants where 900 churn but 100 become successful businesses can be more profitable than a cohort of 1,000 merchants where 500 stay but none grow significantly.

"That's really a different insight than most SaaS companies that they get a customer, they really never want that person to leave."
Archie Abrams
▶ 00:08:33

The implication is radical: if your revenue is tied to customer success (not customer existence), high churn can be a sign of health. You want to maximize the number of at-bats, knowing that the winners will more than compensate for the losses. This is venture capital logic applied to a product growth model.

When This Applies

Revenue is tied to customer success rather than customer existence (payments, commissions, usage-based), high natural churn is expected in your market, and your cohort economics depend on the winners, not the average

Retention is mostly an onboarding problem, not a product problem

Sean Ellis (Author of Hacking Growth · GrowthHackers)

Advocates: Sean Ellis (Author of Hacking Growth, GrowthHackers)

Sean Ellis reframes the retention question entirely. Most teams that struggle with retention diagnose it as a product problem (the product is not sticky enough) when it is actually an onboarding problem (users never experienced the product's value in the first place).

"It's usually much more function of onboarding to the right user experience than it is about the kind of the tactical things that people try to do to improve retention."
Sean Ellis
▶ 00:00:45

Before investing in retention features (streaks, notifications, re-engagement campaigns), ensure that users are reaching the must-have experience during onboarding. If 70% of users churn before experiencing the core value, the solution is not better push notifications -- it is better onboarding that gets users to the aha moment faster.

This connects directly to the activation metric discussion: retention and activation are not separate problems. They are the same problem viewed from different time horizons. Poor activation leads directly to poor retention, and no amount of downstream retention tactics can compensate.

Ellis's insight is particularly important for companies that have layered retention mechanics (push notifications, email campaigns, re-engagement flows) on top of a fundamentally broken onboarding experience. These companies see high costs for low retention gains because they are treating symptoms rather than causes. The first step in any retention improvement effort should be an honest assessment: what percentage of churned users actually experienced the core value? If the answer is below 50%, the problem is upstream, and every dollar spent on downstream retention tactics is largely wasted.

When This Applies

Your retention curve drops steeply in the first week, users who make it past day 7 retain well, or qualitative research shows that churned users never experienced the core value

Evidence from the Archive

Duolingo

Duolingo's streak feature has 9 million users with year-plus streaks -- equivalent to a large city's population...

Duolingo's streak feature has 9 million users with year-plus streaks -- equivalent to a large city's population using the app daily for over a year

Jackson Shuttleworth · Group PM, Retention Team ▶ Watch

Benchmark

Pinterest's picked-for-you algorithmic feed was one of the first in any social product -- more pinning led to better...

Pinterest's picked-for-you algorithmic feed was one of the first in any social product -- more pinning led to better recommendations, creating a virtuous Level 2 cycle

Sarah Tavel · General Partner, former Pinterest ▶ 00:05:45

Shopify

Shopify represents 10% of US e-commerce with $235B GMV in 2023 -- roughly the economy of Finland -- built on a model...

Shopify represents 10% of US e-commerce with $235B GMV in 2023 -- roughly the economy of Finland -- built on a model that accepts high churn

Archie Abrams · VP Product, Head of Growth ▶ 00:07:06

GrowthHackers

Lookout (mobile security): moved from 7% to 40% very-disappointed in two weeks with zero product changes --...

Xobni: Sean originally created the 'very disappointed' question because senior management users were never 'satisfied' with anything -- flipping to 'what would you lose' got honest answers

Sean Ellis · Author of Hacking Growth ▶ 00:00:45

The Synthesis

Retention is not destiny. Retention of the right behavior is destiny.

Right Behavior Retention

Is retention really destiny?

Diagnostic Not Goal

What happens when you treat retention as a goal rather than a diagnostic?

Category-Aware Benchmarks

Why is comparing retention across product categories misleading?

Retention is not destiny. Retention of the right behavior is destiny. The hierarchy: users must experience core value (activation), keep doing it (retention of the core action, not just logins), create accruing benefits that make leaving harder, and the business model must convert retained behavior into compounding revenue.

If you optimize for retention without understanding what behavior you are retaining, you end up building sticky products users resent -- notification spam, dark patterns, artificial switching costs -- rather than products users genuinely value. Gamification must be amplification, not substitution.

Different product categories have fundamentally different retention ceilings. A productivity tool used Monday through Friday cannot expect the same D7 retention as a social network. A seasonal product cannot be evaluated on month-over-month retention. The right benchmark is your category's retention curve, not an abstract 'good retention' number.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

What is your product's usage pattern?

Question 2

Where does your retention curve drop most steeply?

Question 3

What is your business model?

Notable Absences

The Bottom Line

The final nuance: different product categories have fundamentally different retention ceilings, and comparing across categories is misleading. A productivity tool used Monday through Friday cannot expect the same D7 retention as a social network that is part of daily life. A seasonal product (tax software, holiday shopping) cannot be evaluated on month-over-month retention. The right benchmark is your category's retention curve, not an abstract "good retention" number. Lenny's retention benchmarks across consumer and B2B products provide the reference points, but they must be applied with category awareness.

The Duolingo deep-dive in Lenny's newsletter provides the most detailed case study of how retention-focused features compound over time. Their growth was not linear -- it was exponential, driven by a systematic investment in retention mechanics that started with streaks and expanded to leaderboards, friend quests, streak freezes, and perfectly timed notifications. But the newsletter makes clear that every one of these features was layered on top of a product that genuinely teaches languages well. The gamification was amplification, not substitution.

Sources

Jackson Shuttleworth — "Behind the product: Duolingo streaks | Jackson Shuttleworth (Group PM, Retention Team)" — Lenny's Podcast, December 15, 2024
Sarah Tavel — "The hierarchy of engagement | Sarah Tavel (Benchmark, Greylock, Pinterest)" — Lenny's Podcast, December 27, 2023
Archie Abrams — "Breaking the rules of growth: Why Shopify bans KPIs, optimizes for churn, prioritizes intuition, and builds toward a 100-year vision | Archie Abrams (VP Product, Head of Growth at Shopify)" — Lenny's Podcast, November 7, 2024
Sean Ellis — "The original growth hacker reveals his secrets | Sean Ellis (author of “Hacking Growth”)" — Lenny's Podcast, September 5, 2024