Due diligence is the investigation you run between signing a letter of intent and closing, to find out whether the business is what the seller says it is. Every buyer knows they need a checklist. A good one runs past a hundred items. The trouble is that a hundred-item list, worked top to bottom, turns into a pile of documents nobody reads in time, and the item that would have killed the deal sits in the middle of the pile.
This lesson turns the checklist into four workstreams, each with an owner and a deadline, and adds the instrument that experienced buyers rely on and first-time buyers skip: a scorecard comparing what you asked for with what you actually received.
Principle
Diligence is not a checklist, but you need one. The list makes sure nothing is forgotten. The workstreams make sure each item gets read by the person qualified to read it, in time to matter. Yale's 2022 primer on diligence in search fund acquisitions, written for traditional search funds, organizes the work into preliminary and confirmatory phases across eight pillars (M19, 2022). For a self-funded buyer with no team, four workstreams are easier to run:
| Workstream | What it tests | Typical owner |
|---|---|---|
| Financial | Earnings, tax returns, bank deposits, working capital, debt, the quality of earnings report | You, then your CPA or quality of earnings provider |
| Legal | Entity and ownership, contracts and their assignability, litigation, liens, licenses and permits, the lease, intellectual property | Your deal attorney, with your own public-record searches |
| Operational | Customers, suppliers, equipment, systems, pricing, the sales pipeline | You |
| People and culture | Owner dependency, management depth, key employees, compensation, retention risk | You, with an employment attorney on anything unusual |
Each item on the master list belongs to exactly one workstream and one owner, with a due date inside your exclusivity window.
The biggest diligence mistake is confirming what you want to believe. By the time you are under a letter of intent, you have spent months and money and you want the deal to work. That is when you stop looking for what could kill it. A study of five failed takeovers of small and medium-sized firms by outside buyers, based on interviews with both sellers and acquirers, traced the failures to five managerial errors rooted in the acquirers' cognitive biases (B12, 2014). Write down, before diligence starts, the three findings that would make you walk. Then go looking for them first.
What the seller delivers is a finding, separate from what the documents say. Every request you make has three possible outcomes: you get what you asked for, you get something else, or you get nothing. The second and third outcomes are information.
Here is the most common example. You ask for three years of the business's tax returns. The seller sends his personal Form 1040. If the business is a corporation or a partnership, which includes most multi-member LLCs, it files its own return (Form 1120-S, 1120 or 1065), and the personal return shows only what flowed through to the owner. That substitution may be innocent: a seller who has never sold a business and does not know the difference. It may be a seller who keeps the entity return away from you because it does not match the earnings in the broker's memo. Either way, the gap is a finding, and you log it before you ask again. (A single-member LLC taxed as a sole proprietorship reports its business income on Schedule C of the owner's 1040, so in that case the 1040 with its Schedule C is the right document. Know which kind of entity you are buying before you judge the delivery.)
Your lender will check the same thing on its own. Under SOP 50 10 8.1, effective October 1, 2026, lenders on a 7(a) change of ownership must verify the seller's financial information against IRS tax transcripts, and for initial acquisitions and business expansions with a business purchase price of $3 million or more, the lender must obtain a quality of earnings report that includes a cash proof (a reconciliation of bank statements to the income statement and tax returns) (Y25). The lender "must use the report's findings to calculate the Debt Service Coverage", and if that coverage does not support the structure, "the loan amount must be reduced accordingly." Above that threshold, the report's earnings are the lending number, not yours. That report is performed for the lender's benefit and must not be prepared by or for the seller; if you commissioned your own beforehand, the lender may have one of its approved vendors review it, and may not rely on it without that review (Y26). Better that you find the gap first.
Diligence catches sins of commission and misses sins of omission. On episode 2 of the Still Searching podcast, Drew Williams described what surfaced after he took over his company: issues the owner left behind not through a sin of commission, but a sin of omission, which simply never got brought forward. Old customers who came back for warranty work on jobs from two years earlier, and did not care that a new entity now owned the name. Documents can expose what someone did and recorded wrongly. They rarely expose what nobody wrote down. That is why diligence is paired with protection in the purchase agreement (representations, warranties, indemnities and a right to offset claims against the seller note, all drafted by your attorney) and with time spent at the business, learning from the seller before you close. Lesson 7.3.3 covers the liabilities nobody lists.
