All podcasts / Lenny's Podcast / Summary

5 questions to ask when your product stops growing | Jason Cohen (2x unicorn founder)

2026-01-25 - 106 min - source - Read full transcript

Lenny Rachitsky (host)Jason Cohen

Key insights

Diagnosing stalled growth requires a strict-order sequence of five questions, not a flat checklist.

Cohen's framework asks, in order: are customers leaving, is pricing correct, are existing customers growing (NRR), are acquisition channels saturated, and do you actually need to grow. Fixing a lower-order issue doesn't matter if a higher-order one - like churn - is still broken, the same way tuning the bottom of a leaky funnel doesn't help if the top is the real problem.

growth-diagnostics

A single formula reveals the hard ceiling that churn imposes on a business.

Maximum sustainable customer count equals new customers added per month divided by the monthly cancellation rate. At 100 new customers/month and 5% churn, a company can never exceed 2,000 customers, because cancellations grow in absolute terms as the base grows while marketing-driven acquisition typically does not.

churn-and-retention

'Too expensive' is almost never the real reason customers cancel.

A customer who cancels already saw the pricing page, decided it was worth it, and paid - so price alone didn't stop them from buying in the first place. Cohen's death-certificate analogy argues teams should dig past the stated reason ('too expensive,' 'project ended') to the deeper cause, such as a missing integration or an unmet promise.

churn-and-retention

Asking 'what made you cancel' instead of 'why did you cancel' roughly doubled usable survey responses in a real case study.

Cohen cites Groove's cancellation-email test: switching the question phrasing raised usable responses from about 10% to 20%, because 'why' invites a canned excuse while 'what' prompts the customer to describe the actual product or situational issue.

churn-and-retention

Raising B2B SaaS prices usually doesn't reduce signups the way textbook demand curves predict.

Price functions as a quality and positioning signal: mid-size and larger companies often distrust products priced too low as not enterprise-ready. Raising price can move a product into a segment that actually wants and trusts it, producing a 'mesa' shaped demand curve rather than a simple downward slope.

pricing-strategy

How a price is framed against what the customer already values changes willingness to pay far more than the underlying value delivered.

In Cohen's 'Double Down' thought experiment, pitching an ad-cost-halving product as 'cut your cost in half' caps what a buyer will pay at a fraction of their savings (about $5K/month on $40K spend), but reframing the identical product as 'double your leads' lets the same buyer justify roughly 8x the price ($40K), because the buyer pays for more of what their organization already prioritizes (growth) rather than for savings.

pricing-strategy

Net revenue retention (NRR) alone can hide a business that is dying at the customer-count level.

NRR blends upgrades against cancellations and downgrades into a single ratio, but percentage losses and gains aren't symmetric - a 20% loss requires a 25% gain just to break even - and NRR says nothing about whether enough customers remain to expand. Cohen notes essentially no large public SaaS company has NRR below 100%, and the median at IPO is around 119%, which is why logo (customer-count) retention must be tracked alongside it.

churn-and-retention

Growth channels don't plateau into a stable S-curve - they eventually decline, a pattern Cohen calls the 'elephant curve.'

Channels saturate as the same audience is repeatedly exposed to the same message, and the channel's own effectiveness can decline over time, as seen historically with magazine ad circulation and conference attendance. Relying indefinitely on one channel, or just adding another feature and 'flogging marketing harder,' eventually stops working.

channel-saturation

Breaking out of a saturated acquisition channel usually requires a genuinely new channel type, not incremental tuning of the existing one.

Examples: Constant Contact restarted growth by running in-person small-business workshops in cities; HubSpot deliberately built an agency-partner channel that grew to roughly 50% of revenue within four to five years; WP Engine sells heavily through web-design agencies. These required committing to a new, sometimes unscalable-looking bet rather than optimizing existing paid or organic channels.

channel-saturation

AI/LLMs are good at summarizing themes in customer feedback but bad at surfacing the specific, actionable detail that actually drives a decision.

Because an LLM behaves like an averaging machine, it's well suited to identifying common topics in survey responses but tends to smooth over the surprising, specific complaint - like a missing integration - that would actually change a roadmap. Cohen's workaround is to have the AI extract themes, then force it to attach every underlying verbatim detail and customer link to each theme so a human can still spot the trigger.

growth-diagnostics

'If you're not growing, you're dying' is worth interrogating rather than assuming as a business law.

