Every serious sales organization runs on a discipline with a name, MEDDIC or a cousin, and the discipline is a demand: document the deal. Log the metrics, the economic buyer, the decision criteria, the pain, the champion, as you go, on every opportunity, every week. The premise is that the salesperson’s judgment is a black box to be pried open and made legible, so a manager can see inside and a company can own what it sees. It sounds reasonable. It is wrong, and the fastest way to see how wrong is to carry the same demand into rooms where the answer is already obvious.
The one question that sorts it
Ask one question and every profession sorts itself: is the document the deliverable, or a description of one? For some, the document is the whole product. A medical examiner’s output is the autopsy report, read and contested in court; a claims adjuster’s is a defensible determination someone pays out on; a scientist’s paper is the reproducible result. They run on deep tacit judgment too, but their job is to convert it into a record others can act on, so rigor is the work, not overhead. For the other kind, the document only describes the outcome. A surgeon’s deliverable is a healed eye, a firefighter’s a dead fire, a comic’s a room that laughed, and the judgment that produced each stays tacit and works directly. Nobody needs to read how the eye was fixed for the eye to be fixed. A salesperson’s deliverable is a closed deal, so the pipeline entry is not the product, only a thin description of one. The pipeline’s whole error is treating the second group as the first.
Narrate the incision
Make an expert in the second group document in real time and you meet the first failure mode. Tell a surgeon to log each of the roughly two hundred micro-adjustments of their hands as they make them and you have blinded the patient. Gary Klein spent years asking fireground commanders how they decide, and the answer came back as a version of I do not make decisions, I just know: the expert reads the smoke and acts on a recognized pattern, with no discrete steps to log. Hubert Dreyfus mapped it decades ago, the novice follows explicit rules, the expert has dissolved them into intuition, and dragging the expert back to articulate the rules regresses them toward the novice. You do not measure the performance by demanding the log. You degrade it.
The fiction written after
Then log it afterward, say the enforcers, once the work is done. Now you get the second failure mode, a fiction. A record produced after the fact is a reconstruction cleaned up for the reader, not a transcript of the tacit process that did the job. The most prestigious example on Earth is the judicial opinion. A judge issues a fully documented chain of reasoning, and a century ago the legal realists took it apart: Judge Joseph Hutcheson’s confession, The Judgment Intuitive, admits the decision arrives first as a hunch, and the opinion is written afterward to justify it for the file. The police report is the same, written at a desk from a stop that happened at the roadside; the pipeline entry filled in after the call is the same again, a tidy account of metrics and champions assembled to satisfy a manager, describing a deal that was felt out in real time across a table. The form is not lying. It is doing the only thing a post-hoc form can, invent a legible account of an illegible act.
It is not a document, it is surveillance
And the word documentation hides what a modern CRM actually is. It is not a document but a live database, read by the manager inside the system in real time, the rep’s logged activity accumulating long before anything is printed for a pipeline meeting. A document is something you author; a feed is something done to you. Which raises the one profession that exposes the arrangement: the criminal, whose whole craft is anti-documentation, to leave no legible trail, and the only party who ever forces a record out of them is the state, after capture, as a confession. That is what the pipeline is. Not a tool the salesperson wields, but a record extracted from the performer for an authority, a confession about work whose substance no confession can hold.
Why they want it anyway
So why keep demanding the impossible? Not stupidity, but a real hunger for legibility. Michael Polanyi named the fact in a line, we know more than we can tell, and James Scott showed what institutions do about it, force the illegible into legible form so it can be measured, audited, and owned. The rep’s judgment is the company’s most valuable asset and it walks out the door every night inside the rep’s head, so the company builds machinery to filet the expertise out of the expert and keep it on a server. But the filet kills the fish. Tacit judgment does not survive extraction: interrupt it and you get a worse performance, reconstruct it and you get a fiction, because the knowing was never separable from the knower.
What the form cannot hold
The deal, when it closes, is closed by the same faculty that removes the cataract and reads the smoke and finds the judge’s hunch, the reading of a human being across a table in real time. That is exactly what the form cannot hold. So the pipeline fills with metrics and stages and captures everything about the deal except the part that did the work, which is why it feels so busy and predicts so little, and why the best reps treat the CRM as a tax. Keep a record when a real downstream reader needs one, but do not mistake the record for the thing. The autopsy is the examiner’s deliverable. The healed eye is the surgeon’s. The closed deal is the salesperson’s, and the documentation of it, however rigorous, however surveilled, however dutifully confessed, is not the deal.
Touchstones: Michael Polanyi, The Tacit Dimension (1966), we know more than we can tell; Gary Klein, Sources of Power (1998), recognition-primed decision-making among firefighters; Hubert and Stuart Dreyfus on skill acquisition and the expert’s regress under forced rule-following; Joseph C. Hutcheson, The Judgment Intuitive (1929), and the legal realists on the hunch that precedes the opinion; James C. Scott, Seeing Like a State (1998), on legibility and metis. A companion to There Is No Such Thing as a Marketing Qualified Lead.