The month a quality of earnings report halved my price at $49K MRR: a founder diary (2026)

The letter of intent had one number in it and I had read that number roughly four hundred times. Four times ARR. I was at $49,000 MRR in 2026, which is $588,000 of annual recurring revenue, which is $2,352,000, which is a number I had already spent in my head on a house and a very long holiday. The quality of earnings report arrived eleven weeks later and the number in it was $1,162,960.

Quick answer (2026): This is a composite founder diary about what a quality of earnings report does to a solo bootstrapped SaaS company during an acquisition. Three things surprised me. ARR is a run rate and not revenue, so my trailing twelve month recognised revenue was $517,300 rather than the $588,000 on my own dashboard. The single largest adjustment was not a disputed add-back and not the deferred revenue, it was the salary I had never paid myself, normalised to a market rate. And "one time" turned out to have a definition I had never read, which is why most of my add-back list did not survive. This is a founder's account and general reflection, not legal, tax or accounting advice.

The number in the letter and the number in the report

A quality of earnings report is not an audit. Nobody is signing an opinion and nobody is checking whether my books comply with anything. It is a buyer paying an accounting firm to answer one question: if we owned this business next year, what would it actually earn us. Everything in it flows from that question, and once I understood that, every adjustment stopped feeling like an insult and started feeling like arithmetic.

That took me about three weeks.

It is worth being precise about the difference, because I had assumed a quality of earnings report was a kind of light audit and it is nearly the opposite. An audit looks backwards and asks whether the statements are fairly presented under an accounting framework. It is performed for the company, it produces a formal opinion, and its job is compliance. A quality of earnings review looks forwards and asks whether the earnings are repeatable. It is performed for whoever is paying, it produces findings rather than an opinion, and its job is to work out which parts of last year will still be there next year under a new owner.

So an audit can bless a set of books that a quality of earnings review then takes apart, and there is no contradiction in that. My books were fine. My books had never claimed to answer the question being asked.

The practical consequence is that almost nothing in the report was a finding of error. There was one genuine misposting, worth $1,900, and I fixed it in an afternoon. Everything else that moved the number was a difference of framing, applied consistently, by somebody with no reason to flatter me.

The request list ran to sixty one items. Bank statements for thirty six months, the Stripe payout ledger reconciled to the bank, a customer level revenue file by month, contracts for anything above a threshold, and every invoice behind the expenses I wanted added back. Producing it took me eleven working days that I had not budgeted for and could not delegate, which is its own lesson about doing this at a size where you are the finance department.

The firm was not hostile. The associate who ran it was patient with me in a way I did not deserve in week two. But a quality of earnings report is written for the buyer, and the buyer is not interested in the story I tell on my own landing page.

ARR is a run rate, and a run rate is not revenue

Here is the mistake I made, and I think most founders who price a business off ARR make it.

My MRR twelve months earlier was about $38,000. It ended the period at $49,000. I had been reporting $588,000 of ARR because that is what $49,000 times twelve is, and every dashboard I own agrees with me.

But nobody paid me $588,000. They paid me the sum of what I actually billed each month while I was climbing from $38,000 to $49,000, and that number is $517,300. The $588,000 is a forecast of a year I have not had yet. It is a perfectly good operating metric and it is not a revenue figure, and when you multiply a forecast by four and call it a price you are asking a buyer to pay you for growth you have not delivered.

The gap was $70,700. On a growing business, ARR always overstates trailing revenue, and the faster you are growing the wider the gap. Mine was 13.7 percent.

I have never seen that spelled out on any of the accounting firm pages that rank for this topic. They all say "we normalise revenue." That sentence is doing an enormous amount of work.

The salary I never paid myself

This is the line that cost me the most money and it is the one I want to warn people about, because I have not seen a single advisory page put a number on it.

I am the only person in the company. I write the code, answer the tickets, do the sales calls and reconcile the bank. I pay myself in distributions and I had recorded roughly nothing as wages, because that is what a lot of small owner-operated companies do.

The buyer is not buying me. The buyer is buying a business that will need somebody to do all of that on day one, and that somebody will want paying. So the report added a market salary as an operating expense that had never appeared in my books.

They used a published national median. The Bureau of Labor Statistics puts the 2025 median pay for software developers, quality assurance analysts and testers at $134,040 per year. That is the number that went in. Employer payroll taxes sit on top of it and the report included those too, though I am leaving them out of every figure below to keep the arithmetic clean and conservative.