Case
Coastal Comfort HVAC is the fictional company that runs through this curriculum. Your letter of intent offered $3,800,000: $3,150,000 in cash at closing, $350,000 in Seller Note A on full standby for the life of the SBA loan, and $300,000 in Seller Note B, paying over seven years (lesson 7.1.1). Because the price is $3,000,000 or more, the lender had to obtain a quality of earnings report. You commissioned one on Day 55, three days after the letter was accepted, from a firm on the lender's approved vendor list, and the lender's vendor reviewed it. You ran diligence across workstreams and kept a request-versus-delivery scorecard. Here is what it showed when you wrote up your findings on Day 123:
| Area | Requested | Received | Gap |
|---|---|---|---|
| Financial | Three years of statements, trailing twelve months, bank statements, merchant reports, distributor statements | Complete after two requests | None in delivery: the cash proof ties deposits to reported revenue within 2 percent (the margin problem is F-4) |
| Tax | Entity returns, three years | Latest year at once; earlier two after a reminder | None: returns tie to the profit and loss statement |
| Contracts | Maintenance agreements, master agreement with Beachline (the largest builder customer), top-20 customer list | Agreements yes; no Beachline master agreement exists | Finding: the largest commercial relationship rests on one estimator's friendship and job-by-job purchase orders |
| Legal | Litigation search, license status, lease file | Lease file late and thin | Finding: the lease expires in 14 months with no renewal option |
| People | Employee census, compensation, tenure | Complete | Finding: 8 of 24 field technicians are one extended family, including the senior commercial lead |
Two of those three findings came from the scorecard, not from reading a document. A master agreement that did not exist could not be read; it could only be noticed missing. A lease file that arrived late and thin was a signal to look harder, and the problem was in the dates.
The findings memo, dated Day 123, listed eight findings. It is your written summary, not the first time anyone saw them: each finding went to your lender as it closed, and the purchase agreement was signed on Day 130.
| Finding | What diligence showed | Response | |
|---|---|---|---|
| F-1 | Owner economics | Confirmed. Ray personally holds the relationships on the 6 largest commercial accounts | The transition plan names account-by-account handoffs |
| F-2 | The fiscal 2025 spike | Confirmed as an event, not a trend. The 2026 year-to-date run-rate tracks the fiscal 2024 baseline plus about 6 percent | The price stays on recurring earnings that exclude the spike |
| F-3 | Beachline | Worse than screened: no master agreement at all | Introductions with Beachline's construction manager before closing; $50,000 moved from cash at closing into Seller Note A |
| F-4 | Quality of earnings | Revenue and add-backs held; the fiscal 2025 margin did not. A distributor rebate (-$52,000) and a December cutoff error (-$21,000), plus the qualifying-agent stipend (-$24,000), take recurring adjusted EBITDA to about $920,000 | Re-traded to $3,500,000, agreed Day 104 by amendment to the letter (below) |
| F-5 | Key technicians (new) | 8 of 24 field technicians are one extended family | A $60,000 retention escrow, funded by the seller out of his cash at closing, for two named leads, paid at 6 and 12 months: the senior commercial lead and Luis Ortega, the licensed senior install lead (F-8) |
| F-6 | Deferred revenue | About 810 maintenance agreements confirmed; 9 percent are multi-year prepayments sold at a fiscal 2025 promotional discount | The deferred-revenue part of the working capital peg raised by $21,000, to $75,000, to count the multi-year portion, which the seller's books carried as a long-term liability outside working capital |
| F-7 | Lease | The dispute behind the legal expense was a common-area-charge reconciliation, resolved. The lease expires in 14 months with no renewal option | The landlord signs a 5-year renewal at 8 percent higher rent as part of the assignment |
| F-8 | Licensing | Ray is the only qualifying agent and cannot stay after closing; you cannot qualify the license yourself for two to three years at best | A permanent, paid qualifier: Luis Ortega, already licensed, at a $24,000 yearly stipend with 90 days' notice and a place in the retention escrow; you apply to be approved as financially responsible officer; a second license targeted within 18 months (lesson 7.3.2) |
