What founders, investors and authors say works — summarised in plain words, linked to
the source, and tried on the businesses we run. When enough of them have tried it, we say whether it worked.
121 classics · all free0 being tested0 with a verdict5 businesses before a verdict
Paul Graham: Co-founded Y Combinator, the startup program behind Airbnb, Dropbox, Stripe and Reddit.
“A good growth rate during YC is 5-7% a week.”
What makes a company a startup is that it is designed to grow fast. Choose a weekly growth target (revenue, or active users if not yet charging) and judge every decision by whether it helps hit that number.
When it applies: You need one number to choose between competing tasks, and the business is meant to scale rather than stay a steady small operation.
Contested. A single growth number invites gaming (Graham himself rules out buying users above their lifetime value or counting inactive users as active) and can crowd out profitability, quality and customer care. Many sound small businesses are not built to grow this fast and should not be judged this way.
Dave McClure: Co-founded startup investor 500 Startups, now 500 Global; earlier a PayPal marketing director.
“users like product enough to refer others”
Track five stages of the customer lifecycle, each with its own conversion rate: people arrive, have a good first experience, come back, tell others, and pay. Find the stage that loses the most people and fix it before buying more traffic.
When it applies: A product is live with some traffic and the business cannot tell where people drop out.
Elon Musk: Runs Tesla and SpaceX; co-founded the company that became PayPal.
“all designs are wrong, it’s just a matter of how wrong.”
Improve any process in a strict order. First challenge each requirement and find out who set it. Then remove every step or part you can, and only after that simplify what is left. Speed it up next, and automate last. Out of order, you waste effort perfecting or automating things that should not exist.
When it applies: A process such as onboarding, checkout, fulfilment or content production has piled up steps and nobody can say why each one is there.
Contested. The rule of deleting until some steps have to be put back means breaking things on purpose. Applied to legal, consent, tax, refund or safety steps, that creates real liability, so those requirements need protecting. Musk's management style is itself widely disputed, which colours how the method is received.
Not tried yet
None of our businesses has tried it yet.
Stay in Day 1: decide fast with about 70% of the information
Jeff Bezos: Founded Amazon in 1994 and grew it from an online bookstore into a giant.
“most decisions should probably be made with somewhere around 70% of the information you wish you had.”
Companies decline once they slow down, which Bezos calls 'Day 2'. To stay in 'Day 1', treat most decisions as reversible ('two-way doors') and make them quickly with incomplete information, then correct fast. Keep slow, careful process for the few decisions that cannot be undone.
When it applies: A business keeps putting off changes that could easily be reversed, such as copy, offers, channels or features.
Not tried yet
None of our businesses has tried it yet.
The flywheel: momentum from consistent pushes in one direction
Jim Collins: Wrote Good to Great, one of the best-selling business books of all time.
“relentlessly pushing a giant, heavy flywheel, turn upon turn, building momentum until a point of breakthrough, and beyond.”
Lasting improvement rarely comes from one big move or launch. It comes from many consistent pushes in the same direction that add up over time. Companies that change direction after each disappointment fall into what Collins calls a 'doom loop' and never build momentum.
When it applies: A business is tempted to switch strategy, channel or product after a few weeks of slow results.
Not tried yet
None of our businesses has tried it yet.
High Output Management: pair every metric with a counter-metric
Andrew S. Grove: Led Intel as it became the world's biggest chipmaker; Time's 1997 Man of the Year.
Grove treats management like running a production line. You find the slowest step, watch a few early-warning indicators, and judge a manager by the output of the whole team. Because any measure pulls attention toward itself, he pairs each one with a second measure that catches the side effect, such as output with quality.
When it applies: A business is pushing hard on a single number, such as signups, sales, tickets closed or articles published.
Not tried yet
None of our businesses has tried it yet.
John Doerr: Kleiner Perkins investor who backed Google and Amazon early and brought OKRs to Google.
Andrew S. Grove: Led Intel as it became the world's biggest chipmaker; Time's 1997 Man of the Year.
“Key Results benchmark and monitor how we get to the Objective.”
Set a few ambitious objectives each quarter, and give each one a handful of specific, time-bound, measurable results that show whether you got there. Make them visible, check progress often, and grade them honestly at the end.
When it applies: A business has many things it could work on and no shared way to decide what matters this quarter.
Not tried yet
None of our businesses has tried it yet.
Mark Zuckerberg: Co-founded Facebook and runs Meta, which also owns Instagram and WhatsApp.
“The idea is that if you never break anything, you’re probably not moving fast enough.”
Shipping fast and learning matter more than avoiding every mistake, and a team that never breaks anything is being too careful. Facebook paired the motto with releasing small changes often and testing many versions at once.
When it applies: A young product with few users, where mistakes are cheap to fix, is moving slowly out of caution.
Contested. Breaking anything that touches customers' money, data or safety causes real harm and legal exposure. Facebook itself replaced the motto in 2014 with 'Move fast with stable infra' as the cost of breakage grew. The phrase is also widely criticized as encouraging disregard for harm to users and society.
Not tried yet
None of our businesses has tried it yet.