The thinking behind the work

Dick Fosbury's flop only worked because of the foam pit underneath. Growth playbooks work the same way: what you can't see is the retention that pays for them. What to check before you copy.

Watt's efficient steam engine made Britain burn more coal, not less. AI will do the same to marketing: cheaper creative means far more tests, and every test needs media spend. How to budget for it.

In 1936 a magazine polled 2.4 million Americans and called the election wrong. Your NPS, reviews and surveys have the same flaw: they only hear from people who stayed. Who to ask instead.

Napoli won two titles with Maradona, then waited 33 years for the next. Growth that lives in one media buyer, one creator or one founder leaves when they do. How to check and fix it.

A Coke cost 5 cents from 1886 to 1959, held there by a contract and vending machines, not by customers. Many subscription apps are priced the same way. How to find what holds your price and test a new one.

Leicester won the 2016 Premier League with 42.4% possession, the lowest of any champion. Marketing dashboards full of reach, impressions and clicks measure possession. What to put at the top of the report instead.

Two clubs passed on a 13-year-old Messi over a monthly medical bill. Barcelona bet on the early signal. How growth teams can act on a promising first cohort with predicted LTV, then prove it.

AC Milan hired Arrigo Sacchi, a former shoe salesman, and won two European Cups. Marketing hiring overweights logos and certificates. The questions that reveal how a candidate decides what to test, kill and scale.

Every big club scouted the same academies and missed Jamie Vardy on a factory shift. Most ad budgets compete for the same audiences on the same three platforms. Where to look for cheaper users who still pay back.

Klopp's counter-press wins the ball the second it is lost. The best moment to win back a churned subscriber or an abandoned cart is the same: right away. How to build a winback that triggers on the exit.

After a decade as a CMO, the best decisions were the things we did not do. Five no's that protect a growth budget, why each one matters, and how to write them into your own plan.

A mobile game asked to double its UA budget. The audit said stop. How a D30 to D90 retention leak was found and fixed, and daily revenue rose 15.8% on the same spend.

Most teams kill campaigns on gut feel and week-old data. A three-tier rule based on predicted D90 ROAS from the first three days, written before a single dollar moves.

A growth audit checklist for consumer apps and games. Score positioning, retention, unit economics and channel separately, with pass and fail lines, and always check channel last.

Margin turns positive and every founder wants to go all in. Three conditions a channel must pass before it gets more budget, including enough conversions and your real breakeven ROAS.

The classic D1 40%, D7 20%, D30 10% soft-launch gate matches the top 10% of mobile games, not the minimum. What the 2026 GameAnalytics data shows, and a better kill-or-scale framework.

The best-looking channel in a games studio's media plan turned out to be buying installs that would have come anyway. How a geo holdout test found it, and why cutting it raised profit 42%.

In 1902 Hanoi paid a bounty per rat tail and ended up with more rats. Last-click attribution works the same way. How to spot channels that clip credit from sales that were already coming.

Most founders don't have a marketing problem. They have a retention problem their ad budget is hiding. Why the first session decides, what the data says, and what to check before spending more.

Calling an A/B test early is a coin toss with a dashboard. How a tutorial test in a mobile game lifted ROAS 14.5%, and the rules that make a result survive past the screenshot.

In 1693 Edmond Halley priced life insurance from early data. Most growth teams wait 12 months to trust LTV. Why day 7 to day 30 predicted LTV should drive budget decisions.

Your product, pricing and marketing already assume an ideal customer. If you haven't defined that person, check how much they look like you. A practical way to define an ICP for a consumer app.

Pareto counted land deeds in 1896. The 80/20 rule is an observation about uneven outcomes, not a slogan. How to verify your real 20% of campaigns, customers and channels, and cut the rest.

The more dashboards a company has, the fewer decisions it makes. Why reports turn into "let's keep watching", and how a decision rule written before the number arrives fixes it.

Flat revenue looks the same whether the cause is marketing, product or broken measurement. The order to check them in, and how to tell which one you have.

A high blended ROAS usually means you are underspending. How to find marginal ROAS, your real breakeven line, and when not to scale.