The date sales already promised

The email arrives on a Tuesday. A deal closed, congratulations all round, and in paragraph three: the customer has been told the integration ships by end of March. Engineering is learning about the commitment and its date in the same sentence. What happens in the next hour determines whether this becomes a delivery problem or stays what it currently is — a recording problem.

The commitment already exists

Start from the fact everyone in the room wants to argue with: the promise is made. It is in a contract, or in a customer's planning, or in the relationship — and engineering's acceptance was never a precondition for its existence. Refusing to "accept" it does not un-promise it; it just moves the promise off every internal record while keeping it fully alive in the external one. Now the organisation carries a real commitment that no plan accounts for, no capacity was reserved for, and no one is accountable for — the maximum-risk configuration, achieved in the name of process purity.

The angry meeting has the same defect. It renegotiates blame, not the date, and it teaches sales one lesson only: involve engineering even later next time.

Book it as what it is

A pre-promised date is an early commitment — a promise made before readiness — and the honest handling already exists: record the exception. Four fields, filled in the open:

• Why it was made early: the deal, named. This is a legitimate reason — real revenue on a real signature outranks process comfort, and pretending otherwise is its own theatre.
• What is still unknown: whatever shaping has not happened — scope, dependencies, the capacity figure the delivering team never gave.
• Who accepts the risk: sales leadership, by name. Not "the business", not the account executive's manager's manager — a person who can later be asked whether they would accept it again. This field does the most work and meets the most resistance, in that order.
• What would change the decision: the condition — "if discovery finds the vendor API cannot batch, the date moves and the customer hears it by the 15th."

Then run the demand through the same machinery as everything else — shaped, sized by the team, dependencies confirmed — knowing the date arrived before the evidence. The exception does not bless the gamble. It makes the gamble countable.

Displacement, out loud

Capacity did not grow when the deal closed. The pre-promised demand is an edit to the quarter, not an addition — something above the line moves down, visibly, and its stakeholder is told by name. Skip this step and the displacement still happens; it just happens silently, to a victim discovered at quarter end, which converts one honest hard conversation now into two dishonest ones later.

The only argument that changes a sales process

Here is what recording buys, three quarters in: a list. Every pre-promised date, its exception, its owner, and its close — kept, kept with differences, missed. When the exceptions are counted, the conversation with sales leadership stops being "please involve us earlier", which is a plea, and becomes "pre-promised dates missed at three times the rate and cost us these two renewals", which is a business case. Nobody has ever changed a sales process with a plea. The record is not revenge — it is the only shared set of facts from which the process can improve.

And notice the quiet symmetry: the same record protects sales. When engineering's own estimate slips on a demand that was ready, the ledger says so too. A commitment system nobody can game in either direction is the thing that finally makes the Tuesday email arrive as a question instead of a verdict.