Cohen argues that "too expensive" is almost always the top-cited cancellation reason and almost never the true one. A customer who found you, survived the homepage, cleared the pricing page and bought already decided the price was acceptable. Something else broke the promise, and blaming price lets you avoid finding it.

Buying is itself proof the customer didn't think the price was too high
Losing budget is a different reason than being too expensive
Citing price as the cause is a fallacy that stops you from finding the real breakage (e.g. a missing integration)
Dig past the 'approximate cause' the way a doctor digs past 'stopped breathing'

“And that is never ever ever the reason. How do I know? Because they already looked at your homepage, read all the stuff, saw what…”
Jason Cohen · 20:30

#churn#pricing#customer-research

Myth Buster36:00

Raising prices usually doesn't lower signups

The textbook demand curve says higher prices mean fewer signups, but Cohen says in the real world raising prices often leaves signups unchanged or higher. Founders under-price because they guessed and never revisited it, quoting Patrick Campbell. Fear of raising prices is usually unfounded.

Patrick Campbell (4,200 startup data points): prices are too low because you guessed and never changed them
Microeconomics 101 demand curve rarely describes real SaaS pricing
Raising prices frequently keeps signups flat or increases them
Founders avoid raising prices for emotional, not economic, reasons

“Your prices are way too low because you just guessed and you haven't changed them.”
Jason Cohen · 36:00

“So what usually h what often happens is you raise prices and signups don't change.”
Jason Cohen · 37:00

#pricing#growth#positioning

Hot Take· 4

Hot Take22:00

There's no such thing as a single root cause

Cohen rejects the idea of 'root cause' analysis, arguing complex systems fail through many interlocking factors, not one. Using a healthcare death-certificate analogy, he shows a chain of causes (stopped breathing back to undiagnosed diabetes) and says a good analysis is an array of interventions, not one culprit.

Complex systems have many interlocking causes, each a possible point of intervention
The 'approximate cause' (stopped breathing) hides the real driver (undiagnosed diabetes)
Even 'five whys' can be too simplistic if it implies one cause sits at the bottom
An analysis should produce a set of things to fix, not a single root cause

“Another as a sidebar, I hate the idea of a root cause. Complex systems do not have one root cause.”
Jason Cohen · 22:00

#decision-making#diagnosis#mental-models

Hot Take30:30

AI is an averaging machine, so it misses the actionable details

Cohen warns against letting an LLM do the thinking on customer-feedback synthesis. Because an LLM predicts the most likely output, it's good at themes and summaries but bad at surfacing the specific, non-average details that actually trigger action. His workaround: ask for themes, then force it to list every specific detail under each theme with its source.

LLMs excel at averages: summarization, topics, themes
They're weak at 'what is interesting and not average', which is what's actionable
The predictable themes ('couldn't figure out the integration') aren't the surprises; the details are
Workaround: extract themes, then have it list every detail with the customer and link so a human can spot the trigger

“the LLM is an averaging machine, right? It's predicting the most likely that's an averaging kind of a thing.”
Jason Cohen · 30:30

“But when you're asking for like what is interesting and not average, it's actually pretty bad at it.”
Jason Cohen · 31:00

#ai#customer-research#llms

Hot Take1:20:30

Do you actually need to grow, or is that just the investors talking?

Cohen's final and most philosophical question challenges the mantra 'if you're not growing, you're dying'. For a bootstrapped founder paying themselves well, stasis can be fine, and growing might mean a company or market they don't want. But he reframes the phrase as being about you the person: for the kind of people who build things, stagnation really can feel like dying.

'If you're not growing, you're dying' may be a line investors use to push growth that isn't warranted
Bootstrapped firms (e.g. 37signals) can rationally maximize profit over revenue and stop growing
Reframe the phrase as personal: are you, the person, still growing and fulfilled?
If every other option is exhausted and it's truly stagnant, something dramatic may need to change

“Is that true or is that the kind of thing that like investors use to like make founders grow or try to grow even when…”
Jason Cohen · 1:20:30

“what I like to say is maybe the you and if you are not growing, you're dying is you the person”
Jason Cohen · 1:22:30

#strategy#founder-mindset#growth#bootstrapping

Hot Take1:34:30

A/B testing mostly doesn't work

In the contrarian corner, Cohen argues A/B testing fails on anything important, such as strategy, vision, or insight, and that most 'wins' on the small stuff are false positives. When the real effect is rare and tools aren't statistically perfect, false positives outnumber true ones, so stacked 'winners' often add up to no real improvement. He cites Shopify's team finding a third of effects vanish on re-check.

