Most deals worth doing leave both sides ahead
Good products create value that did not exist before. Some contests still have exactly one winner.
Do things that allow you to win and that allow someone else to win. It is not a zero-sum game.
GitHub hosts open source projects for free and charges for private repositories. Every popular library that lives there brings developers to the site, and some of those developers go on to pay for private code at work. The projects get free hosting and a wider audience; GitHub gets its customers. That arrangement is the plainest reading of this line from its co-founder. The talk or post where he first said it is hard to trace, so it is best taken on its own terms.
A zero-sum game is one where every gain for one side is a loss for the other, like poker or a fixed pie cut at a table. Much of business feels like that from the inside: a competitor takes a contract, a customer negotiates a discount. Preston-Werner’s claim is that most of the useful work in building a company is different. A good product creates value that did not exist before, so the customer gains more than the price and the company earns more than its cost.
In October 2010 he also published an essay on optimizing for happiness, arguing for running GitHub on its own revenue instead of venture money, partly so that decisions could favor staff and users over a quick exit. The two ideas fit together: a company that does not need a fast sale can afford arrangements where others do well too.
The line fails where the game really is zero-sum. A single contract has one winner. A candidate accepts one offer. In a price war between two near identical products, the customer gains and both sellers lose. Pretending otherwise in those cases is naive.
It can also be misused as cover. “Everyone wins” is a phrase some companies repeat while taking more than their share. The test is concrete: after the deal, can the other side point to what it gained, and would it sign again?