Cash-intensive businesses: how to assess whether the takings are plausible
Most guidance on cash-intensive businesses tells you that cash is a risk factor and stops there. That leaves the analyst holding an alert on a barber shop with no method for deciding whether £14,000 a month is plausible. This note is about the arithmetic of plausibility, and about resisting the lazy conclusion that cash itself is the finding.
Start from the right question
A car wash banks cash. A takeaway banks cash. A market trader banks cash, often in denominations that look untidy on a deposit slip. None of that is a finding. It is a description of the trade.
The question that actually matters is narrower and much more useful: do the declared takings hang together with everything else you can observe about the business? Opening hours, staff count, premises size, supplier invoices, the card terminal record, the rent, the rates. If those things point roughly the same direction, you have a plausible business banking cash. If they pull apart — if the invoices support one scale of operation and the deposits support another — then you have something worth writing down.
I have sat in enough quality assurance sessions to know how often that distinction collapses under time pressure. An analyst opens an alert, sees a cash-intensive SIC code, and writes a conclusion that amounts to the customer deposits cash and cash is high risk. That is not an assessment. It is a restatement of the customer's industry. In the teams I have run, that narrative comes straight back.
Why sector-wide de-risking is a poor answer
Before the method, the caveat, because the method is dangerous without it.
When a bank decides it no longer serves car washes, or money service businesses, or Somali-owned convenience shops, it does not reduce the flow of illicit funds. It moves the flow somewhere less visible and it strips legitimate traders of their ability to bank. The FATF has been explicit for years that wholesale de-risking runs against the risk-based approach it promotes, and the FCA has taken a similar line with UK firms. The people who suffer are overwhelmingly small operators with thin margins and no alternative provider.
So hold both ideas at once. Cash businesses carry genuine laundering risk, because cash breaks the audit trail and because integration into a legitimate till is one of the oldest methods there is. And the overwhelming majority of cash businesses on your book are exactly what they say they are. Your job is to tell them apart with evidence, not to reach for a sector-level shortcut.
Build a rough model of the business before you look at the deposits
This is the step analysts skip, and skipping it is why so many cash reviews end up circular. If you look at the banking first, you anchor on the number and then hunt for reasons it might be fine. Build the model first, then compare.
You want a range, not a figure. Three inputs get you most of the way.
Capacity. How many customers can this business physically serve in a day? A single-chair barber shop doing cuts at twenty minutes cannot serve ninety people. A hand car wash with four staff and one bay has a ceiling set by how long a wash takes. A takeaway's ceiling sits in the kitchen, not the till. Walk it through in your head: opening hours, minus realistic idle time, divided by service time, multiplied by service points.
Average transaction value. Often the easiest thing to verify, because most of these businesses publish a price list. A menu, a shopfront board, a Facebook page with prices from eighteen months ago. Take the mid-range item, not the most expensive, and assume people buy drinks or extras at a modest rate.
Cost base. Supplier invoices, wholesale purchases, rent, wages, utilities. A takeaway buying about £3,000 of stock a month is not producing £40,000 of food. This is the single most powerful test I know for cash businesses, and it is under-used because it requires you to actually request and read the documents. Food cost as a proportion of revenue sits in a fairly predictable band for a given cuisine and format. Car washes buy chemicals and water. Barbers buy very little, which is precisely why barber shops are harder to test and why you have to lean more heavily on capacity.
Multiply capacity by average transaction value, sanity-check against the cost base, and you have a plausible monthly range. Now open the statements.
What the deposits should look like, and what to make of it when they don't
Compare your modelled range to the banked total. Three outcomes.
Deposits sit inside the range. Document the model, note the inputs you used, close it. That file will save the next analyst forty minutes when the same rule fires in six months.
Deposits sit below the range. This is common and usually uninteresting from a laundering perspective — cash retained for wages paid informally, or takings not fully banked. It may be a tax matter rather than a money laundering matter, though in some jurisdictions the two overlap in ways that affect your reporting decision, so check your own framework rather than assuming.
