13 new growth playbooks, each with its number. The biggest moves changed the offer, not the funnel.

In September we published 32 growth playbooks. The lesson then was that the page that wins is the one a machine can quote.

The library has kept growing, and 13 of its new entries are here. Each one names a company or a person, gives a number, and links to a source you can open.

The first edition was mostly about getting found. The new entries are mostly about something else.

Pricing and business model changes that moved revenue: what the new entries show

Six of the 13 new entries got their number by changing the offer: whether paying is optional, what the price is, when you can buy, who the product is for, or what the answer costs to deliver. The product and the traffic were mostly left alone. A seventh entry marks where that move stops working.

The cleanest pair sits at opposite ends of the market: a $9 membership and an enterprise plan.

Email Marketing Membership has cost $9 a month since February 2021, and the price will not move. "I will never raise the price of EMM," Liz Wilcox says.

She puts revenue at "about $40-something thousand" a month, or about half a million dollars a year. On air, host Jay Clouse worked the monthly figure out to about 4,400 paying members, and she agreed. She puts her churn rate at about 5%. There is one product, with no upsell, no mastermind and no coaching.

Her buyer is someone looking at $9 and thinking "this might be all I have to spend this month." At that price, a bad week is no reason to cancel.

The referral offer is priced for the same buyer. Bring two friends and your membership is free. She pays affiliates 50%, or $4.50, and the affiliates are her own customers. New members come from those referrals and from "almost 400 podcasts and virtual stages" where she has been a guest (Liz Wilcox, on Creator Science).

At the other end, 54% of Anthropic's new enterprise logos in 2026 came through self-serve. The commercial team came back from the 2025 winter break to demand far past what it had hired for. Its four constraints included headcount it could not add fast enough without lowering the recruiting bar.

So in January 2026 Anthropic launched an enterprise self-service MVP. It went to production in February. The buyer lands on an enterprise plan with real ACV, terms of service, invoicing, provisioning and training enrollment. Intercom's Fin, a support product, was retooled to guide the purchase.

Four months after the January launch, 54% of new enterprise logos that year had come through the self-serve funnel (Eleanor Dorfman, Anthropic, written up by Jason Lemkin at SaaStr). The figure counts logos, not revenue.

Neither of them bought more traffic. Wilcox wrote a price her buyer never has to decide on again. Anthropic made its largest plan something a buyer can finish alone, and it is a real plan, not a starter tier.

The offer is a decision about who is buying. Get that decision right and the funnel you already have converts more people.

Superhuman's numbers say the same thing from the product side. In Rahul Vohra's 2018 account, 22% of surveyed users said they would be "very disappointed" without the product. When he narrowed the count to the personas inside that group, the score was 33%, with no product change. Three quarters of roadmap work later, it was 58% (Rahul Vohra, First Round Review). Choosing who the product is for moved the number before the product did.

Where it stops working. RevenueCat's 2025 report put "not enough usage" at 37% of subscription cancellations, ahead of price at 35%. Daphne Tideman's advice is to find which habit broke, not to cut the price.

Her evidence is one client. The cohort that signed up on a 50% discount had a £50 lower lifetime value and higher churn than full-price sign-ups (Daphne Tideman, RevenueCat). If people leave because they stopped using the product, a lower price brings back the wrong people.

The check takes ten minutes. Write down what a customer must do to pay you. Is paying optional? Is the price built for the buyer you actually have? Can a serious buyer finish the purchase without waiting for a human? Then look at how your last ten cancellers used the product in their final month. If they had stopped using it, the price is not the problem.

Two samples, so you can see what the entries look like

AstroSage serves the same buyer the same question at nearly 90% margin, against 45 to 55% for human marketplaces. Astrology marketplaces connect users with human astrologers and take a commission. Punit Pandey, of AstroSage, puts that model at 45 to 55% gross margin.

AstroSage answers with in-house AI astrologers instead, at "nearly 90% margins." AI revenue has grown about 20% month on month for 18 consecutive months. Users rate the AI astrologers 4.6 out of 5, against 4.3 for its human astrologers (Punit Pandey, AstroSage).

The buyer and the question did not change. What changed is what it costs to deliver the answer.

PMM Camp opens its cart on the 1st to the 5th of every month, and never converts below 30%. Tamara Grominsky runs an application-only membership for senior product marketers, with 280+ members. The sales page always says "join the waitlist." Approved applicants get four emails over five days at the start of each month. One of those emails never changes. She rewrites the other three every month.