What the quality of earnings report found. Before the letter went out, your earnings bridge took the seller's $1,308,000 of reported EBITDA to about $1,017,000 of recurring earnings. The arithmetic was right. But take the Harbor Pointe project ($930,000) and the pulled-forward replacements ($300,000) out of fiscal 2025 revenue and about $5,445,000 is left, earning $1,017,000: a margin of about 18.7 percent. Fiscal 2024, normalized the same way, earned about $789,000 on $5,325,000, about 14.8 percent. Removing low-margin project work should not lift the margin of what is left by almost four points on lower revenue. You could not price that question before the letter, so you named it in your override memo, wrote it into the report's scope and kept the quality of earnings condition in the letter. The draft report arrived on Day 98. Revenue was real. The margin was not, in two places, and you added one cost of your own:
| Line | Amount | Finding |
|---|---|---|
| Your recurring adjusted EBITDA before the letter | $1,017,000 | From the seller's documents |
| Q-1 Distributor volume rebate | -$52,000 | The equipment distributor pays a year-end volume rebate, credited to cost of goods. Harbor Pointe and the pull-forward pushed fiscal 2025 purchases into its top tier: the rebate was $88,000, against $34,000 in fiscal 2024. At recurring volume the tier pays about $36,000 |
| Q-2 December cutoff | -$21,000 | Invoices for installs completed in December 2025 were entered in January 2026, so fiscal 2025 cost of goods and year-end payables were understated. The fiscal 2024 cutoff tested clean |
| Q-3 Qualifying-agent stipend (your pro forma) | -$24,000 | The yearly cost of replacing Ray as qualifying agent (F-8). Qualifying was part of Ray's job; once he leaves, it is a new cost |
| Recurring EBITDA after the quality of earnings work | about $920,000 | The number the lender uses and the deal closes on |
The rebate is the lesson. Your bridge removed Harbor Pointe and the pull-forward at the job-cost level: their revenue and their direct costs. It did not remove them at the rebate level, because the rebate arrives as one year-end credit to cost of goods that no job-cost report shows. The spike bought the top tier; take away the spike and $52,000 of the rebate goes with it. The rest of the margin gap held up: a 7 percent price-book increase in July 2024 ran a full year for the first time in fiscal 2025, and about $920,000 on $5,445,000 is a margin of about 16.9 percent, which the price book and the service mix explain. Test the margin of your result against the prior year, then ask what sits in cost of goods that moves with volume.
The re-run. The method is lesson 6.5.2's, on the new number. The value midpoint is 3.75 times $920,000, or $3,450,000. The bad year (revenue 15 percent below its recurring level, about 40 cents of each lost dollar being contribution) now leaves about $583,000 of EBITDA. At your 1.10x floor on $583,000, an all-SBA structure supports a price of about $3,013,000. Your walk-away fell from $3,600,000 to $3,000,000. At the letter's $3,800,000, bad-year coverage on the letter's structure falls to about 0.99x. Below 1.0x is a stop.
The re-trade, Day 99 to Day 104. A re-trade is a change to the price after the letter is signed. On Day 99 you sent the report's summary to Ray and Dana Whitfield, his broker, with a revised price of $3,440,000: the letter's own multiple of 3.74 applied to $920,000. Ray's first answer was no. Dana walked him through the report, and the rebate and the cutoff were his own books. On Day 104 you agreed $3,500,000 (about 3.80 times $920,000) by amendment to the letter: $2,750,000 cash at closing, Seller Note A of $500,000 on full standby (the $50,000 from F-3 is inside it), and Seller Note B of $250,000. Ray gave up $300,000 of price and moved $100,000 more of it onto standby. You paid $60,000 above your opening number to keep the deal, and cash at closing stayed $250,000 under the new walk-away. The report, the amendment and F-3 were all in your lender's hands before the Day 110 financing commitment, which was issued on the closed structure (lesson 7.4.1).
Three rules come out of those six days. A quality of earnings condition in the letter is what makes a re-trade legitimate: condition (c) let you change the number without breaching anything, and a buyer without it negotiates with no lever. Re-trade with the report, not with feelings: send the finding, the arithmetic and one number, and nothing about how disappointed you are. And sellers often refuse, and deals die here. If Ray had held at $3,800,000, the memo would have ended in a walk. That is the system working.