Cohen's final diagnostic question is whether the company - or the founder personally - actually needs to keep growing. Some bootstrapped businesses are legitimately fine optimizing for profit or stability instead of growth, but he argues the phrase remains more reliably true of the person than of the company, since most people drawn to building products and startups are wired to want to keep learning and advancing.

purpose-and-ambition

A/B testing is largely a waste of time below meaningful traffic scale, and it cannot validate strategic decisions.

Cohen's contrarian take: A/B testing can't tell you if a strategy or vision is right, and because true effects are statistically rare, most 'wins' from routine tests (copy, buttons) are false positives that vanish when re-checked. He cites a Shopify example where a 100-person, statistically sophisticated experimentation team found roughly a third of their tested wins disappeared on holdout re-verification; Lenny counters it's still valuable at very high traffic and scale.

growth-diagnostics

Books referenced

On Writing Well: The Classic Guide to Writing Nonfiction - William Zinsser - Cohen's most-recommended writing book; he cites Zinsser's line 'I never knew writing could be easy' as summing up the craft.
Crossing the Chasm - Geoffrey Moore - Cohen's top product recommendation - he argues most people have only seen a summary blog post, not read the actual model for defining a market.
Hidden Multipliers - Jason Cohen - Cohen's forthcoming first book (pre-order at hiddenmultipliers.com), about small decisions - like moving churn from 5% to 4% - that have outsized impact on revenue and profit.

Media referenced

ER - show - Cohen's favorite recent watch - rewatching the 1994 medical drama's 15 seasons with his 16-year-old daughter, who says it still holds up.
The Pit - show - Lenny's recommendation for ER fans - a recent award-winning medical drama on Netflix.
Ecosystems as a growth channel - article - Emily Kramer piece referenced for ideas on growing through influencers, content, and partners rather than paid/organic channels.
The elephant curve - article - Cohen's own smartbear.com post naming the pattern where growth channels don't plateau into a stable S-curve but eventually sag and decline.
Post on ChatGPT's App Store as a growth channel - article - Upcoming Lenny's Newsletter piece flagged as a potentially large new distribution channel via OpenAI's app submissions.

Companies

WP Engine - Cohen's most recent company, a unicorn, sold heavily through web-design agencies rather than direct - his own example of a channel-diversification bet.
Smart Bear - Cohen's ~20-year blog (asmartbear.com) where he has written 150-350 posts, all free, with no ads or courses.
Groove - Case study Cohen cites: switching the cancellation-survey question from 'why did you cancel' to 'what made you cancel' roughly doubled usable response rates (10% to 20%).
Double Down - Anonymized/illustrative company in Cohen's pricing story - halving a customer's ad spend could be priced at $5K/month framed as 'savings,' or $40K/month (8x more) framed as 'doubling your leads,' for the same underlying product.
Buffer - Example of a company that deliberately declined to move upmarket into enterprise social tools because it conflicts with its identity as a company for individuals and small teams.
Constant Contact - Restarted growth for its small-business email marketing tool by physically running in-person workshops in cities, teaching restaurateurs and dentists email marketing.
HubSpot - Famously tested selling through agencies instead of only direct; agencies grew to roughly 50% of revenue within four to five years.
Duolingo - Cited as a consumer example of successful net revenue retention/upsell design - many small ways to pay more (gems, cosmetic upgrades).
AG1 - Consumer example of expanding revenue per customer via a second product (a sleep supplement) sold to the same existing base.
AT&T - Negative consumer example - a company Cohen doesn't want to spend more with because it isn't delivering more value in return.
Amazon - Positive consumer counter-example to AT&T - a company where increased spend does correspond to increased value delivered.
Rippling - Referenced via a prior podcast guest, CPO Matt McGinnis, for advice on when to quit a struggling startup.
Shopify - Referenced via a prior podcast guest (its CTO) - even a 100-person, statistically sophisticated experimentation team found roughly a third of their 'winning' A/B test effects disappeared on holdout re-checks.

