Skip to content

08Service

Operations Due Diligence

What the data room asserts, tested against how the operation actually runs.

Operations due diligence is an operator’s independent read on a target’s operating model, delivered before the transaction commits you to it. It answers the questions the data room is designed to keep abstract: can the delivery engine scale at the rate the thesis assumes, or does it already run hot at current volume? Is quality a measured system or an asserted virtue? How concentrated are decisions, relationships and knowledge in the founder? Where does margin quietly leak — rework, credits, write-offs, unbilled hours? And does cash behave the way the model says, or the way the billing process actually works? The output is a board-grade written assessment, ranked by financial exposure.

The craft here is not deal experience — it is operating experience, pointed at someone else’s operation. For nineteen years my job has been reading operating models from inside and finding where they leak: auditing quality across 2,000+ campaigns and 450 clients and moving the score from 95% to 99%; building the measurement a global network relied on to protect more than USD 20 million in billings; compressing a billing cycle from roughly two months to fifteen days across 75 entities, which required understanding exactly where cash gets stuck and why. Due diligence applies that same discipline to a target: the same questions, the same evidence standards, compressed into a deal timetable.

The assessment is built for the people who must act on it. For an investment committee: the operational risks that should price into the deal, ranked by exposure, with the evidence trail behind each. For the deal team: the specific claims in the CIM that did and did not survive contact with the operation. For whoever owns the asset after close: what the first hundred days should actually contain — not the template plan, but the sequenced repairs this specific operation needs, in the order the risk demands. Where a full review will not fit the timetable, a narrower pre-LOI red-flag read exists for exactly that purpose.

01The problem

Deal teams diligence what they can see. Financial DD prices the past; commercial DD sizes the market; legal DD checks the paper. The operating model — whether the delivery engine can actually carry the growth case, whether the quality story is a system or a performance, how much of the company is really one founder — usually gets a management presentation and a site visit. Then the deal closes, and the surprises that surface in the first two quarters are almost always operational. By then they are not findings. They are your problems.

02Signs you need this

When this is the right call.

  • 01

    The thesis assumes the operation can double without breaking

  • 02

    Quality and delivery claims in the CIM have never been independently tested

  • 03

    The founder is the operating system, and the model prices that at zero

  • 04

    Margin depends on rework and utilisation assumptions nobody has verified

  • 05

    Working capital behaviour is a model input, not an observed fact

  • 06

    The hundred-day plan is a template waiting for a signature

03The method

How the work goes.

  1. 01

    Read the operation as it runs

    Beyond the data room: management working sessions, process walk-throughs, and — where access allows — live work and system evidence. I trace how work actually moves from sale to delivery to cash, because the distance between the described process and the real one is itself a finding, and often the largest.

  2. 02

    Test the claims

    Every material operating claim gets pressure-tested against evidence: quality scores against how they are actually measured, delivery metrics against their definitions, utilisation against the roster, capacity against the growth case. Where the CIM says one thing and the workflow says another, the workflow is the truth.

  3. 03

    Price the operational risk

    Findings are ranked the way an investor needs them: by financial exposure and by how hard each is to fix. Founder-dependence, key-person concentration, hidden rework, margin leak, cash-cycle drag — each one sized, evidenced and tied to what it means for the thesis, not listed as observations.

  4. 04

    Write the assessment

    A board-grade written report: what holds, what does not, what it costs, and what the first hundred days should contain. Written to be defended in an investment committee — every material claim traceable to evidence — and short enough that it will actually be read by the people who decide.

04In depth

What this work really involves.

What financial diligence cannot see

EBITDA is an output. Financial diligence can tell you what the machine produced; it cannot tell you whether the machine will survive the plan you are buying it for. The gap hides in operational mechanics that never reach a data room: the four people who actually hold the client relationships, the quality process that exists mainly on the slide describing it, the delivery team already at capacity while the model assumes forty percent growth, the rework running through the cost base with no line of its own. None of this is fraud — management often cannot see it themselves, because nobody with operating scars has looked. That is the specific job here: an operator reads the operation, so the price reflects the machine as it is rather than as it presents.

Will the delivery engine carry the thesis?

Almost every growth thesis is an operating assumption wearing a financial costume. Double revenue in three years means: double throughput without quality collapse, hire and ramp at a rate the onboarding process has never sustained, and keep the cost of delivery flat while both happen. Whether that is achievable is legible in the current operation, if you know where to look: how close to capacity the engine already runs, whether throughput scales with headcount or with heroics, how long a new hire genuinely takes to reach standard, which processes break at the next volume step because they were patched at the last one. I test the thesis against the engine and give the answer a shape: what the plan truly requires, what the operation can hold today, and the cost of closing the distance.

