Jason Cohen says monthly new customers divided by cancellation rate sets a company's growth ceiling, making churn the worst problem.
Cohen says "it's too expensive" is never the real reason customers cancel, and "what made you cancel?" doubled Groove's usable responses.
Cohen says pricing is almost always too low, citing a founder who moved from $300 a year to $300 a month without losing signups.
How to diagnose a product that has stopped growing
Growth rarely stops all at once. It slows, and the reasons are diagnosable, so Jason Cohen works through the questions he asks in order when a product stalls. Churn comes first because cancellations compound with company size and set a hard ceiling on how big a business can get. Then come pricing, expansion revenue and channel saturation, each with its own test. The last question asks whether the company needs to grow at all, which Cohen treats as a serious option rather than a joke.
Takeaways
Churn caps a company's maximum size. Maximum size equals new customers per month divided by the cancellation rate. At 100 new customers a month and 5% churn, the ceiling is 2,000. Cohen calls anything above 3% monthly cancellation terrible, and notes cancellations scale with company size while marketing does not.
Ask what made customers cancel. Groove doubled usable cancellation responses from 10% to 20% by switching the wording. Cohen says 'too expensive' is a proximate cause, like a cause of death, and rejects a single root cause. Since most churn happens early, onboarding is a good default bet.
Prices are almost always set too low. One founder went from $300 a year to $300 a month with no drop in signups. Price selects the market, since a $100 product looks too cheap to a company with $400m in revenue. Enterprises buy at $75,000 to $150,000. Structure matters as much as the number.
Expansion revenue must offset churn. A 20% revenue loss needs a 25% gain to recover. Only about two of more than 100 public SaaS companies have net revenue retention below 100%, and the median at IPO is 119%. Referrals from existing customers also scale with company size, offsetting cancellations.
Channels follow an elephant curve. A channel climbs an S-curve, saturates and then sags. HubSpot tested selling through agencies instead of direct, and that channel reached 50% of revenue within four or five years. Cohen says people nod at the advice to move on and then do not act.
Growth is not always the goal. Cohen questions the line that if you are not growing you are dying, calling it something investors use to push founders when they should not. He cites 37Signals and says he wants to grow only within work he is proud of.
Most A/B tests are a waste of time. Cohen argues testing cannot settle strategy, since you cannot A/B test whether Uber is a good idea. He also says most positive results turn out to be false positives.
Tap a name to go there · tap a line to see why they are linked