Deposits sit meaningfully above the range. Now you have work to do, and the work is to look for the explanation you have not thought of before you look for the one you have. Businesses get busier. A new takeaway near a stadium has match days. A car wash near a new housing estate grows. A trader may have started a second pitch. Ask.
The card-to-cash ratio, used carefully
Card acquiring data, where you have it, is the closest thing to an independent witness. If the card terminal processes £4,000 a month and the account receives £31,000 in cash, that is a ratio of roughly one to eight. Is that possible? In some trades, yes. Hand car washes and market pitches genuinely skew heavily to cash. In a town-centre takeaway with a delivery app presence, a ratio like that is much harder to square, because the app payments arrive electronically and customers who order in person increasingly tap.
What makes the ratio powerful is movement over time. A business that ran at sixty per cent card for two years and then drops to twenty per cent card, with total turnover flat or rising, has changed something. Perhaps the terminal broke. Perhaps they switched acquirer and you are only seeing part of the picture. Perhaps something else. The shift is the signal, not the level.
Case note
A composite from two reviews several years apart. A hand car wash, incorporated about three years earlier, single director, premises on a leased forecourt. Between March and August one year the account received cash deposits averaging about £9,400 a month, split across four to six branch visits. Card acquiring through the same institution ran at about £1,100 a month. The rule that fired was a simple cash-to-turnover threshold.
The capacity model did not obviously fail. Six staff visible on the business's own social media, a posted price of £8 for an exterior wash and £15 for a valet, open seven days. At a blended £11 and about thirty cars a day, you get to roughly £9,900 a month. The numbers hung together well enough that the first analyst closed it, reasonably.
What surfaced eleven months later was a second pattern. From the following February, deposits rose to about £21,000 a month while the card volume stayed at £1,050. The price list had not changed. The forecourt had not changed. Staffing on the wage runs had gone from six to seven. Thirty cars a day had become, on the arithmetic, something above sixty-three, on the same site, in February. When the relationship manager asked, the director said trade had improved. No revised price list, no second site, no explanation for why card takings had not moved at all while cash had more than doubled. That combination — one input moving hard while every correlated input stays frozen — is what went into the narrative, and it is what made the narrative defensible. Not the cash. The divergence.
Documents, and how to ask for them without being a nuisance
Most of what you need is ordinary business paperwork. Filed accounts, VAT returns where applicable, a sample of supplier invoices covering a defined month, a copy of the current price list, the lease, and a wage summary. If you are running a full review rather than a triage, that set overlaps heavily with what an EDD file should contain anyway, so you are not doing duplicate work.
Two things I have learned about the asking. First, request a specific month and be precise about it, because "send us some invoices" produces a random handful that tests nothing. Second, expect informality. A market trader may hold receipts in a carrier bag and have no accountant until January. That is a records issue, not evidence of anything, and treating it as evidence of something is how analysts end up writing conclusions their own QA cannot support.
Where the deposits come in structured shapes — repeated amounts just under a reporting or reporting-adjacent level, spread across branches or days — that is a separate pattern with its own logic, and what structuring looks like on a bank statement is worth reading alongside this. Cash intensity and structuring often appear in the same file, but they are different findings and should be written as such.
Writing it up
Whatever you conclude, the file should let a reader reconstruct your arithmetic. Inputs, the range they produce, the banked figure, the gap, and what the customer said about the gap. I would rather read a two-paragraph assessment with real numbers in it than a page of hedged prose about cash being inherently high risk.
If it goes to a report, the divergence is your spine. Guidance from bodies like JMLSG in the UK and the Wolfsberg Group internationally consistently frames suspicion as something built from inconsistency rather than from sector membership, and that framing is what keeps the report useful to the financial intelligence unit receiving it. The structure of that write-up follows the same discipline as any other — there is a worked example of a SAR narrative in these notes that shows the shape.
And if it does not go to a report, say so plainly and say why. A closed alert with a documented plausibility model is a genuinely valuable artefact. It is also the thing that stops the same business being reviewed from scratch, by a nervous new analyst, every single quarter.
Worth remembering: The finding is never that a business handles cash — it is that one number moved while every number that should have moved with it stayed still.