Verdict: pursue with modified structure. After F-3, $50,000 moved from cash at closing into Seller Note A. After F-4, the price itself moved, to $3,500,000.
An SBA-financed deal cannot make price depend on future results through an earnout, and this deal uses no contingent price of any kind. So the ugliest findings are priced three ways. First, the price is rebuilt on the report's recurring earnings, which leave out the Beachline project, the pulled-forward replacements and the rebate they bought. You are not paying for the spike, so you do not need a mechanism to get money back if it does not repeat. Second, more of the price now waits: after F-3 and the re-trade, $500,000 sits in a note that pays nothing until the SBA loan is paid off. Third, the purchase agreement gives you a right to set off indemnity claims against the seller notes, drafted by your attorney and approved by your lender.
If Beachline's construction manager had refused to meet you, or the family had no retention path, the same memo would have ended in a walk. That would have been an equally correct outcome.
Steps
1. Build the master list, then assign it
Start from a broad checklist of a hundred or more items across the four workstreams. Delete what does not apply to this business. Assign each remaining item an owner and a due date. Put the dates inside your exclusivity window, with the last ones at least two weeks before the purchase agreement must be signed.
2. Write your kill list first
Before the first request goes out, write the three to five findings that would end the deal, drawn from your red flags (RF-01 to RF-12) and the five kill conditions of the Walk-Away Doctrine. Put those items at the top of each workstream.
3. Send requests in writing, numbered
One numbered list, sent by email, with the exact document named ("the business's federal income tax returns, Form 1120-S, for the last three years, as filed"). Vague requests produce substitute deliveries.
4. Keep the request-versus-delivery scorecard
For every item record: request number, date requested, date due, date received, what was delivered, and a match grade.
| Grade | Meaning | What you do |
|---|---|---|
| Match | Exactly what was asked | Review it |
| Partial | Some years or pages missing | Ask once more, in writing; log the reason given |
| Substitute | A different document | Log it as a finding; ask why, then ask again for the original |
| Missing | Nothing, or "we don't have that" | Log it as a finding; decide whether its absence is itself the answer |
| Late | Delivered after the due date | Note the pattern; a seller who is late on everything is telling you something |
Review the scorecard every week. Patterns matter more than any single line.
5. Reconcile across workstreams
The financial workstream's revenue should match the operational workstream's customer list. The legal workstream's lease should match the rent on the profit and loss statement. Conflicts between workstreams are where the important findings live.
6. Price or protect what you find
Every finding ends in one of four places: a price change, a structure change (more of the price moved into seller paper, a closing condition), a protection in the purchase agreement, or a walk. Write which one, and why. In a deal financed with an SBA 7(a) loan, a seller earnout is not on the list (lesson 7.4.1).
7. When the quality of earnings report moves the number, re-trade or walk
- Re-run the walk-away on the report's earnings, the same way you built the first one (lesson 6.5.2). Rebuild the bad year too.
- Set the revised price as your letter's own multiple on the new earnings, so the only thing that changed is the fact.
- Decide your limit before you send anything: cash at closing stays under the new walk-away, and bad-year coverage stays at or above 1.10x on the structure you would sign.
- Send the seller and the broker the report's summary, the arithmetic and one number. No adjectives.
- Document any agreement as an amendment to the letter, and give it to your lender before the financing commitment.
- If the seller refuses and the gap cannot be closed inside your limits, walk. You lose the cost of the report, not the cost of the business.
8. Plan for what diligence cannot find
List the obligations a business like this usually carries that never show up in documents: warranty claims, verbal promises to customers or staff, informal discounts. Ask the seller about each one in writing, and give the list to your attorney for the representations section.
Deliverable
A diligence workbook for your deal, with:
- The master list, sorted into four workstreams, each item with an owner and a due date.
- Your kill list at the top.
- The request-versus-delivery scorecard, updated weekly.
- A findings log: each finding, its workstream, its red flag number, and its resolution (price, structure, protection or walk), closed with a one-line verdict: pursue, pursue with modified structure, or walk.
- A short list of sins of omission to cover in the purchase agreement.
This lesson is education, not legal, tax or accounting advice. A licensed CPA should review the financial workstream and a deal attorney the legal workstream of any business you intend to buy.