Customers
Deals and pipelines
Why the probability sits on the stage, why a lost deal needs a reason, and why winning creates a project rather than becoming one.
A deal is something you are trying to win. It belongs to a pipeline, sits in a stage, carries a value in integer minor units, and has an owner. That is nearly all of it — the interesting parts are the three things it deliberately does not do.
The probability is on the stage, not the deal
Every stage carries its own likelihood of closing, and the forecast weights each deal by the stage it is in. There is no per-deal percentage to set.
Why
A per-deal percentage is a number somebody guessed once and never revisited, and a forecast built from a thousand stale guesses is arithmetic dressed up as evidence. A stage probability is a claim about your process — *deals at proposal close forty per cent of the time* — which can be checked against your own history and argued about in a room. When it is wrong, one number is wrong rather than four hundred.
The win rate counts closed deals only
Wins divided by wins plus losses. Deals still open are not in the denominator.
Dividing by every deal ever opened makes the rate fall every time somebody adds a prospect — which teaches people not to add prospects, and a pipeline nobody puts anything into is worse than no pipeline.
A loss needs a reason. A win does not
Moving a deal into a lost stage asks why, and refuses without an answer. Moving one into a won stage asks nothing.
The reason is the only part of a lost deal worth keeping. *We won* explains itself, and demanding an explanation is friction at the moment somebody is least inclined to tolerate it. The field is free text on purpose — every drop-down of loss reasons is somebody's taxonomy, and the useful part is always the sentence.
Winning creates a project. It does not become one
A won deal offers to create the delivery project, carrying the company across. The two records stay separate and are linked.
| The deal answers | The project answers |
|---|---|
| What was sold, for how much, by whom | What it costs to deliver |
| When it closed and why | Who is working on it and how far along |
| Which pipeline and stage it moved through | Which cadence its work follows |
Why they are not one record
The margin question needs both halves, and a single record answers it with whichever was edited last. A lost deal must also not leave a project behind, and a project routinely outlives the deal that started it — a client on their fourth phase has one project and four deals.
What a pipeline must have
- At least one won stage and one lost stage. A deal that can never leave is a deal that sits in the forecast forever.
- Stages in a stated order. Position is what makes *further along* mean anything.
- A currency on any deal that has a value. A bare number is not a price. Values are never converted between currencies — the forecast reports per currency, the same rule the workforce cost report follows.