Independent stock audits: counts that stand as evidence

Some stock numbers only need to be right. Others need to be provably right — to a statutory auditor, a lender whose facility is secured on the stock, an insurer assessing a claim, or the other side of a dispute. The difference is not counting skill; it is independence and documentation. A count performed by the people who manage the stock, however diligent, is an assertion. A count performed by an independent firm, under controlled cut-offs, with a documented method and a traceable reconciliation, is evidence.

CPCON provides that second kind. We are not an audit firm — we issue no opinion on your financial statements — which is exactly what makes the model work: your auditor observes and tests our count under ISA (UK) 501 without any conflict, and third parties get a number produced by a firm with nothing at stake in what it turns out to be. This page explains why independence carries evidential weight, where the obligation comes from, how an audit-grade count is built step by step, how sampling is decided, what the documentation pack contains, and why doing it internally is a false economy when the number has to convince someone.

Why an independent count exists: evidence, not assertion

The whole field rests on one principle borrowed from the logic of audit evidence: the weight you can place on a number rises with its independence from the party making the claim. Your own team may count meticulously, but their count is still an interested assertion — the result may touch reported profit, a bonus, a bank covenant or a sale price. None of that makes them dishonest; it makes their count uncorroborated. An independent firm with no stake in the outcome supplies the corroboration. That is the single reason auditors, lenders, insurers and counterparties weight a third-party count so much more heavily than an internal one, and it is the reason this service exists.

Independence cuts both ways, which is part of its value. A neutral counter has no incentive to inflate the result to flatter the balance sheet, and equally no incentive to write down good stock to be cautious. We report what the count and reconciliation support — no more, no less — with the trail attached so anyone can follow how the number was reached.

The UK rules that frame it: ISA (UK) 501 and the Companies Act

Two pieces of UK framework sit behind year-end stock counting, and it is worth stating each plainly because they are routinely confused.

The first is ISA (UK) 501, Audit Evidence — Specific Considerations for Selected Items. Where inventory is material to the financial statements, it requires the auditor to obtain sufficient appropriate evidence of its existence and condition by attending the physical count (unless attendance is impracticable). At that attendance the auditor evaluates management’s counting instructions and procedures, observes their performance, inspects the inventory, and performs their own test counts. The key point for clients: the standard does not ask the auditor to do the count. It asks them to be present at a count and to test it. An independent, well-documented count is exactly what they attend and test — and a chaotic internal count is exactly what gives them cause to widen their procedures and their fee.

The second is the record-keeping duty in the Companies Act 2006 s.386. Companies dealing in goods must keep accounting records that include statements of stock held at the financial year-end, and the stocktakings from which those statements were prepared (s.386(4)). In plain terms: the law expects you to be able to show the count that sits behind your year-end stock figure. A documented independent count produces that record as a matter of course, so satisfying ISA (UK) 501 attendance and the s.386(4) record obligation become the same piece of work rather than two separate scrambles.

When an independent count is instructed

Independence is not needed for routine stock control — that is what ordinary internal counting and cycle counting are for. It becomes worth the cost the moment the number has to convince someone with their own interest at stake. The common triggers:

Anatomy of an audit-grade count

What makes a count “audit-grade” is not effort on the day; it is the discipline wrapped around the counting. The four stages below are the spine of every engagement, and the order is not negotiable — a brilliant count with a broken cut-off is worth nothing as evidence.

1. Scope and instructions. Locations, stock populations, valuation basis and count rules fixed in written instructions beforehand — the document your auditor or lender reviews first. Ambiguity here is what turns into argument later; an agreed scope is the foundation the evidence is built on.
2. Cut-off control. Book extract frozen at a recorded timestamp; all movements in the count window logged. Cut-off failures — goods counted but not yet booked, or booked but already despatched — are the classic way counts lose evidential value, so this step is non-negotiable and documented.
3. Independent counting with recount trails. CPCON counters, supervisor blind recounts on sampled sections, and damaged, obsolete and slow-moving stock flagged separately — feeding the net realisable value discussion, not buried in quantities. Blind recounts mean the recounter does not see the first figure, so agreement actually proves accuracy.
4. Reconciliation and reporting. Full book-to-physical reconciliation with cause-coded variances, a signed adjustment schedule and the statement of stocktaking your accounting records require. A net difference on its own tells an auditor nothing; an explained, traceable difference is the whole point.

Methodology and sampling: full count versus tested sample

Not every line warrants a full count, and a competent audit-grade engagement says so up front rather than either counting everything blindly or sampling everything loosely. The approach is driven by materiality and risk, and — crucially — it is written into the instructions before anyone walks the floor, so the auditor can evaluate the method and, if they choose, extend it.

• Full count. High-value and high-risk populations — the lines that move the financial statements — are counted in their entirety. There is no sampling shortcut on the stock that matters most to materiality.
• Value-weighted / statistical sampling. The long tail of low-value lines may be tested on a sample, selected so that selection probability reflects value or risk, with the method documented and the sample size defensible. This is testing with a known basis, not spot-checking on a hunch.
• Condition and NRV flags. Damaged, obsolete and slow-moving stock is flagged separately at the point of count, because it feeds the net realisable value judgement under UK GAAP rather than the existence question. Quantity and value are different problems and we keep them apart.
• Recount thresholds. Variances over an agreed threshold trigger an immediate, documented recount on the floor, so material differences are investigated while the stock is still in front of the counter, not weeks later from a spreadsheet.