Anyone who does not buy is taken off the waitlist about five days later and invited to re-join. She found that 95% of buyers buy in their first open cart, so the removal costs little and gives her a real denominator. First-cart conversion is about 30%: "I never drop below 30." One month hit about 80%.

The counter-example is on the record. In fall 2025 she switched to evergreen enrolment, the always-open model. "I completely broke my business… no one was joining." She went back to monthly windows in December. She now forecasts revenue twelve months out (Tamara Grominsky, on Creator Science).

Family 1: The offer (new entries)

Every entry gives the number in the source's own terms, the mechanism you can copy, and the caveat where a figure is self-reported or old.

1. Email Marketing Membership: $9, never raised, customers as the sales force. The details are above (Liz Wilcox, on Creator Science). Her own math before she started: 1,000 members at $9 is $9,000 a month, and she gave herself two years to get there. Caveat: the 4,400 is Jay Clouse's arithmetic on air, which she agreed to. The monthly and yearly revenue figures are hers. Check: price your referral reward in what your customer wants, not in cash.

2. Anthropic: 54% of new enterprise logos, self-serve. The details are above (SaaStr). The rest of the rebuild kept the stack and threaded Claude through it: Clay, LeanData, Salesforce, Gong, Ironclad and Slack. A "call prep" Skill replaces 30 minutes of research with a one-pager. Claude surfaces six coaching moments per rep per week. SaaStr's summary: "almost none of it required new software." Caveat: forecasting is the stated weak spot. The team still spends "at least 10 minutes at the top of every forecast call discussing how they should be forecasting." Check: can a buyer at your largest plan size finish the purchase without a call?

3. AstroSage: same buyer, same question, a different margin. The details are above (Punit Pandey, AstroSage). The questions are narrow and they repeat. At Astrotalk, an Indian marketplace that takes a cut of each consultation, the average consultation is around ₹210, and close to 65% of revenue comes from marriage questions (Inc42). Caveat: every AstroSage figure is the company's own, from a February 2026 post. Check: in a per-minute expert business, count how many questions are repeats.

4. PMM Camp: a monthly window, a purge, a forecast. The details are above (Tamara Grominsky, on Creator Science). Jay Clouse's own membership, The Lab, began testing the same monthly model in September 2026 after watching hers. Caveat: nine months of data, all hers. Check: if your membership is always open and lacks top-of-funnel volume, try one closed window and count the denominator.

5. EmailEngine: the same software, with paying made required. For his first eighteen months, Andris Reinman sold an optional commercial licence: in his words, "a solution to a licensing problem that most of my potential customers did not actually have." Then he made a paid licence key required, kept the €250 price, and made $2,000 the first month. He has since raised the price to $1,450 a year, and he says the increases "did not seem to reduce the number of new customers." It is now a $20k MRR solo business (Andris Reinman). We told the full story in a separate post. Caveat: he had the idea in 2013 and started the company in 2020. He credits a decade of open-source credibility, and luck. Check: if people already run your free version at work, ask what they would lose if it stopped.

6. Superhuman: choose who it is for, then build. The survey asks four questions. How would you feel if you could no longer use it? What type of people benefit most? What is the main benefit? How can we improve it? Segment to the "very disappointed" group and the personas in it. Spend half the roadmap on what they love. Spend the other half on what holds back the "somewhat disappointed" users who value the same thing. Survey new users continuously, never the same person twice. Score: 22% to 33% by segmenting, then 58% within three quarters (Rahul Vohra, First Round Review). Caveat: the account is from 2018. It is new to this library, not new. Check: run the four questions on users active in the last two weeks.

7. RevenueCat: when the problem is usage, a discount makes it worse. "Not enough usage" was 37% of cancellations in the 2025 report. In the 2026 report it is the second-biggest driver, at 26 to 40% depending on category. The diagnosis runs in four steps:

1. Define your "enough": what long-term subscribers repeat in their first billing cycle.
2. Look at usage in the 30 days before cancellation.
3. Segment churners by usage frequency.
4. Compare cancellation reasons across price tiers.

Then find which of four breaks you have. Either the trigger never fired, the action never stuck, the reward went flat, or the value never landed. "Not one of them responds to a discount." One targeted re-engagement email sequence drove a 100%+ lift in users completing a workout within three days, against a no-email control (Daphne Tideman, RevenueCat). Caveat: the £50 LTV and the 100% lift are single, unnamed clients. Check: before any winback discount, look at the churners' last month of usage.