Due diligence & business model analysis · for people buying businesses · Singapore
Every deal has two prices.The business can only afford one.
There is the price you agree to pay. And there is what the business can actually afford, out of the cash it makes each year. When those two are far apart, the business is the thing that fails.
I go through the business before you buy it. That costs you nothing. I am paid later, and only if the business gets better — a quarter of the improvement in the cash it frees up, once that beats the day you bought by six percent.
The call is free. If I am the wrong person for it, I will say so on the call.
I read the accounts of private companies before money moved — analyst, then partner, at a UK investment firm. I have published twelve research papers taking businesses apart from their financial statements.
Nobody is paid when it works.
Not because anyone is dishonest. Because of how they are paid. My fee works the other way round.
On the left, what happens to the business. On the right, what the adviser earned while it happened. Every figure is published, named and linked.
The business
The most common way to lose is to pay too much.
Across 3,000+ deals, 2012–2022, the buyer ended up worse off 57.2% of the time. One of the two reasons KPMG gives: they overestimated the benefits and paid for them.
KPMG, 2025 — public companies, deals over US$100m
The loan is repaid out of the same cash that pays the staff.
Across 484 businesses bought with borrowed money and tracked for ten years, the chance of going bankrupt rose about eighteen percentage points.
Ayash & Rastad, Finance Research Letters, 2021 — buyouts funded with borrowed money
None of this requires anybody to be dishonest.
Everyone in that room can be decent and good at their job. They are still paid more for a bigger deal that closes, and nothing at all for one that does not.
The fee
Your adviser is paid only if you win the auction.
The paper puts it plainly: an adviser working for the buyer is paid only if their client wins the bidding. And the way to win a bidding contest is to be the one who pays most.
Morkoetter & Wetzer, University of St Gallen, 2017
Two to three per cent of the price, plus a bill every month.
2% to 3% of the price when it closes, plus $25,000 to $75,000 every month while you wait. They are paid for the months it takes. Not for whether it turns out to be a good deal.
DealRoom, accessed 2026 — an industry price list, not a study
The fee is only paid if the transaction actually happens.
That is not me being cynical. It is how the academic study itself describes it: the fee depends on the size of the deal, and is paid only if the deal happens.
Morkoetter & Wetzer, University of St Gallen, 2017
Where these numbers do not apply
These are all studies of large deals. Nobody has measured how often this happens to a business your size, so be suspicious of anyone who quotes you a figure for small companies. What these studies show is why it happens. That reason does not change just because the deal is smaller.
Agree a price the business cannot afford, and it pays the difference afterwards — out of the staff, the stock, and the owner's own pocket.
So one question is left before anybody signs: how much can this business actually afford to be bought for? The usual way to answer that is to forecast its future. I stopped doing that.
Where these numbers come from
- 01KPMG, “The M&A Dance”, 2025 — 3,000+ public-company deals over US$100m.
- 02Ayash & Rastad, Finance Research Letters vol. 38, 2021 — 484 debt-funded purchases over ten years.
- 03Morkoetter & Wetzer, University of St Gallen WPF 2015/15, 2017.
- 04DealRoom, 2026 (an industry price list) · MidStreet, 2022 (their own worked examples) · IBBA & Pepperdine Market Pulse, Q3 2012.
less than you think.
Two examples of what I read out of a set of accounts. This is exactly what I would do on the business you are buying.
Who actually pays, and why they keep paying
- What the statements show
- Revenue is one line on the income statement. Underneath it is a list of customers, how long each has been buying, and how much of the total comes from the largest one.
- What they do not
- A million a year from forty customers who cannot easily leave is a completely different business from a million a year from three who can. The revenue line looks identical. What you are buying is not, and neither is what it is worth.
- What I go and check
- If the biggest customer left on Monday, what is still standing on Friday? You do not need a forecast to answer that. You need the customer list and an afternoon. Then I want the reason each customer stays, said in plain words: it costs them money to switch, it costs them time to find someone else, it is habit, you have something nobody else can get, you have a relationship with a regulator or a landlord, or you can do something the others cannot do yet.
What the money costs, and how long it stays cheap
- What the statements show
- An insurer collects premiums years before it pays claims. In the meantime it is holding other people’s money, and the accounts record that as a liability.
- What they do not
- An owner does not ask what next year’s profit will be. They ask three things the accounts can actually answer: how big that pool of money is, what it costs to hold, and how long it stays that cheap. Those three decide the profit. No forecast can rescue you from getting them wrong.
- What I go and check
- All three are in the filings today. It is the same question on an ordinary business — does the cash come in before the bills go out, and for how long has that been true?
↗ Money That Isn’t Yours but Works Like It Is, 2026 — read it in full
I have done this in public twelve times.
Twelve research papers, free and dated, each taking one business apart from its own published financial statements — how it makes money, what protects it, and how long that is likely to hold. The same reading I would do on the business you are looking at.
The framework underneath them is written down too: Copying Without Pride ↗
Twelve businesses, taken apart from their own filings
All twelve on Substack ↗Twelve papers · 27 June – 5 August 2026 · free, no signup
Nobody paid me to write these and nobody edited them. That is what makes them useful to you: you can see how I read a set of accounts before you show me yours.
