Methodology

Most commercial diligence is a reading exercise. Someone reads the management presentation more carefully than the last person did, calls the references the seller chose, and writes it up.

We work in 4 layers, and each exists because the layer below it can be wrong. Reconstructed data can be technically accurate and tell you nothing. Primary evidence can be honest and unrepresentative. Observation catches what neither can, and none of the 3 decides anything alone.

The layers run on frameworks built from over a decade of the founder's enterprise SaaS commercial experience and 100+ commercial engagements delivered inside software companies: a standard way of rebuilding revenue from source, interviews designed to surface disagreement rather than agreement, a grading of every finding by the source behind it, and a standing model of commercial red flags that makes one company comparable with the next. What follows is the architecture. The implementation is the practice, so parts of it stay unwritten.

The frameworks stay constant. How they are applied does not. Which layers run, how deep they go and what you receive at the end are set by the case: the question you need answered, the access available, and the scope of the work we agree with you. Before you have a seat at the table, we work from what a company reveals without meaning to. Under exclusivity, from what a seller agrees to open. After you own it, from everything.

Layer one. Reconstruction

What goes in: Where access allows, billing and subscription exports, CRM exports and contract terms. In every case, the entire observable surface of a software company: hiring patterns, release cadence, documentation depth, pricing page history, review site movement, the specifications in their own job adverts.
What comes out: Cohort retention rebuilt from source rather than accepted from a slide. Net revenue retention separated from logo retention. Concentration, and how it has moved. Discount discipline at renewal. Win rate and sales cycle by segment. The share of pipeline carrying a named next step and a date.
What it changes for you: You stop negotiating against numbers you cannot check. Figures trace back to source, so when you move on price you are moving on evidence rather than on a doubt a seller can talk you out of.

Layer two. Primary evidence

What goes in: Conversations, deliberately weighted away from the reference list. Churned customers, because they say what current customers are too polite to. Lost deals, because losses are unrehearsed. Former sales staff, because they know which deals closed for reasons the CRM does not record. Channel partners, because they see the competitive set from outside. Current customers last, and never only.
What comes out: Why customers actually bought, in their words rather than the company's. What renewal costs to earn. Which competitor was genuinely evaluated. What would have to break for them to leave. Every finding carries a confidence level and the class of source behind it, so a reader can see which conclusions rest on one voice and which rest on many.
What it changes for you: You hear the version of the company nobody prepared for you. This is often where the condition on the offer comes from, and it is rarely the risk the seller flagged in the room.

Layer three. Observation

What goes in: Live sales calls, where the seller grants access, across different stages of the funnel: discovery, demo, negotiation.
What comes out: Whether the representative qualifies or accepts any meeting offered. Whether value is articulated or features are demonstrated. Whether the deal is won on value or on discount. What happens when a prospect raises price. Who has to be on the call for it to close, and how the conversation changes when they are not.
What it changes for you: You find out whether you are buying a company or a person, before you own the answer. No financial, legal or technical workstream reaches this question, and in a software acquisition it can matter more than any of them.

Layer four. Judgment

This is where the method stops and the reading begins.

Findings are scored against a standing model of commercial red flags, built from a decade of enterprise software sales and refined across every engagement. The model makes assessments comparable. It does not make the decision.

The decision is what the layers mean for this buyer, at this price, with this plan for the company afterwards. The same set of findings supports a walk for one buyer and a proceed at a lower price for another, and no framework resolves that. It is the part of the work that has to be earned rather than run.

What it changes for you: You get something you can take to an investment committee: a position, the conditions that would make the price work, and the reasoning underneath. Not a list of risks with the call handed back to you.

Where the method ends

Three things this method does not do, stated plainly because knowing the limits is part of trusting the output.

It does not predict.: It establishes what is true now and what would have to change. Everything past that is a forecast like any other.
It does not replace financial, legal or technical diligence.: It answers the commercial question and hands the rest to people who do those properly.
It does not work without some seller cooperation at the full assessment level.: The Commercial Thesis Review exists for the stage where you have none, and is honest about being narrower as a result.

How we use AI, and why it is not enough on its own

We use AI throughout the work, and we are trained to use it well across the leading models, from OpenAI to Anthropic. The models are available to anyone. What we run on them is not: our own proprietary skills and plugins, built from real engagements and the commercial best practice that came out of them. They exist nowhere else, because the work they were built from is ours.

That is pre-work you benefit from before your engagement begins. The frameworks, the questions and the checks are already built, so the time goes into your company rather than into working out how to look at it.

It is a combination, not a substitute in either direction. Whichever model you choose, however capable, 3 things decide the outcome of commercial diligence, and none of them comes from the model alone.

Context.: A model knows what it is given. Ours is given what our engagements taught us to ask for: the questions that separate a working commercial engine from a well-told one, and the patterns that repeat from one company to the next. The facts that decide a deal still sit where no model can reach them: why a deal closed when the CRM does not say, what a churned customer tells you when nobody prepared the conversation, who has to be on a call for it to close. Knowing what to go and get, and getting access to it, is much of the work.
Experience.: Whether a retention figure is strong or a warning in this segment, and whether a pipeline problem is fixable in a quarter or is the business, is calibration. It comes from over a decade carrying a number in enterprise SaaS and from running the commercial motion inside software companies, not from reading about them. A model can tell you what the number is. Experience tells you what it means in this company.
Judgment.: Layer four is the part no model does for you: deciding what the findings mean for this buyer, at this price, with this plan for the company afterwards. A model will give you an answer. It will not stand behind it in front of your investment committee.
The founder in every case.: The skills take the gathering off the founder's desk. What is left is the part that decides the outcome: hours of focused, senior thinking on your case from a founder who knows the commercial engine end to end, from the first sales call to the board pack, and who works today as a C-level executive in commercial operations across several SaaS clients. That is what you are paying for, and no tool supplies it.

What we agree with you before the work starts

Every engagement starts with a way of working agreed with you in advance, so scope, data and delivery are settled before anything is opened.

Scope.: Which questions the engagement answers, which layers it runs and how deep, set by your case and the access available rather than by a fixed package.
Data.: How your information and the target's is handled: who sees it, where it is kept, which tools may process it, and what happens to it when the work ends. Data privacy and confidentiality are respected in every engagement, on the terms agreed here.
Progress reviews.: The key meetings where findings are reviewed with you as they form, so the verdict builds with you rather than arriving at the end.
Deliverables, formats and dates.: What you receive, in which format, and by when. Formats follow how your team works: HTML report pages organized by section, a presentation for your investment committee or board, a working model in Google Sheets or Excel, or a combination.

Last updated September 11, 2026