Quality claims versus quality systems

Every CIM asserts quality; almost none can evidence it. The difference is testable in a day if you know the questions, because I have spent years building the real thing: quality scored against a defined standard across 2,000+ campaigns, defect rates traced to root cause, measurement that moved a network from 95% to 99% and could be defended line by line. A real quality system shows you its standard, its sampling method, its scores over time and its worst quarter. A performance shows you a plaque, an anecdote and a suspiciously smooth trend. The distinction prices directly: a business with systemic quality can absorb the growth case; one running on inspection heroics will convert your volume plan into client-facing defects within two quarters — a risk that belongs in the model, not in the surprise column.

The founder-dependence map

In founder-led targets, the most important diligence question is rarely asked precisely: if this person stepped back tomorrow, what exactly stops? I map it concretely. Which client relationships are genuinely institutional and which are personal. Which decisions the organisation makes without the founder, and which queue behind them — the approval flows tell the truth even when the org chart does not. What lives only in their head: pricing logic, supplier history, the reasons processes bend for certain accounts. The output is a dependence map with a transition cost attached: what must be institutionalised, in what order, and how long it realistically takes. Deals rarely die of founder-dependence; they die of pricing it at zero and discovering it at month four, with the earn-out already contested.

Where margin quietly leaks

Operating margin has a stated version and a real version, and the distance between them is usually rework and its relatives: work done twice because the handoff failed, credits and makegoods issued to keep clients calm, delivery hours never billed because scoping was optimistic, write-offs approved quietly at quarter end. Most targets do not measure any of this as a category — which is convenient, because unmeasured leak is invisible in a data room. I know where it pools because finding it was my job: the makegoods QA framework I built at a global advertising network, protecting more than USD 20 million in client billings, existed precisely because this class of leak compounds when nobody measures it. In diligence I reconstruct the leak from the artefacts — credit notes, revision counts, utilisation gaps — and size what the thesis is actually inheriting.

Cash-cycle mechanics, observed rather than modelled

Working capital models assume cash behaves the way policy says. Cash behaves the way the billing process actually works — and the two diverge in ways an operator can read directly. Where invoices genuinely originate and how many touches they take. Where approvals stall and disputes incubate. Which clients have quietly negotiated their own reality regardless of stated terms. How much work-in-progress sits unbilled because delivery and finance reconcile monthly at best. I have rebuilt this machinery at scale — a billing cycle taken from roughly two months to fifteen days across 75 entities and 2,000+ employees — so I know both what broken looks like and what fixable is worth. Sometimes the finding is risk; sometimes it is upside the model never priced. Both belong in the assessment.

From findings to the first hundred days

A diligence report that ends at the findings does half the job. The last section of my assessment is the operating agenda for ownership: the repairs sequenced by risk rather than by visibility, the two or three metrics that must exist by day thirty because they are how you will know whether the thesis is on track, the founder-transition steps with owners and dates, and the quality or cash-cycle fixes that fund themselves fastest. This is deliberately not the template hundred-day plan — those exist to reassure committees, not to run companies. It is the specific plan this operation’s evidence demands. Where the buyer wants operating support after close, that is a separate conversation about a separate engagement; the assessment stands on its own either way.

05What it looks like

What an engagement looks like

  • A defined diligence window, aligned to the deal timetable
  • Works from the data room, management sessions and live operating evidence
  • A board-grade written assessment, risks ranked by financial exposure
  • Available as a pre-LOI red-flag read or in full post-LOI depth

Outcomes

  • Operational risk priced into the deal instead of discovered after it
  • Management claims tested against evidence, not presentation
  • A hundred-day plan grounded in diagnosis rather than template
  • A written read the investment committee can defend

Questions

Common questions.

Operational due diligence is an independent assessment of a target company’s operating model, run before a transaction commits you to it. Where financial DD examines the numbers and commercial DD examines the market, operational DD examines the machine: whether the delivery engine can carry the growth case, whether quality is a system or a story, how dependent the business is on its founder, where margin leaks through rework and credits, and how cash actually moves. The purpose is pricing and planning: risks sized before they are yours, and a first-hundred-days agenda grounded in evidence. The deliverable is a written, board-grade assessment an investment committee can interrogate.

Deal teams are excellent at what deal teams see: the model, the market, the paper. The operating model resists that toolkit because its truths are not in documents — they are in how work moves, where decisions queue, and what the workflow does when the process map is not watching. Reading that requires having run operations, not analysed them. The complement works cleanly in practice: the deal team owns the transaction and the model; I supply the operating read that either validates the assumptions or reprices them. On several classes of finding — quality reality, delivery capacity, cash-cycle behaviour — an operator’s week is worth more than an analyst’s month, because pattern recognition does the work.