The companies are large and listed, because that is where the filings are public and you can check my work against the source. What I study is the other end of the range — small and medium businesses in Asia and the United States, under five billion dollars in market value. The reading is the same at either end. The size of the company does not change the method. What matters is whether the facts are public.
It is a number of years.
Pay five times earnings and you have said this business goes on doing what it does for about five more years. That is answerable today. A forecast is not.
The usual way to price a business is to forecast its cash, then add one number standing in for every year after that. In McKinsey's own textbook that single number is most of the answer — and how big it is depends on whether the analyst chose to forecast five years or ten. The business has not changed.
So I do not forecast. I ask how many more years this business can keep doing what it does now — and then I go and check the things that decide the answer.
Damodaran, NYU Stern, 2016 ↗·Mauboussin & Johnson, Credit Suisse First Boston, 1997 ↗
Used in public, on ten industries
The Ship Has a Price Tag · The Invisible Franchise · The Invisible Architecture · and nine more ↗
And one number to put on all four
Add up everything the business has tied up — stock, equipment, premises, the cash it needs to trade. Put the cash it produces in a year against that total. I call the result the cash reinvestment rate. Twenty per cent means the business rebuilds a fifth of itself every year out of its own earnings, without borrowing a penny.
That one number tells you whether growth is free or whether growth has to be borrowed — the difference that shows up in month fourteen, when the repayments start and the shelves need restocking in the same week.
Run it on the business you are looking at.
Runs in your browser · nothing is sent anywhere
7 numbers · about a minute
to have me look.
Free to find out. Paid only if the business you buy actually gets better.
Find out before you pay.
What the business actually stands on, what it owns, and whether the price can be paid out of its own cash — before you sign anything.
It costs you nothing to find out.
No fee to go through the deal, and none for the three years I keep reading the accounts afterwards. No retainer, nothing monthly, nothing at close.
I only do well if the business does.
A quarter of the improvement in the cash the business produces, and only after it is six percent better than the day you bought. If you overpay, it probably never gets there — so I am paid nothing.
What I charge
25%of the improvementabove a 6% hurdle
A quarter. Not of the business, not of the price. Of one thing only.
Free cash flow divided by total equity: the spare cash the business produces, against the owners’ stake on its balance sheet. Whatever that works out to on the day you buy is the starting line.
It has to beat that starting line by six percent before I am owed anything at all. If it never does, I am paid nothing. And if you overpay, it probably never will.
a quarter · of the improvement · after it clears the first six percent
Arithmetic only — how the fee is calculated. Not a projection, not an expected result, and not something I have ever delivered for anyone.
On the day you buy, work out free cash flow divided by total equity, and call that 100. I am owed nothing until it reaches 106 — the same measure, six percent better. Say it reaches 200. I am paid a quarter of everything above 106: (200 − 106) × 0.25 = 23.5. If it never reaches 106, this comes to nothing. Every year, for three years.
The terms
That number ends up more than six percent better than it was on the day you bought
I am paid a quarter of the improvement above that line.
It never clears the line, or you walk away, or you never buy at all
I am paid nothing. Not for going through it, not for the three years after.
What the standard check costs
Having an outsider verify a small company's profits normally runs $10,000 to $35,000 — and $60,000 to over $100,000 at one of the four biggest accounting firms, whose partners bill $700 to $1,000 an hour. In Singapore, the market research alone starts at SGD 10,000 before anyone opens the accounts.
There are cheaper providers than any of those. ClearView publishes $3,900, and what that buys is a formatted report in ten working days — not somebody still reading the statements three years later.
ProjectionHub, 2024 ↗DueDilio, 2026 ↗Assembled, Singapore, 2026 ↗ClearView QoE, 2026 ↗
What you get, and in what order
- 01
A call, free, about forty minutes.
You tell me what the business does and what you are being asked to pay. I tell you whether there is anything here worth looking at — and if I am the wrong person for it, I tell you that on the call.
- 02
You send what you already have.
Three years of accounts if they exist, the customer list, the lease, the licences. No data room, no portal, no forms. If the books are a mess, that is normal, and it usually tells me the most.
- 03
I go and look.
The four checks, on the actual business — who buys, why they stay, what it owns, and what those things would fetch today. Where it matters I go there myself and count what is actually on site.
- 04
A written findings note, and a call to walk you through it.
Plain English, no jargon, every claim tied to a document or a thing I saw. It includes what I could not verify and what would make me walk away. I will tell you how long it will take before you commit to anything.
If you are looking at a business right now, send me a message about it. One paragraph is enough: what it does, roughly what size, and the thing about it you keep coming back to.
It goes to my phone. Nobody else reads it.
Run by Felix — one person, in Singapore. LinkedIn · X · Substack
Not ready to say anything about your business? Read the papers first ↗ You can read every one without telling me your name.
Scope
I do not issue valuations or fairness opinions, and I will not put a single number on what a business is worth. I do not audit or certify accounts — only a licensed audit firm can do that, and if your deal needs one I will say so. I do not give investment, legal or tax advice; what I find goes past your own lawyer and accountant before it goes anywhere else. I am not licensed or regulated, and you should take that into account.