A diligence list rarely fails because it is too short. By the tenth or fifteenth transaction, the list is comprehensive, the data room index is standardised, and the advisers know what to pull. The deals that go badly are almost never the ones where somebody forgot to ask. They are the ones where the question was asked, the answer came back, and everybody accepted it.
So the useful conversation for people who buy companies repeatedly is not which due diligence questions to ask in an acquisition. It is which answers have historically correlated with trouble eighteen months after closing. In my own work sourcing and evaluating investments, three areas have earned more of my attention than the rest: where the revenue is concentrated, where the judgment sits, and the distance between what gets reported and what gets collected.
Provenance matters more than content
Sellers prepare for the primary question. Every management team running a process has rehearsed the top-five-customer answer, the churn answer, and the why-is-the-founder-selling answer. You learn very little from any of them, because you are hearing a product, not a fact.
What you learn from is how the answer arrives. Ask for the same number from three directions — revenue by customer out of the accounting system, the same cut out of the CRM, and cash receipts by customer for the same period. They will not match exactly. That is fine. The diagnostic question is whether anyone inside the company already knew they did not match, and could explain the difference without going away for a week.
A business that can reconcile its own numbers on request is a business where somebody has been doing that work all along. A business that produces a figure once, from a spreadsheet nobody else can operate, has given you an opinion in the costume of a number. That is true whether the figure flatters the seller or not.
Concentration is about who decides, not who pays
The top-five-customers percentage is where concentration analysis starts and too often where it stops. Three refinements have been worth more to me than the headline figure.
First, look at concentration in gross profit rather than revenue. A large customer that has negotiated itself down to a thin margin is contributing volume and operating leverage but very little to the earnings you are buying. Concentration in margin is usually the more dangerous version, and it is rarely the number in the deck.
Second, count decision-makers, not logos. A dozen accounts can sit behind one procurement office, one group purchasing organisation, one distributor, or one referral relationship. The same applies to acquisition channels. A company with thousands of customers and a single source of new ones is concentrated, and increasingly that single source is a platform or a search result rather than a person. I have written elsewhere about how deliberately a company chooses its customers; the reverse question in diligence is whether it has ever had a choice at all.
Third, read the renewal mechanics rather than the contract term. Auto-renewal with a notice window behaves nothing like an annual re-bid, even where the stated term is identical. Check assignment and change-of-control language on every material agreement, and then ask when each of those customers was last spoken to about the transaction. A contract that survives the closing legally can still be lost commercially in the first quarter.
The failure mode specific to experienced buyers is the confidence that concentration will be diluted after closing. You have done it before, so you underwrite it. Diversification is real work and it reliably takes longer than the hold assumption in the model. If the plan is to fix concentration, it belongs in the plan with an owner and a cost, not in the narrative.
Key-person dependence hides in the exceptions
Never ask whether the business could run without the founder. Nobody has ever said no. Ask instead for the last twenty exceptions: discounts outside policy, credit extended beyond terms, a specification changed mid-build, a hire above band, an escalated customer. Get the approver for each one.
If one name appears on eighteen of the twenty, you have a dependency, regardless of what the organisation chart says. That is not a criticism of the person. Founders accumulate exception authority because they are usually right, and because saying yes quickly is a competitive advantage in a small company. But it means the operating judgment has not been written down, and you are buying it as a retention risk rather than an asset.
Then go one layer down and ask those managers to state the rule rather than the outcome. People who own a decision can tell you the threshold at which their answer changes. People who merely execute it can only tell you what happened last time. That difference is the core of management diligence, and it does not show up in a CV or a management presentation.
Watching matters more than interviewing. On boards, I ask to observe a meeting before I accept a seat, because an hour of watching people disagree with each other tells you more than a year of reading packets. The same instinct applies here. Sit in on a genuine operating review if you can get access to one, and notice who talks, who gets interrupted, and who is asked for the number.
The gap between reported and collected
A quality of earnings report will surface most of what is wrong with the revenue. The problem is what buyers do with it. Under time pressure, a QoE gets read for its adjusted figure and its net working capital peg, and the exceptions section — the part written in careful, hedged language — gets skimmed. Read the exceptions first and the summary last.
The specific gap to chase is between what the income statement reports and what the bank account collects. Look at the ageing, and then look for the balances that never move from one bucket to the next. Look at credit memos issued after each period closed, which is where revenue pulled forward gets quietly returned. Look at days sales outstanding by quarter rather than by year, because an annual average will hide a two-year drift. Look at whether deferred revenue is shrinking while reported revenue grows.
On the add-backs, one question has been more revealing than any schedule: which of your own adjustments would you refuse to accept if you were the buyer? Sellers who have thought honestly about their business will name one or two without much hesitation. Sellers who defend every line have told you something about how they will behave during the transition, which is information you cannot get from the file.
And separate recurring from merely repeating. Contracted revenue with a renewal obligation is a different asset from habitual re-purchase, even where the historical retention curves look the same. The second one is a customer relationship. It transfers with the people who hold it, and you have already established who those people are.
What to do with a flag you actually believe
Experienced buyers rarely miss red flags in a business acquisition. What they do is find the flag, price a little of it, and then talk themselves into fixing the rest, because they have fixed things like it before. That confidence is usually earned and occasionally fatal.
There are three honest responses to a material finding: reprice it, structure around it with an escrow, earn-out or holdback, or walk. Fixing it is not a fourth option — it is a project, and a project needs an owner, a date and a line in the model. If it cannot survive that translation, it is not a plan. The work of turning a thesis into something operable begins well before the closing dinner.
Before signing, I find it useful to write the memo you would write if the deal had already gone wrong. Assume three years have passed and the outcome was poor. What does the memo say you already knew? The flags you can name in that exercise are the ones you are choosing to accept, which is a legitimate decision, as long as it is a decision. A disciplined review process after the fact is what turns those choices into judgment rather than folklore.
Keeping your own record
The most valuable diligence asset a repeat buyer can build is not a better checklist. It is an honest internal record of which flags actually predicted trouble in your own deals and which ones you worried about for nothing. Most firms never keep it, so every partner carries a private, unexamined theory of what matters. Write it down, revisit it after each exit, and let the list earn its length. Our questions should get sharper with each transaction — and a few of them should get dropped.