One salary. The lowest defensible one, for only the engineering half of what I actually do.

The bitter part is that the same fact had already been a problem in a different direction. The IRS position on owner-operated S corporations is that they "must pay reasonable compensation to a shareholder-employee in return for services that the employee provides to the corporation before non-wage distributions may be made". My accountant had been raising it politely for two years. I had been ignoring it politely for two years.

The same unpaid salary was a tax exposure and a valuation deduction. One fact, two bills.

The two year test that killed my add-back list

I submitted $31,600 of add-backs. I got $13,400.

Add-backs are the expenses you argue a new owner would not repeat, so they get added back to earnings. I had four. A rebrand at $6,200. A failed migration contractor at $7,200. Conference travel at $8,400. Legal fees at $9,800.

The rebrand and the contractor survived. Travel and legal did not, and the associate explained the rejection by pointing at a standard I had never read. Public companies are bound by Item 10(e) of Regulation S-K when they present non-GAAP measures, and it says a registrant must not adjust a measure to eliminate or smooth items identified as non-recurring "when the nature of the charge or gain is such that it is reasonably likely to recur within two years or there was a similar charge or gain within the prior two years".

My company is private. That rule does not bind me or the buyer. But diligence practice borrows the test, and read plainly it is brutal and fair: I had gone to a conference the previous year, and I had a legal bill the previous year. Different conference, different lawyer, same line on the income statement. Both failed on the second half of the sentence before anyone even argued about the first half.

I had been using "one time" to mean "I am not planning to do that again." The test means "there is no similar charge in the two years behind you and none reasonably likely in the two years ahead." Those are not the same sentence and only one of them survives contact with a buyer.

$18,200 rejected.

The annual plans I had been so pleased about

At $32K MRR I pushed hard into annual prepayment, and it was the correct decision. It fixed my cash flow and it dropped my churn.

It also meant that at the measurement date I was holding $41,000 of cash for service I had not delivered yet. That is not my money. It is a liability, and it sits in the balance sheet as deferred income under the same caption that the SEC's own balance sheet schedule uses for material items of deferred income.

The buyer's position was simple and I could not argue with it. They will deliver eleven more months of service to those customers and collect nothing, because I already collected it and already spent it. So the deferred balance came off the purchase price at close, dollar for dollar.

Not multiplied. Just gone. I had spent that cash on the growth that produced the ARR number I was so proud of.

What each line actually cost

Here is the bridge, as the report presented it.

The conversation then moved off a multiple of ARR and onto a multiple of that, which is a change of metric and not just a change of number. At six times adjusted EBITDA, less the deferred revenue funded at close, the price was $1,162,960 against the $2,352,000 I had been carrying around. Just over half.

That headline gap mixes two different things, the change in metric and the individual adjustments, so it does not decompose cleanly and I am not going to pretend it does. But holding the six times multiple fixed, this is what each adjustment was worth:

Look at the second row against the other three.

The salary I never paid myself cost $804,240. The other three combined cost $574,400. The line I had never once thought about was worth more than the revenue restatement, the entire add-back fight and the deferred revenue put together, and it was the only one where I never even got to make an argument, because there is nothing to argue: somebody does have to do the work.

I spent three weeks fighting over $18,200 of travel and legal receipts. I spent zero weeks on the $134,040.

What actually happened

I did not sell. The revised price was not wrong, it was just below the number where I would rather keep the business, and we shook hands and stopped. I paid for half the report by agreement and I consider it the best money I spent that year.

Here is the ledger, rounded and self-reported, for the months around it. It picks up where the previous month's takedown notice left off.

The direct cost was about $9,000 of my share of the fees. The real cost was eleven weeks of attention in a quarter where I shipped almost nothing.

The one thing I would tell you

Put yourself on payroll now, at a market rate, and run the business on what is left.

Not for the tax reason, though that reason is real and my accountant was right. Do it because it is the only way you will ever see your own numbers. For two years I had been reading a profit line that quietly assumed one senior engineer worked for free, and every decision I made on top of that line was made on a business that does not exist.

The buyer did not invent that adjustment to push my price down. They just refused to keep making the assumption I had been making. The $134,040 was always there. It was coming out of my life instead of my income statement, and the only thing the report changed was that I could finally see it.

I would rather have found that out from my own books in year two than from somebody else's spreadsheet in week eleven of a deal.