Six things, weighted to the thesis. The delivery engine: capacity, throughput mechanics, and whether scale-up is achievable at the assumed cost. The quality system: standards, measurement, defect handling — evidence versus assertion. Founder- and key-person dependence: the relationships, decisions and knowledge that leave when people do. Margin integrity: rework, credits, unbilled work and the other leaks that rarely have a line in the accounts. Cash-cycle mechanics: how billing actually runs and where cash actually sticks. And the organisation itself: decision rights, cadence, and whether the management layer runs the business or narrates it. Each area produces findings ranked by exposure and difficulty — aimed at the price and the plan.

A full assessment typically fits inside two to four weeks depending on access and the size of the operation — designed to run inside a standard exclusivity window, in parallel with the financial and legal workstreams rather than after them. Where the timetable is tighter, a pre-LOI red-flag read compresses to roughly a week: narrower, blunter, aimed only at the findings large enough to change the price or kill the deal. What I will not do is stretch the language to fit a timetable that cannot support the evidence — if access constraints mean a finding is provisional, the report says so plainly. A diligence product is only worth its evidence discipline.

A written assessment built for an investment committee: an executive read of two or three pages that states plainly what holds and what does not; findings ranked by financial exposure with the evidence behind each — what was examined, what it showed, how confident the conclusion is; the specific CIM claims that did and did not survive testing; and a first-hundred-days agenda sequenced by risk. No hundred-slide theatre — the length is governed by what the evidence supports and what a committee will genuinely read. Where it is useful, I present the findings to the committee directly and take the questioning; a read that cannot survive interrogation is not one you should price from.

Yes — a red-flag review built for exactly that stage: roughly a week, working from the data room, management sessions and targeted questions, aimed only at the operational issues big enough to reprice or kill the transaction. Founder-dependence the thesis ignores, a delivery engine visibly at capacity, quality machinery that is presentation rather than system, cash-cycle behaviour that contradicts the model. It will not produce the full evidence trail or the hundred-day agenda — that depth needs post-LOI access — but it answers the question that matters early: is there an operational reason to walk away, renegotiate, or proceed with specific conditions attached? The red-flag read is built to convert to full depth once exclusivity opens the doors.

More than a data room, less than a disruption. The workable minimum: management working sessions with the operators who actually run delivery, quality and billing — not only the CFO; a walk-through of the core workflow from sale to delivery to cash; system evidence where it exists — quality scores, utilisation, revision counts, credit notes, invoice ageing; and a sample of live or recent work. Sensitivity is normal at this stage, and the process respects it: sessions are framed as operational discussion, requests run through the deal team, and nothing reaches the target that signals conclusions. Where access is genuinely constrained pre-close, the report marks each finding’s confidence accordingly — evidence discipline includes being honest about the evidence you could not get.

I will answer that precisely, because diligence providers often blur it. My background is not deal-side; it is operating-side, and that is the point of the service. Nineteen years running operations, quality and governance at scale — auditing quality across 2,000+ campaigns and 450 clients, building measurement that protected more than USD 20 million in billings, re-engineering a billing cycle across 75 entities — is the craft of reading operating models against evidence, which is what an acquirer actually needs pointed at a target. The financial, legal and commercial workstreams belong with your existing advisers. What I add is the workstream deal processes structurally lack: an operator who has run machines like the one you are buying, reading this one before you own it.

Weigh three things. First, operating scars over deal counts: many ODD providers are former consultants or deal professionals; the best operational due diligence comes from people who have actually run operations at scale, because pattern recognition is the product — ask every candidate what they have operated, not how many processes they have reviewed. Second, evidence discipline: ask to see a sanitised report structure — findings ranked by exposure, traceable to evidence, confidence stated, rather than a hundred slides of observation. Third, independence from the fix: a provider positioned to sell the post-close transformation has an incentive to inflate findings. Whoever you choose, insist the assessment stands alone and survives committee questioning without its author in the room.

Yes, where it makes sense — and the assessment is designed so you are never dependent on it. The natural continuations: supporting the hundred-day agenda the report defined; a fractional COO engagement inside the portfolio company to install the cadence, measurement and quality systems the diligence found missing; or a standing advisory line to the board or the CEO through the ownership transition. Two disciplines keep this honest. The diligence conclusions are never shaped to generate follow-on work — the report would say the same things if we were certain never to meet again. And post-close work is scoped as its own engagement on its own merits, with the diligence handed over cleanly either way.