A red flag is not a reason to stop. It is a reason to stop moving at the speed you were moving, and to go and find out which of two explanations is the true one.
11 min read · Level 3
The received wisdom is that experience teaches you to spot red flags, and there is something to it - a practitioner who has seen a hundred processes recognises shapes faster. But the more useful skill is almost the opposite, and it is rarely taught: knowing which flags to stop investigating. Every observation can be pursued one level deeper, every explanation can be doubted, and a process that treats depth as a virtue in itself will keep discovering things for as long as it is funded, without ever becoming more able to decide. The flags that deserve depth are the ones where the serious explanation, if true, would change the answer rather than the price. A working capital quirk that would cost 200,000 to fix on a 4,000,000 business is not that; a revenue figure that might not be revenue is. This cuts against how most people feel, because the small resolvable questions are satisfying to close and the large ambiguous ones are not - which is why so many files are thick with resolved detail and thin on the two things that actually mattered.
A red flag is an observation with two plausible explanations, one ordinary and one serious, which you cannot yet tell apart. That is the whole definition, and it is what separates a flag from a finding.
Receivables growing much faster than revenue is a red flag. The ordinary explanation is that a big customer pays on ninety-day terms and signed in October. The serious explanation is that revenue is being recognised on invoices nobody intends to pay. Both fit the same number, and the number alone cannot distinguish them.
What a flag demands is not a decision but a specific further question. It converts an open-ended process into a narrow one: you now know exactly what you are trying to establish, and usually exactly which document would establish it. A process with three live flags and three named follow-up questions is in far better shape than one with no flags at all, which usually means nobody has looked hard enough to raise any.
Because the alternative to a flag is a vague unease, and vague unease does not survive a room. It gets restated as 'I'm just not sure about the numbers', which anybody can respond to, and which the momentum of a live process will flatten within a day. A flag with a number attached - receivables up 52.2% on revenue up 16.7% - cannot be answered with reassurance. It has to be answered with a document.
And because flags have to be counted as well as investigated. Any real business generates two or three; a business with none has been examined superficially. The signal is not the presence of a flag, it is the accumulation of them, and particularly the accumulation of several that would all be explained by the same underlying fact.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
The manufacturer again. Revenue this year 4,200,000 against 3,600,000 last year - growth of 16.7%. Receivables 700,000 against 460,000 - growth of 52.2%.
Convert both years into days. Days sales outstanding is receivables divided by revenue, times 365.
This year: 700,000 / 4,200,000 x 365 = 60.8 days Last year: 460,000 / 3,600,000 x 365 = 46.6 days
Customers are taking 14.2 days longer to pay than they were a year ago. That is the flag, and it is now a number rather than a feeling.
What it is not is a conclusion. Three explanations fit it, and they have completely different consequences. The business may have won a large customer who pays slowly, which is a working capital cost and nothing worse. It may have extended payment terms to hold on to customers who were about to leave, which means the revenue growth was bought and will not repeat. Or it may be invoicing work that has not been accepted, which means part of the receivable is not a receivable at all.
Each of those is distinguishable, and the distinguishing document is specific. An aged receivables analysis showing which invoices are overdue and by how much. The top ten customers' payment terms, this year against last. And the subsequent cash receipts - which of those 700,000 of invoices were actually paid in the weeks after the year end. That last one is close to decisive, and it costs a single request.
Here is the part worth internalising. In the first lesson, the same business showed cash receipts of 3,780,000 against 4,200,000 of booked revenue - a 420,000 gap - and the top three customers at 60.0% of revenue. Take those three observations together: revenue not converting to cash, receivables stretching, and the growth concentrated in a handful of relationships. Individually each has an innocent reading. Together they have a single explanation that fits all three, which is that the growth was bought from a small number of customers on terms the business could not really afford. That is what accumulation means, and it is not visible from any one flag.
The first mistake is treating a flag as a verdict. Walking away from every business with a working capital wobble means never buying anything, because every real business has one.
The second is the opposite: treating a flag as noise because an explanation was offered. An explanation is a hypothesis until a document supports it, and the ordinary explanation is offered with equal fluency whether or not it is the true one.
The third is failing to count. Flags are assessed one at a time as each arises, each is individually survivable, and nobody ever writes the list out on one page. Several flags that would all be explained by the same underlying fact are not three small concerns; they are one large one that has been broken into pieces by the order in which they happened to be discovered.
The fourth is raising a flag without naming the document that would resolve it. 'The receivables look stretched' produces a conversation. 'Please provide the aged receivables analysis and the cash received against year-end invoices in the following six weeks' produces an answer.
The fifth is letting a flag age. A question asked in week two and still unanswered in week nine has not been answered; it has been outlasted, and the process has quietly decided to proceed without it.
Receivables up 52.2% against revenue up 16.7%, or days sales outstanding at 60.8 against 46.6. A quantified flag survives a meeting. An impression does not, and the difference is not about rigour - it is about what can still be raised on the day everybody wants to proceed.
The ordinary one and the serious one, in your own words. Doing this first is what stops the seller's explanation from becoming the only one you ever consider, and it takes two minutes.
For every flag, write the one piece of evidence that would tell the two explanations apart. If no such document exists, say so explicitly - a flag that cannot be resolved by any available evidence is a permanent uncertainty, and it belongs in the decision rather than in the process.
Open, explained, resolved by document, unresolved. One page, read in full at every checkpoint. The accumulation is the signal, and accumulation is invisible when each item lives in the notes of the week it arose.
Read the page and ask what one thing, if true, would account for three of these. That question finds the underlying issue faster than investigating each flag separately, and it is the reason the list has to exist on one page rather than in sequence.
Throughout any diligence process, and afterwards - the same discipline applies to monitoring something you already own, where the flags arrive one quarter at a time and the accumulation is even harder to see.
Do not manufacture flags to demonstrate thoroughness. A list padded with items nobody genuinely finds concerning destroys the signal the list exists to carry, and it trains everyone including you to skim it. Three real flags beat fifteen dutiful ones, and the discipline is being willing to write a short page.
A sequence that recurs across private transactions: a quantified concern is raised, the seller or their adviser offers an explanation that is entirely plausible, everybody in the room relaxes, and the supporting document is neither requested nor supplied. The explanation then appears in the buyer's own file as the resolution. It is rarely a deception - the explanation is usually what the seller genuinely believes - but the flag has been closed by a statement rather than by evidence, and the file no longer records that the distinction was ever available.
Record how each flag was closed, not just that it was. 'Explained by management' and 'resolved by document' are different statuses, and only one of them is verification.
Create a free account to read the rest and apply it to your own investing.