Starting and building
Zhang Yiming
Delayed gratification
Zhang Yiming (Contemporary · 1983–) — Delayed gratification.
What actually happened?
One passage of his gets quoted constantly: I have met a great many clever people, but very few willing to delay gratification, and people at different orders of that willingness cannot usefully discuss a problem at all. He spent five years grinding on product with almost nothing said in public — no speaking slots, no profiles, just the work. The sentence is not advice about patience. It is a claim that this one trait sorts people into groups that cannot hold a conversation with each other.
Delayed gratification
Giving up a present pleasure for a longer-run value is a scarce capability and, in his account, the sharpest way to tell people apart. When he picks a partner or an employee the first measure is not raw ability but how far out that person is willing to place the payout.
Look at a record of past moves: always the highest immediate pay, or sometimes the steepest slope. The answer is already written down.
Against vagueness
He asks people to speak without maybe, roughly, I feel and it seems, forcing precision. If you cannot say it clearly you have not thought it through, and something not thought through is most likely wrong before anyone starts building it.
Rewrite a hedged claim as: I think X, because Y, confidence Z per cent. Half the sentences do not survive the rewrite.
The company is the product
Develop the company as though it were a product: you are not only building the thing users touch but also the organisation, the incentives and the information flow behind it. Optimising one part in isolation is the usual way of manufacturing a problem somewhere else.
A team rewarded on its own metric will improve that metric and quietly damage two others nobody is watching.
User value has a time dimension
Will this product still be worth something in three years is the question he puts against short-term numbers, and he refuses trades that lift this quarter by taking from the experience that produces every later quarter.
An extra advertising slot lifts revenue this month and is deducted from retention later. Put both curves on one chart before deciding.
How do I use it today?
At the next team meeting, ban maybe, roughly and I feel from everyone's speech, and require every judgement in the form: I think X, because Y, confidence Z per cent. The point is not precision for its own sake. It is that a claim in that shape can be proved wrong, and the ones that cannot be stated in it were never judgements at all.
Deep read
Read alongside
Further
Lines to keep
People at different orders of delayed gratification cannot discuss problems usefully.
A startup's biggest advantage is execution, not the idea.