The First Number You Say Out Loud Becomes the Schedule

A founder on a discovery call asks how long it will take. The answer arrives in about a second, because the person answering has priced a lot of projects and has a feel for it. Eight to ten weeks, something in that range.

Nothing has been scoped at that point. No feature list on paper, no data model, no line drawn between what is in the first version and what waits for the second. What got said was a reflex, and a reflex is not an estimate. It is now the only schedule in the conversation, and it belongs to the founder rather than to the person who said it.

We did this earlier this year. The number that came out on a call was less than half of what the scoped plan came back as, once everything discussed on that same call was written down and priced properly. Both numbers were given in good faith. Only one of them had been checked.

The first number is the one that gets planned around

A founder does not file a discovery-call number under "preliminary." They plan against it, because planning is their job and that number is the only input they have.

Eight to ten weeks becomes a pilot they mention to three people. It becomes a hiring conversation, an investor update with a quarter attached, a launch aimed at a season that actually matters to the business. None of that is naive. It is what you do with information from the people who are going to do the work.

So when the scoped plan arrives, it is not landing on a blank page. It is arguing with commitments that already exist. And it usually loses that argument, because the first number came from the same shop. It carries the same authority and it got there first. The founder's most reasonable reading is not "they looked closer and learned something." It is "they are padding now that I am interested."

That is the real damage. Not the schedule. The suspicion.

Why the two numbers are so far apart

The gut number prices the thing the estimator pictured while the founder was talking. The scoped number prices what was actually said.

Sitting between those two are all the parts of a conversation that register as detail and behave like scope. An admin area, which is a second interface with its own permission rules and its own screens. A content side, because someone on the client's team needs to change words and images without calling us. A structure of organizations sitting above individual users, each able to invite the level below it, which is schema work rather than a screen. Notifications, in a case where the platform choice ruled out the obvious mechanism, so now two outside services are in the build. The question of which entity signs, who owns the accounts, and whose card is on file for the services the app depends on.

None of those was a change of mind. Every one of them was on the call. They did not make it into the picture in the estimator's head, because a picture is a screen and scope is a system. You can hold five screens in your head during a conversation. You cannot hold a permission model.

There is a second reason, and it is less flattering. A fast number is a friendly thing to say. It keeps the call warm. Nobody wants to answer an excited founder with "I would need a week to tell you that," even when a week is the honest answer.

The gap is the vendor's problem, and vendors sit on it

The uncomfortable part is what happens next.

The gap gets discovered early and disclosed late. It gets discovered the first time someone actually writes the scope document, because that is the moment the admin area stops being a word and becomes a list of screens. It gets disclosed when the shop can no longer avoid it, which tends to be the week a contract is supposed to get signed. That is the worst available moment for everyone in the room.

Our own version of this sat open on an internal board for five weeks. Nobody was hiding it. It stayed open because the conversation is unpleasant and there is always something with a nearer deadline, and because a number that has never been retracted keeps quietly working as a promise. Silence is cheap for the vendor and expensive for the founder, which is exactly the shape of a thing that does not get fixed on its own.

What the honest version looks like

It is not better guessing. Three moves cover most of it.

Say the number with its contents attached. "Eight to ten weeks for the five screens we just walked through. I have not scoped the admin area or the content tools, and in my experience those are not small." Same number, now it has edges. A founder can plan against that without being misled, because the parts that are missing are named.

Write the corrected number the day it exists. Not the week the contract needs signing. The distance between those two dates is the entire problem.

Bring it before the contract, as our mistake, out loud. A gap surfaced before signature is a conversation about scope and phases, and it usually ends with a smaller first phase and a better project. The same gap surfaced at week twenty is a dispute. By then the founder has made real decisions on a schedule we handed them and never took back.

What founders should do with this

When a shop gives you a number on the first call, do not read the speed as confidence. Read it as a picture in someone's head. Then ask two things.

First: what is inside that number, and what is not. Use the words. Admin. Permissions. Content editing. Integrations. Migrating the data you already have. Watch whether the number moves. If it does not move at all, the estimate is not tracking scope, and you have learned something more useful than a timeline.

Second: when do I see a scoped version, and what would make it different from this one. Then hold them to the date. If the scoped plan shows up the same day the contract does, the process ran backwards and you are being asked to sign something nobody had time to argue with.

There is a third thing, and it is about character rather than process. If the number changes, the shop should be the one telling you, unprompted, before you think to ask. A partner who makes you discover this gap on your own will make you discover the next one the same way.

The number is still worth asking for

None of this is an argument for refusing to answer.

A founder who asks "roughly how long" is really asking whether to keep talking. That is a fair question and it deserves a real answer at the moment it is asked. A shop that will not give you an order of magnitude until you pay for a discovery phase is running its own version of the same dodge, just with better manners.

The answer is not silence. It is a number with edges on it, and a date when a better one arrives.

Every project has a first number. The only thing worth deciding is whether it is also the last one anybody says out loud.