What Your Win-Loss Notes Are Telling You

A win-loss record tells you which option a rep picked from a dropdown at the moment they closed the deal. A dropdown returns only the choices somebody wrote before those deals existed, so the report built out of it counts picks rather than reasons. What moves positioning is the buyer’s own wording, and almost nobody keeps it.

What should a win-loss record contain?

Two things the form has no field for: the buyer’s own wording, and the alternative they named when they said no. A third is where the deal changed direction, which a closing note misses because it is written from the end.

A reason code carries none of that. Price, timing, competitor, no decision. Four boxes configured by somebody forecasting which losses were coming, and a rep clearing records at quarter end picks the nearest. Count those picks a year later and the count gets presented as evidence.

What each kind of record can carry:

Record	Who produces it	What it supports	Where it thins out
Dropdown reason code	The rep, clearing the record	A count by category	Categories fixed before these deals existed
Free-text note at close	The rep, from memory	Recall of one deal	Written from the end, to close a record
Written questionnaire	The buyer, unpaid	Whatever they will draft	A drafting task for somebody with no reason to do it
Debrief interview	Someone off the deal	The comparison set, and where it turned	Needs a person and a calendar slot
The call recording	Nobody. It exists already	Objections in order, in the buyer’s phrasing	Consent, plus somebody has to listen

Why do win-loss surveys come back empty?

Because a questionnaire moves the writing onto the person with the least reason to do it.

We hit that on a ghostwriting programme rather than a win-loss one, and the mechanism carries over. A written survey went to a leadership team at a global consulting firm. Six responses came back, none specific enough to build a paragraph on. The thinking existed, in recordings of conversations those people were already having.

A buyer who has just chosen somebody else has less reason still. More in why executives will not fill in your form.

Who reads the records, and when?

Somebody who was not on the deal, reading the whole set in one pass rather than a record at a time.

We learned that ordering on a 255-page content build, where generation ran as one pass and verification as another. Generation is cheap and arrives sure of itself. The whole job of verification is to disagree with it, which takes longer. Run them together and the sure half wins, and a debrief typed an hour after the call is the sure half.

On a batch of our own posts, one-line source entries produced invented supporting detail that survived a repair pass, because the writers closed the gap themselves. A short CRM note gets completed the same way by whoever reads it next.

What do you do when the same reason appears twice?

Change something, and put a name against the change.

Our editorial rules on one client newsletter were written at issue 20 rather than issue 1, because that is when it became clear the same notes were being given twice. Eleven issues have since gone out under nine standing rules and a QA script.

The second time an objection arrives in the same words, it becomes an edit to a page or a new qualification question.

The categories in your loss-reason dropdown were written before any of the losses happened.

What has to be true of the CRM first?

It has to hold a history rather than a snapshot.

One audit we ran found roughly 98% of a client database sitting in the stage those contacts were created in, with no automation to move anybody out. It proved the contacts existed and said nothing about what happened to them.

Notes die the way stages die, because both wait on a person remembering rather than a trigger firing. Attach the debrief to a stage change that already gets recorded, or you keep only the deals somebody felt strongly about.

The longer versions of both are in the recordings you already have and auditing the CRM first.