You can't A/B test strategy or vision; you don't A/B test whether Uber is a good idea
Stacked 'winners' often produce no measurable improvement a year later
When the sought effect is rare, false positives outnumber true positives even at 95% accuracy
Even Shopify's sophisticated team sees ~a third of effects disappear on hold-out re-checks
It can pay off at massive scale where a fraction of a percent is millions of dollars

“AB testing doesn't work very well and it doesn't work on most things. It won't work on strategy or vision or insights like nothing actually…”
Jason Cohen · 1:34:30

“it's sort of like the poker table. If you don't know who the psy is, it's you, right?”
Jason Cohen · 1:36:00

#experimentation#ab-testing#data#product

Explainer· 3

Explainer16:30

The math that caps how big your company can ever get

Jason Cohen explains why cancellations, not marketing, set a hard ceiling on company size. Marketing delivers a roughly fixed number of new customers a month, but cancellations are a percentage of your whole base, so they grow automatically as you grow. The ceiling is simply new customers per month divided by the cancellation rate.

Marketing grows linearly and doesn't care how many customers you already have; cancellations scale with your size
Max customers = new customers per month / monthly cancellation rate (e.g. 100 / 5% = 2,000)
As you approach that ceiling, growth stalls because nearly as many leave as you add
The same math works for revenue churn (dollars in / dollar cancellation rate), not just logo churn

“So let's suppose you add 100 customers a month and you have 5% cancellation. So 100 divided by 5% is 2,000. So a company like…”
Jason Cohen · 16:30

“Whereas, cancellations grow automatically as you grow, right? So, cancellations always overtake marketing for this reason.”
Jason Cohen · 15:30

#churn#saas-metrics#growth#retention

Explainer54:30

Why NRR flatters you and logo churn keeps you honest

Cohen shows that net revenue retention over-counts because percentage losses and gains aren't symmetric. Like a stock that drops 20% then rises 20% and lands at 96, a 20% cancellation offset by 20% in upgrades doesn't net to even. That's why he still watches logo churn: NRR can look positive while too few customers remain to matter.

A 20% loss requires a 25% gain just to break even
NRR treating a 20% cancellation as offset by 20% upgrades is wrong; it only reaches 96%
NRR can be positive while logo churn quietly shrinks the base too fast
Median NRR at SaaS IPO is ~119%; only ~2 of 100+ public SaaS companies are under 100%

“So if it goes down 20% and up 20%, it does not come back to zero. It's worse.”
Jason Cohen · 54:30

“This is why NR isn't quite right because NR is saying that a loss of 20% from cancellations is offset by 20% from upgrades. As…”
Jason Cohen · 55:00

#saas-metrics#nrr#retention#churn

Explainer1:08:00

The elephant curve: why growth channels sag, not plateau

Cohen rejects the tidy S-curve story of a channel maturing and flattening. In reality a channel starts as an S-curve and then its 'butt' sags downward as the audience saturates and the channel declines, a shape he named the elephant curve. Circulation figures, conference attendance, and ad channels all quietly decay before anyone admits it.

Channels don't plateau, they sag: an S-curve that then bends down
Audiences saturate (people have seen it enough) and channels decline over time
Vendors hide the decline: magazines report rising circulation right before folding
You can't rely on marketing forever, since channels are finite and decaying

“What happens is it starts with an S-curve and then it starts sagging. Its butt starts sagging down. So, I wrote an article about this…”
Jason Cohen · 1:08:00

#marketing#channels#growth#saturation

Story· 4

Story19:00

The cancellation dropdown that was pure noise

At Smart Bear, Cohen noticed one cancellation reason was picked far more than the rest and suspected it was just the first item on the list. When he randomized the order for every user, all options got picked equally, revealing the data was meaningless. He notes other companies have found the same thing.

The most-picked reason was simply the one at the top of the list
Randomizing option order flattened all responses to equal, exposing the bias
Fixed-order multiple-choice cancellation surveys generate noise, not insight
This is a global phenomenon other companies have reproduced

“So then then we randomized the list so everyone saw a different order of the list and now all the items were picked equally like,…”
Jason Cohen · 19:00

#customer-research#surveys#churn

Story38:00

He 12x'd the price and nothing changed

Cohen tells of a founder selling to enterprise and government for $300 a year, which he judged far too cheap. On a dare, they switched the price from per-year to per-month, a 12x increase, and signups stayed at one or two a week. Cohen's point: nothing observable changed, so the founder still wasn't near the right price.

