A Thought Experiment About Trusting a Number
Picture a deal team putting a number in front of an IC. Call it an operational health score of 62. Someone asks the natural question.
Why 62 and not 70?
What happens next is the whole game.
Two versions of the answer
In the first version, the answer lives in the analyst’s head. They remember the inputs, the weighting, why the cultural compatibility index pulled the score down. The explanation is a reconstruction from memory. The IC is trusting the analyst.
In the second version, the answer lives in the work. Every input is traceable. Every weight is documented. The IC member who asked the question can walk the chain from score to evidence without the analyst in the room. The IC is trusting the method.
Both produce a score of 62. Only one survives being interrogated three years later by someone who was not in the room.
Numbers and money follow. They do not lead.
A diligence score is a compressed representation of a judgment. That is its value. That is also its risk. The compression strips the reasoning. When the reasoning is stripped, the number becomes a trust proxy for the person who produced it. If you trust the analyst, you trust the 62. If you do not, you have no way to verify it.
Scores have worked this way for a long time because there was no practical way to keep the reasoning at scale. There is now.
What changes when the reasoning is kept
When the reasoning is kept, the score stops being a verdict and starts being a door.
A credit score works this way. The number tells you where someone stands. The report behind it tells you why. You can pull the handle and walk through to the inputs.
A diligence score can work the same way when three things are true.
The method is reproducible. A second team, following the same methodology, would produce the same score from the same inputs. The method has to be explicit, recorded at the point of application, not reconstructed from memory afterward.
The evidence is chained. Every input to the score traces back to a source. Every source is stored once, in one place, with a timestamp. The chain from raw evidence to final score is preserved and walkable.
The method is versioned. The methodology used to produce the score today is documented well enough that someone can reproduce it three years from now. That matters because diligence scores get re-read. When a deal goes sideways in year two, the first thing the IC reaches for is the original diligence.
What this looks like in practice
We built a scoring system and ran it against a data room we had just ingested. The confidence levels came back lower than expected. When we dug in, the issue was that some source documents had not imported cleanly. Old spreadsheet formats that did not convert. PDFs where the text extraction missed sections.
The system did not hide the problem. It flagged it through lower confidence, which forced us to investigate the source. A score you can trust is one whose trust is built into how it was made. When the chain is visible, problems surface early. When the chain is hidden, problems surface at the IC meeting, or worse, in year two.
The bar the craft has not cleared yet
Most scoring systems in commercial diligence are closer to the first version than the second. The score lives in a model. The reasoning lives in the analyst’s head. The evidence is scattered. The methodology is implicit.
That cost has dropped. The tools to keep the reasoning, chain the evidence, and version the methodology exist now. The bar the craft has not cleared is an infrastructure bar, not an effort bar. Clarity is the moat. The practices that build clarity into their scores first will be the ones everyone else is comparing themselves to.
-Regis
Regis Hadiaris is Managing Partner, AI and Product Innovation for The Wisory. He is responsible for IntelliQ, the company’s proprietary platform designed to enhance the quality, speed, and precision of strategic and investment decisions.
