Jan Koum

The pattern they actually represent

Three structures did the work that everyone remembers as character.

The first was a piece of paper. His co-founder Brian Acton wrote it, and Koum kept it taped to his desk: "No Ads! No Games! No Gimmicks!" Not a value in a deck. A note in somebody else's handwriting, in the room where the decisions got made, naming three specific things rather than an abstraction like focus. A refusal you can read at the moment of temptation is a different object from a refusal you agreed to in a planning session six months ago.

The second was the price list. From 2013, WhatsApp charged about a dollar a year after a free first year. Revenue therefore tracked how many people had the product, not how long they stared at it. Under that model a feature that lifts engagement adds engineering, support and reliability risk, and adds nothing to the top line. Games and feeds did not have to be resisted. They did not pay. The company had published the reasoning on its own blog in 2012, a year before the fee arrived.

The third was arithmetic. In February 2014 Sequoia reported 32 engineers, one for every 14 million active users, and not a penny spent on marketing. Facebook's announcement that month put the product at 450 million monthly users. At that ratio you cannot build a long feature list even if you decide to. The roadmap is bounded before anyone has to be principled about it. They got there by refusing to rebuild what already existed. Identity was the phone number already in your address book. No username, no password, no social graph to build, populate or moderate.

None of the three sits inside a person. All three can be copied by someone with ordinary willpower.

The blockage it speaks to

You are not failing to say no to bad ideas. You are failing to say no to good ones.

Look at the backlog. Every item has a person behind it — a customer who asked on a call, a competitor who shipped it last month, an investor who raised an eyebrow. Nothing on the list is stupid, which is why refusing feels less like discipline than negligence. The intention under the yes is responsiveness, a real virtue you have been rewarded for since your first customer.

The cost is quiet. Because the definition of done keeps moving, done stops being something anyone can check. A version you would call finished never arrives, so eighteen months of genuine work produces no event: nothing to launch, nothing to price.

Then the structural half, which is where the week fails you. Nothing on your calendar manufactures a no. Standups add. Customer calls add. Board meetings add. Between the backlog and the roadmap there is one mechanism: you, at the end of a long day, being the only person in the room who wants less.

That is why "be more focused" collapses inside a fortnight. It asks a character trait to do a job that belongs to a structure. Koum had three structures doing that job, and none required him to want less than anyone around him.

Three moves you can steal

Write the three things you will not build this quarter, and put the paper where the work happens. A refusal decided in advance, readable at the moment of temptation, turns a judgement call made with a customer on the phone into a check against a prior commitment. Three named things and a date can be broken. "Stay focused" cannot, because it cannot be tested.

Put a ninety-day revenue number beside your top five backlog items and delete the ones that earn nothing. Money under the pricing model you charge today, not engagement, not retention. Every honest "none, but people would use the product more" goes in a second list. If that list is longer than the first, you have a pricing problem rather than a scope problem, and that is the week's real work.

Name three systems you are building that already exist, and kill one by Friday. Auth, billing, search, notifications, a contact list. Rebuilt infrastructure never looks like scope creep from the inside; it looks like craft. But every system you own will ask for features later, and not building it is the only scope cut that costs you no relationship.

Where the pattern breaks

The note held as long as its authors owned the decision. It was never in the ownership. WhatsApp was sold in 2014. In January 2016 the subscription that made adding unprofitable was removed, in the same post that announced business tools: under the founders, two years before Koum left. The Business app shipped in 2018. Acton left in 2017 and Koum in April 2018, saying only that it was time to move on. In June 2025 Meta announced ads in Status and promoted channels. Nothing in that sequence required anyone to break a promise. It only required the constraint to live in two people's judgement rather than in the terms of the deal.

What to do this week

Take one sheet of paper, write today's date at the top, and under it the three things you will not build before the quarter ends. Name artefacts, not attitudes: not "no distractions" but "no built-in messaging", "no second onboarding flow", "no Android build". Sign it.

Then put it in the two places the decision actually gets made. Tape it where you write code or specs. Paste the same three lines at the top of the roadmap document, above the backlog, so that anyone who opens it reads them first.

You may rewrite the list when the quarter ends. You may not amend it in the middle of one. Three named things and a date make the refusal falsifiable, which is the one property good intentions have never had.