The Definition Nobody Writes Down Is the One That Breaks Your Forecast

Emily Goh · 2026-10-29

Ask your Sales leader and your CFO what "closed won" means. You will probably get the same answer: "the deal is signed."

Then ask which quarter a deal belongs in when the customer signed on 30 September, your side countersigned on 2 October, and the rep moved it to Closed Won on 5 October. Now you may get two answers.

That gap is why closed won breaks forecasts. It is the one definition everyone assumes is already settled, so nobody writes it down.

Here is the short answer to how to define closed won: write one sentence for each of five questions. What event makes a deal won? Which date counts? Which amount counts? What has to be true in the CRM before the stage can change? And what reverses it? If any of the five has no written answer, your teams are each answering it their own way.

Why "signed" isn't a definition

"Signed" sounds precise. It hides at least three choices.

Signed by whom? The customer, or both sides? Signed what? An order form, a master agreement, a purchase order? And signed when? The date on the signature, the date the rep found out, or the date someone updated the CRM?

Each team answers those choices to suit its own job. Sales wants the deal counted the moment the customer commits. Finance wants it counted when it can be booked. Both are reasonable. The trouble starts when both teams report "closed won" and nobody says which version they mean.

That is allowed, by the way. Sales and Finance can keep different definitions. What isn't workable is both of them using the same two words for different things and finding out in the board meeting.

The five questions every closed won definition answers

1. What event makes a deal won? Pick one, and name the document. "The customer has signed the order form" is a definition. "The deal is done" is not.

2. Which date counts? The signature date, the countersignature date, or the date the stage changed in the CRM. This one decides which quarter a deal lands in, so it is often the choice that moves a forecast most.

3. Which amount counts? First-year value, total contract value, or annual recurring revenue. Say whether one-time fees and services are in or out. Two teams can agree a deal is won and still report numbers that are far apart.

4. What has to be true in the CRM first? The rules a record must pass before it can say Closed Won. For example: amount above zero, close date not in the future, a signed document attached, a primary contact on the deal. If a closed won deal can have a close date next year, your definition has a hole in it.

5. What reverses it? Customers cancel, payments fail, contracts get rewritten. Decide what happens then. Does the deal move to Closed Lost, stay won with a flag, or move back to an open stage? Without an answer, reps make one up each time, and last quarter's number quietly changes.

Then add two lines under all five: who owns this definition, and when it last changed. The owner is a person, not a team. The change record is what lets you answer "did the business change, or did the definition?" when the number jumps. (The Definition Drift Tax shows what that missing change record costs.)

Where the two versions live side by side

If Sales and Finance genuinely need different answers, write both. Label them.

Sales: Closed Won (Sales) = customer has signed the order form; dated by customer signature; amount is first-year value.

Finance: Closed Won (Finance) = both parties have signed; dated by countersignature; amount is annual recurring revenue, services excluded.

Now every report can say which one it uses. With calendar quarters, the deal signed on 30 September lands in Q3 for Sales and Q4 for Finance, and nobody has to argue about it, because nobody is surprised.

Check last quarter's deals this week

You can test how much the missing definition is costing you with one export. Steps 1 to 3 take about an hour; step 4 needs Sales and Finance in the room.

1. Export every deal that entered Closed Won last quarter. Filter on the date the deal entered the stage, not the close date: a close-date filter hides exactly the deals with wrong close dates. In Salesforce, use the Opportunity History report; in HubSpot, the "Date entered Closed Won" property. Include amount, close date, created date and owner.
2. Count the ones that are almost certainly wrong. A close date in the future. An amount of zero. A close date before the deal was created. No signed date or e-signature status, if your CRM records one. Some will have a good reason, like a free pilot or a migrated record. Most won't.
3. Count the ones that moved. Deals that have left Closed Won since (in HubSpot, "Date exited Closed Won"; in Salesforce, Opportunity History), and, if field history is on for Amount, deals whose amount changed after closing.
4. Take five deals signed in the last week of the quarter and ask Sales and Finance which quarter each belongs in. Write down every disagreement.
5. Write the five sentences. Start with the questions where the disagreements came from. Put a name next to the definition and a date on it.

The counts from steps 2 and 3 are your evidence. They turn "I think our closed won number is soft" into "these deals in last quarter's number need checking, and here they are."

If your pipeline review keeps turning into a debate about whose number is right, Your Pipeline Meeting Is a Blame Ritual. It Doesn't Have to Be. is the next read.

Where Gohvio fits

Gohvio turns the parts of those sentences that can be checked, like the CRM rules in question 4, into checks. You upload your own guardrails, the definitions your teams actually use, as a document. A person approves them, and every change is recorded with who changed it and when. Gohvio then checks your CRM data against those guardrails and shows what is broken, what it does to pipeline and forecast, and the rows behind every finding. It is read-only today: it reads your CRM and does not write to it.

Stop debating your data. Start making evidence-based decisions.