Product sold to enterprise/government at $300/year got one or two signups a week
Switching per-year to per-month 12x'd the price with no drop in signups
The founder's instinct was to spend the new profit; Cohen said raise prices again
No observable change means you haven't found the price ceiling yet

“So, in other words, we're 12xing the price, right? So he did and he still got one or two per week. Like nothing changed.”
Jason Cohen · 38:00

“Like you just told me you you 12xed the price and nothing observable changed. That means you're not near the price yet, right?”
Jason Cohen · 38:30

#pricing#enterprise#growth

Story42:00

How to charge 8x for the same product by repositioning it

Cohen's 'Double Down' parable: a tool that halves your AdWords cost can only justify a small fee because it's framed as savings. Reframed as 'double your leads for the same spend', the same product justifies 8x the price, because a CMO would rather tell the CEO they grew leads than saved money. The lesson: sell more of what the buyer values, usually growth, not savings.

As a cost-saver, the tool captures only a slice of the saved money (e.g. $5k to save $20k)
Reframed as doubling leads at the same CAC, the buyer will pay the full $40k
Same product, same math, but 8x the revenue purely from positioning
Pricing is structure and positioning, not just the number on the page

“how this company was able to charge eight times as much for the same product just by talking about it differently.”
Jason Cohen · 42:00

“the big lesson for product managers is sell more of what the company values like growth.”
Jason Cohen · 46:30

#pricing#positioning#sales#growth

Story1:13:30

Reigniting growth with a genuinely new channel

When channels saturate, incremental tweaks won't help; you need a genuinely new channel. Cohen cites Constant Contact physically running email-marketing workshops in cities to sign up small businesses, and HubSpot testing agency sales that grew to half their revenue. WP Engine similarly sells many sites through WordPress agencies.

Constant Contact ran in-person city workshops for restaurants and dentists to restart growth
The tactic seemed uneconomical for a $20/month product but was highly effective
HubSpot tested selling through agencies, which became ~50% of revenue in four to five years
WP Engine sells heavily through WordPress-building agencies as a channel

“HubSpot famously, uh, tested selling through agencies instead of direct. It ended up being 50% of the revenue after four or five years.”
Jason Cohen · 1:14:30

#channels#growth#distribution#go-to-market

Tool· 1

Tool19:30

Ask "what made you cancel?" not "why did you cancel?"

A small wording change on a cancellation survey materially improves response quality. Asking 'why did you cancel' invites a one-word answer like 'budget'; asking 'what made you cancel' pulls out the actual situation. Cohen credits a Groove case study where the same email doubled usable responses.

Use open-ended questions, not a dropdown, because dropdowns invite lazy answers
'What made you cancel' surfaces the product/situation cause; 'why did you cancel' invites a simple excuse
Groove's email went from 10% to 20% usable responses on this change alone
Most people won't answer at all, so extract as much signal as you can from those who do

“What you want to do is say what made you cancel? In other words, what about the product or situation or whatever caused the cancellation?…”
Jason Cohen · 19:30

“They got 10% usable responses. They changed it same email to why what made you cancel and it's 20% usable responses.”
Jason Cohen · 20:00

#customer-research#churn#surveys

Takeaway· 1

Takeaway27:00

When in doubt, fix onboarding first

Most cancellation happens in the earliest part of the customer lifecycle, and small onboarding changes have outsized effects there. Cohen uses the YouTube retention curve as an analogy: shifting the steep early drop-off a few points cascades into a much larger gain at the end. Early churn is also the most unprofitable, since you never earn back acquisition cost.

Far more cancellation occurs in the first day/30/90 days than the rest of the customer's life
Small onboarding tweaks have large effects early; later-lifecycle tweaks often don't
YouTube analogy: cutting a 50% early drop to 45% can lift end-retention by 20-30%
Early churn is especially costly because acquisition spend never pays back

“Um all almost all companies have a whole lot more uh cancellation in the first day, 30 days, 90 days, depends, right? But the first…”
Jason Cohen · 27:00

#onboarding#activation#retention

Frameworks in this episode

Net Revenue Retention Above 100% (and Why NRR Lies)

Pricing Selects the Market (Sell Growth, Not Savings)

The 5-Question Stalled Growth Diagnostic

The Cancellation Ceiling (Maximum Customer Cap)

The Elephant Curve (Channels Saturate and Sag)

The 'What Made You Cancel?' Rooter-Cause Diagnosis

Insights & moments

⚡Myth Buster· 2

"It's too expensive" is never the real reason people cancel