Techniques and frameworks

The 5-question growth-stall diagnostic - Cohen's core framework, applied in strict order: (1) are customers leaving/logo churn, (2) is pricing correct, (3) are existing customers growing (NRR), (4) are acquisition channels saturated, (5) do you actually need to grow. Fixing a lower-order question doesn't help if a higher one is broken.
Churn ceiling math - Maximum sustainable customer count = new customers added per period divided by the cancellation rate, because cancellations grow proportionally with company size while marketing-driven acquisition does not.
Proximate cause vs. deeper cause - Cohen's medical-examiner analogy for cancellation reasons: the stated reason ('too expensive,' 'project ended') is like 'stopped breathing' - useful, but you have to dig to the actual underlying cause to find something actionable.
'What made you cancel' framing - Phrasing cancellation surveys around what happened rather than why, to avoid inviting a canned, socially acceptable excuse like 'budget.'
Net revenue retention (NRR) vs. logo retention - NRR blends upgrades against cancellations/downgrades into one ratio and can mask a shrinking customer base; Cohen argues it must be tracked alongside logo (customer-count) retention, not instead of it.
The elephant curve - A growth-channel pattern where results look like a rising S-curve but then sag and decline over time as the audience saturates or the channel itself weakens, rather than plateauing indefinitely.
Sell what the company already values - Pricing/positioning principle: frame the same product benefit in terms of the metric the buyer's organization already prioritizes (e.g., growth) rather than in terms of cost savings, to expand willingness to pay.

Summary

Jason Cohen, founder of four companies including two unicorns (most recently WP Engine), lays out a strict, ordered diagnostic for when a product's growth stalls - and argues the same sequence doubles as a playbook for accelerating growth that hasn't stalled yet. The first and most important question is whether customers are leaving. Cohen reduces this to a stark formula: the maximum number of customers a company can ever have equals new customers added per month divided by the monthly cancellation rate, because cancellations grow proportionally with company size while marketing-driven acquisition typically doesn't. He walks through how to actually diagnose churn - asking "what made you cancel" instead of "why did you cancel" (which roughly doubled usable responses in a Groove case study), and pushing past superficial reasons like "too expensive" or "project ended" to find the deeper cause, using a medical-examiner analogy about proximate versus underlying causes of death.

The second question is whether pricing is correct - and Cohen argues almost everyone's answer is no, because founders typically guess a price early and never revisit it. He challenges the standard demand-curve intuition that raising prices reduces signups, describing cases (including a 12x price increase for an enterprise product) where signups didn't change or even increased, because price also signals quality and market fit. His "Double Down" thought experiment illustrates how framing the identical product around what a buyer's organization already values (more growth) instead of what it saves (lower cost) let a hypothetical vendor charge roughly 8x more for the same value delivered.

Question three is whether existing customers are growing their spend - net revenue retention (NRR). Cohen argues NRR is necessary but insufficient on its own: because percentage losses and gains aren't symmetric (a 20% loss needs a 25% gain to break even), and because NRR says nothing about whether enough logo-level customers remain, it must be tracked alongside plain customer-count retention. He notes virtually no large public SaaS company survives with NRR below 100%, with a median around 119% at IPO.

The fourth question addresses acquisition-channel saturation, which Cohen names the "elephant curve" - channels don't plateau into a stable S-curve, they eventually sag and decline as the same audience gets repeatedly exposed and channel effectiveness itself erodes. Escaping this usually requires a genuinely new channel type rather than incremental optimization, illustrated by Constant Contact's in-person small-business workshops, HubSpot's agency-partner channel (roughly 50% of revenue within a few years), and WP Engine's own agency-driven distribution. The fifth and final question is more existential: does the company - or the founder - actually need to keep growing at all? Cohen argues some bootstrapped businesses are legitimately fine choosing stability or profit over growth, though he suspects the old adage "if you're not growing, you're dying" applies more reliably to the person than to the company.

In shorter closing segments, Cohen shares that he finds AI (especially Gemini) useful for converting charts and images into structured data he can analyze, but argues LLMs are good at summarizing themes in qualitative feedback and bad at surfacing the specific actionable detail that should actually change a roadmap. In a contrarian-corner segment, he argues A/B testing is largely a waste of time below meaningful traffic scale and cannot validate strategic decisions, citing a Shopify example where a large, sophisticated experimentation team found roughly a third of their "winning" tests were false positives on holdout re-checks.

Notable Quotes

"Your prices are way too low because you just guessed and you haven't changed them." - Jason Cohen (quoting Patrick Campbell)

"Cancellations grow faster than marketing... cancellations overpower the growth of the company and slow it to a halt." - Jason Cohen

"AI is good at picking out themes. It is bad at picking out details that are actionable." - Jason Cohen

"If you don't know who the patsy is [at the A/B testing poker table], it's you." - Jason Cohen

"How do we create more value for the customer and then split that with them?" - Jason Cohen