The founding memo named a quality bar but did not define it. This memo is the definition.
We will be brief, because the bar is supposed to be operational, and a bar you cannot remember at 5pm on a Friday is not a bar.
The working definition.
The quality bar is the standard we would want a competitor to be held to.
That is the entire definition. The phrasing matters: not the standard we hope to reach, not the standard we would defend to a customer, not the standard we would accept from ourselves. The standard we would want a competitor to be held to.
The reason we use the competitor framing is that it removes the gravitational pull of self-protection. Most teams know, without being told, how to apply a high standard to other people’s work. The harder thing is applying that same standard to their own. The competitor framing is a deliberate trick — a way to import the rigor we already have for other people and aim it at ourselves.
If you find yourself defending a release against the standard you would have wanted a competitor held to, the release is not done.
Three places the bar shows up.
The definition is one sentence. The behavior that produces it is three things.
One. We write things down.
A decision made out loud and then forgotten is a decision that will be remade — usually badly, usually in a hurry, usually without the original constraint that made it good the first time.
We write decisions down. Not after the fact, in the form of a postmortem. Before the fact, in the form of a memo. The memo is the artifact that forces the team to be specific about what is being decided and why.
The memo culture is not for nostalgia. It is for accuracy. Writing forces a kind of thinking that talking does not. The people who write things down end up with sharper judgment over years, because they are the people who have been forced, on a hundred different decisions, to face the gap between what they thought and what they could actually justify on the page.
This is not a Stripe Press pose. It is a working tool. Memos at the Group are usually under 1,500 words. They are dated. They are not edited silently after the fact. If a decision proves wrong, we say so in a follow-up memo — we do not revise the original out of existence.
Two. We say so plainly when something is not good enough.
The hardest part of the bar is not setting it. It is enforcing it on a particular Tuesday in May when the team has worked hard, the release is mostly there, and shipping it would feel good.
We have all sat in reviews where everyone in the room agreed the work was below the bar and the work shipped anyway. The mechanism is the same every time: someone says the polite version of "I think this needs another pass," and the room hears the polite version and decides the work is fine because nobody said it more directly.
We try to be the place where the more direct version gets said. Not unkindly — but clearly. If a piece of work is not good enough, we want the person who thinks so to say so plainly, in writing if possible, before the release window closes.
This requires two things. It requires a team that can hear the direct version without taking it personally — which is mostly a hiring question. And it requires the senior people in the room to model it first, because if the most senior people in the room only say the polite version, the polite version is the version that gets normalized. Both of those are ongoing work. We expect to get them wrong sometimes.
Three. We treat criticism of our own work as part of the work.
If we ship something and someone outside the team points out a flaw we missed, that is information, not an attack. The right move is to update — internally, publicly, and in writing.
We would rather be told something is wrong now than discover it ourselves in twelve months. That is also true for things we have already written down, including this document. If a memo in this section ages badly, we will write the memo that says it aged badly. We will not quietly edit it.
This is the part of the bar that is hardest to internalize, because it requires accepting a permanent vulnerability: that the work is incomplete until people who do not work here have had a chance to push on it. It also produces by far the best feedback loops. The companies that compound are mostly the ones that figured out how to treat external criticism as fuel rather than friction.
What the bar is not.
The bar is not a launch readiness checklist. Checklists are useful, but checklists describe the floor, not the ceiling. Most things that ship below the bar pass every checklist that was written to catch them.
The bar is not a personal standard. Personal standards vary with mood, with fatigue, with how close the release window is. The bar is a team standard that gets enforced by the team, not by the individual most committed to it on a particular day.
The bar is not an aesthetic. We do not ship work because it looks like quality work. We ship work because it does, in fact, work — operationally, technically, and over time. Aesthetics are downstream of substance, and the bar is the bar for the substance.
The bar is not a guarantee. We will, occasionally, ship work that turns out to be below the bar. When we do, the right response is to say so, publicly if the work was public, and to write down what changed so the next decision is better. Failing the bar honestly is recoverable. Failing the bar and pretending we did not is not.
Why we are writing this down now.
Because the cost of a quality bar is highest when it is invisible. A bar nobody talks about is a bar that drifts, and bars that drift drift downward. The cheapest insurance against that is making the bar a thing the team can quote at each other in a review.
We are writing it down because the people we are hiring now will set the bar for everyone we hire later. If our first ten hires understand the bar the same way, the next ten will, and the next hundred, and the company will be a thing that compounds. If they do not, the bar will become a slogan, and the slogan will be repeated more loudly the further it drifts from the actual work.
We are also writing it down because we expect, at some point, to fail the bar in a noticeable way. When that happens, this memo will exist as the standard we knew we were failing. That is the right kind of pressure to put on ourselves now, before we have anything to defend.
The thing we are building only works if the bar holds. So the bar holds.