Cut-off: the single point most counts fail

If there is one technical reason an otherwise diligent count is rejected as evidence, it is a broken cut-off. Cut-off is the discipline of drawing a clean line in time between what belongs in the count and what does not, so that the physical count and the book figure are measured at the same instant. Get it wrong and the variance you report is not loss at all — it is a timing artefact, and an auditor who spots one cut-off error will reasonably doubt the whole exercise.

The classic failures are mundane and, once named, obvious:

• Goods received not booked. Stock physically on the dock and counted, but the receipt not yet posted — the count shows it, the book does not, and it reads as an overage that is really just paperwork lag.
• Goods despatched not booked. Stock picked and gone but the despatch not yet posted — the book still holds it, the count cannot find it, and a real movement masquerades as a shortage.
• In-transit between locations. Stock moving between sites at the count moment, at risk of being counted twice or not at all.
• Returns and quarantine. Customer returns or quality-hold stock sitting in a limbo location, counted into the wrong status.
• Production work-in-progress. Components consumed into part-finished goods, where the book has issued them but the physical count still sees them on the line.

Our control is deliberately strict because the cost of getting it wrong is so high. The book is extracted and frozen at a recorded timestamp; the count window is defined; and every movement inside that window — every receipt, despatch and transfer — is logged so it can be reconciled out afterwards. Where the operation cannot stop trading for the count, we run controlled documentation of live movements rather than pretending they did not happen. That movement log is part of the evidence pack precisely so the auditor can satisfy themselves the cut-off held.

A worked reconciliation: turning a raw gap into evidence

It is worth seeing how a bare difference becomes an explained, defensible number, because that transformation is the entire value of the service. Suppose the book says a location holds stock valued at £1,000,000 and the physical count comes to £980,000 — a £20,000 gross shortage. Handed over as a single figure, that number tells an auditor almost nothing and tends to be written off as “shrinkage”. Reconciled, it dissolves into causes:

• £9,000 — goods despatched in the count window, not yet posted (cut-off timing; not loss at all).
• £4,500 — a unit-of-measure error on two lines booked in cases rather than eaches (system artefact; corrects on posting).
• £3,000 — damaged and obsolete stock found and flagged for net realisable value write-down (valuation, not existence).
• £2,500 — a genuine, unexplained quantity loss on three high-value lines (the real residual worth investigating).
• £1,000 — a transfer to another site recorded at neither end (corrects once the movement is posted both ways).

Now the £20,000 is not a mystery; it is four explained adjustments and one £2,500 residual that warrants a closer look — the kind of output an auditor can test and a finance team can post. The same logic at scale, across every location and SKU, is what the engagement produces. The mechanics of that decomposition are the subject of our stock reconciliation service, and where the residual losses keep recurring period after period, the shrinkage control programme is what chases them to root cause.

Division of responsibilities: you, your auditor, and CPCON

Confusion about who does what is common, so it is worth being explicit. Three parties have three distinct roles, and the value of the model comes precisely from keeping them separate:

• You (management) own the stock, the records and the assertion in the financial statements. Under the Companies Act you are responsible for keeping adequate accounting records, including the statements of stocktaking behind the year-end figure.
• Your statutory auditor forms an independent opinion on the financial statements. Under ISA (UK) 501 they attend the count, evaluate the procedures, observe, inspect and perform their own test counts. They do not count for you, and they do not certify our work — they test it as part of forming their opinion.
• CPCON plans and executes the physical count and reconciliation as an independent specialist, and produces the documented evidence pack. We issue no audit opinion and we are not the certifying body for any standard — our role is to make the count itself independent, controlled and testable.

Because all three roles are distinct, no one is marking their own homework: management asserts, CPCON independently counts, and the auditor independently tests the count. That triangulation is what gives the resulting figure its weight, and it is why we are careful never to blur into either the management or the auditor role.

Counting around live operations

A frequent worry is that an audit-grade count means shutting the business for a day. Sometimes a full stop-the-clock count on a quiet weekend is the cleanest approach and we will recommend it. But many operations cannot or will not halt, and the count has to happen around live trading or live despatch. That is a planning problem, not a barrier:

• Zoned counting. The site is divided into areas counted and locked-off in sequence, so a counted zone is frozen while the rest of the operation continues elsewhere.
• Controlled live movements. Where stock must move during the count, each movement is documented against the cut-off log and reconciled out, rather than ignored.
• Out-of-hours windows. Counting overnight or before opening removes most of the movement problem at the cost of unsocial hours — frequently the right trade for retail.
• Perpetual-count integration. For operations already running rolling counts, the independent count can verify and re-baseline the perpetual records rather than starting from cold.

The decision about which approach fits is made in scoping, weighing the disruption of stopping against the complexity of controlling live movement — and the answer is driven by what best protects the evidential value of the result.

