# Nobody commits to a direction they do not understand

A company can change its business model, its website and its team in a few months. What it cannot do is expect commitment from people who do not know where it is going. I think culture is what lets you turn without losing anyone, and it works both ways: someone has to explain it, and someone has to want it.

First the business model changes. Then the website. Then a paywall shows up that nobody had mentioned. Then, cuts. Each decision, on its own, can be defended. Together, in a few months, they leave the people inside unable to say where the company is going.

I have seen it, and it is not rare. What stuck with me was not the number of turns. It was people's faces at each announcement: not anger, confusion.

It was not a lack of talent or will. It was a lack of a direction anyone could explain.

## Turning fast is not the problem

With AI, testing a new direction costs a week instead of a quarter. Whoever is in charge [tests their own ideas in an afternoon](https://www.enderpuentes.com/en/notes/product-engineering/una-tarde-con-claude-no-es-una-estimacion), and if they do not work, throws them away. I think that is fine. A startup that cannot change its mind quickly has a worse problem than one that changes it too often.

The problem is something else. Every turn asks people to drop what they were doing and push in another direction. You can ask for that many times. You cannot ask for it without explaining why.

## Culture is what lets you turn without losing people

To me, a startup's culture is not the values document or the offsite. It is anyone, whether in engineering, design, marketing or sales, being able to answer three questions without thinking too hard:

- What are we building?
- For whom?
- What changes out there if it works?

If those answers are clear, a turn reads as a course correction. People adjust on their own, because they know what the new thing is measured against. They can support the founder instead of waiting for instructions. They can say “this does not fit what we want to be” before it becomes a lost quarter.

If they are not clear, each area pulls toward the version it understood. Engineering closes PRs for the old direction. Design ships screens for a product that already changed. Sales keeps promising what was sold before the turn. Everyone is working. Nobody is pushing the same way.

In [The team is the new context window](https://www.enderpuentes.com/en/notes/ai-engineering/el-equipo-es-la-nueva-ventana-de-contexto) I wrote about a shared base of rules so ten developers with AI do not work in ten different ways. This is the same thing, one level up: a shared base of what for.

## Understanding it is not enough

This is the part I care about most, and it is rarely said: culture is not only the responsibility of whoever is in charge.

As a team member, I can understand perfectly where the company is going and still not want to go. Understanding is the first step. The second is internalizing it: looking at what the company wants to be and deciding I want to be part of it. Without that second step, what you get is compliance, not commitment.

You can comply with a direction you do not share. You do the tasks, close the tickets, hit the date. What you will not do is speak up when something is drifting the wrong way, or spend an extra hour thinking about how to improve something nobody asked for. That only happens when the outcome matters to you.

That is why I find it unfair, both ways, for a founder to expect commitment without having explained anything, and for someone on the team to expect to be inspired without having asked themselves whether they want to be there.

## Commitment is not requested, it is built by explaining

From the side of whoever is in charge, what works is simple and tiring: explaining the why at every turn. Not the announcement. The why.

Before announcing it, run it past the people who will carry it. Whoever maintains billing. Whoever talks to customers. Whoever designs. Validating the idea with AI is fine; it will agree with almost anything if you frame it well. But AI does not know what was already tried last year, or how the team will feel when they hear it. People do.

And there is a side effect: whoever validated the turn is who later explains it to everyone else. It does not arrive as a new ticket, a memo or an out-of-context announcement in a team meeting. It arrives with the why, told by someone from the inside.

## Close

A startup can turn as many times as it needs. What it cannot do is expect commitment from people who do not know where it is going.

If you are in charge, explain the why of each turn before announcing it, and check that the three answers are still true. If you are on the team, ask until you understand them, and then decide whether you want to be there. If the answer is no, that is an honest answer too. The one that does not work is staying halfway.