The deliverable: an audit-grade documentation pack

The count is only as useful as the record it leaves behind. The documentation pack is the actual product — the thing your auditor samples, your lender reviews and your finance team posts from. A complete pack contains:

• Count instructions and methodology — the written scope, rules and sampling approach, issued before the count.
• Team and supervision records — who counted, who supervised, and the blind-recount coverage.
• Cut-off documentation — the frozen book extract timestamp and the movement log for the count window.
• Count sheets or scan files with recount trails — the raw evidence, with first counts, recounts and resolutions visible.
• Book-to-physical reconciliation — every variance cause-coded (cut-off, mis-post, genuine loss, valuation), not a bare net figure.
• Signed adjustment schedule — the agreed postings, authorised and ready for the ledger.
• Statement of stocktaking — the summary record for your accounting records under Companies Act 2006 s.386(4).

Where stock is owned by you but held by a third party, the same pack is produced against your SKU master and the custody contract, so the count stands as evidence in a 3PL or consignment disagreement. And where the verification concerns plant and equipment rather than stock, the equivalent record is produced by our fixed asset verification service, reconciled to the fixed asset register instead of the stock ledger.

Doing it internally: a false economy when the number has to convince

The honest position is that internal counting is the right answer for routine stock control — most counting should be internal, and our stocktaking services and cycle-counting programmes exist to make internal counting better. The false economy is using internal staff when the count has to convince a third party. At that point the saved fee buys a number that the auditor must still test more heavily, the lender may not accept, and the disputing party can challenge as self-interested. The table makes the contrast explicit.

There is a softer cost too. A year-end count run by your own operation pulls staff off their day jobs, compresses into a high-pressure window, and lands the reconciliation on the finance team at the busiest point of the year. An independent count absorbs that load — the field work, the recounts, the reconciliation — and hands finance a posting-ready schedule, which is often reason enough on its own.

Counting at short notice

Many independent counts are not planned a quarter ahead; they are triggered by an event — a fire or flood, a suspected theft, a lender bringing a review forward, an insolvency appointment. We hold field capacity for exactly these situations. The honest trade-off at short notice is planning depth, so the first conversation is ruthless about the two decisions that protect evidential value: getting the cut-off under control, and pinning down the scope. Those two right, at speed, preserve the integrity of everything that follows; everything else can be tightened as the count runs.

Where independent counts recur

Some businesses commission an independent count once, for a specific year-end or transaction. For others it is a standing feature of the financial calendar because their stock is material, mobile or held by third parties. The contexts where the need recurs are worth knowing, because they shape how the count is scoped:

• Retailers with material stock — where shop-floor and stockroom inventory is a significant balance-sheet figure the auditor must see counted, and where loss patterns make the count doubly useful.
• Wholesalers and distributors — high-value, fast-moving stock across large warehouses, where cut-off control and location accuracy are the make-or-break of a credible count.
• Manufacturers — raw materials, work-in-progress and finished goods each measured on a different basis, with WIP cut-off a particular technical challenge.
• Businesses financed against stock — where a lender or invoice/asset-based finance provider requires periodic third-party verification of the secured inventory.
• Owners of stock held by 3PLs — client-owned inventory at a logistics provider, counted independently of both the owner and the warehouse keeper against the custody contract.

For retail and distribution specifically, the operational detail of counting in those environments is on our retail and logistics & warehousing pages. Across all of them the constant is the same: the count has to convince someone other than the people who hold the stock, and that is what independence delivers.

The CPCON difference

Independence is only worth anything if the counting underneath it is sound, and that is where directly-employed teams matter. CPCON counts with its own field people, supervised to a consistent method, rather than assembling a casual crew per job — which is what makes the recount trails reliable and the reconciliation defensible. The methodology behind these engagements has been proven across more than 4,500 inventory and verification projects in six countries over 30+ years, and it treats every count as the reconciliation of three views that must agree: the physical stock on the floor, the logical record in the system, and the financial position in the ledger.

On certification we are deliberately exact, because precision here is part of the trust we are selling: CPCON does not hold or claim its own ISO or SOC certification for this service, and we will not imply otherwise. We are not the certifying body and we are not your auditor — the certifying body certifies your organisation, your auditor forms the opinion, and what we provide is the independent, documented field evidence that makes both of their jobs straightforward. Knowing exactly where our role ends is why third parties can rely on what sits inside it.

Where this fits with the rest of the programme

For routine counting through the year, see our core stocktaking services and cycle counting programmes; if recurring unexplained variances are what brought you here, the shrinkage control service is the follow-through that finds and fixes the cause. The engine under every audit-grade count is the stock reconciliation that explains the variances, and the equivalent independent discipline for plant and equipment is our fixed asset verification. Together they cover the full range from everyday control to the count that has to stand up to scrutiny.

The legal floor is worth stating plainly: Companies Act 2006 s.386(4) requires companies dealing in goods to keep statements of stocktakings behind their year-end stock figures, and where stock is material your auditor is required by ISA (UK) 501 to attend physical counting. An independent, documented count satisfies both at once — and spares your own team the year-end scramble.