Inventory Audit: How to Run One Properly in 2026
Two very different exercises share the name inventory audit.
The first is the accountant's version. Count the stock, reconcile it to the books, investigate the gaps. It answers one question: does the inventory on your balance sheet actually exist?
The second is the operator's version. It answers a more expensive question: what is that inventory doing to your cash and your margin? How much money is sitting in stock that will never sell? How much revenue are you losing to the stock you failed to hold? How far does your cover actually stretch against your supplier's lead time?
Almost every guide on this subject covers the first version and stops. This one covers both, because the counting audit is table stakes, and the money audit is where founders find six and seven figures. Here is how to run each, what to measure, and a checklist you can lift straight into a spreadsheet.
What is an inventory audit
An inventory audit is a structured check of your inventory records against reality. At minimum, that means verifying that the units your system says you hold physically exist, in the right locations, in sellable condition, and that their recorded value is right. Done properly, it extends into how that inventory is performing: what is moving, what is dead, what keeps running out, and what all of it costs you.
Lenders and investors sometimes require the accounting version. Nobody requires the operational version. Which is odd, because it is the one that changes your P&L.
The counting audit: getting the record straight
You cannot audit the economics of stock you cannot count, so the reconciliation work comes first. The standard procedures are worth knowing and none of them is complicated.
A full physical count is exactly what it sounds like: every unit counted and matched against the system, usually annually, ideally with receiving and dispatch paused (a cutoff) so nothing moves mid-count.
Cycle counting is the scalable alternative, counting a rotating slice of SKUs weekly or monthly so the whole catalogue gets verified across a cycle without stopping operations.
ABC analysis focuses effort where the value is, counting your high-value A items more often than the long tail.
Reconciliation is the investigative step when counts and records disagree: tracing the gap back through receipts, dispatches, returns and adjustments until you find whether the cause is shrinkage, mis-picks, unrecorded damage or simply bad data entry. And matching supplier invoices against goods actually received catches the quiet losses, short shipments you paid for in full.
Run this layer on a rhythm: cycle counts continuously, a full count once a year, and every discrepancy investigated to root cause rather than adjusted away. If your physical stock and book stock disagree by more than a few percent, you do not have a counting problem, you have a process problem, and no amount of recounting fixes a broken receiving or returns workflow.
That is the audit most articles describe. Now the part they leave out.
The operator's audit: six numbers that find the money
Once the record is trustworthy, audit what the inventory is doing. Six measures, each computable from a sales export and a stock snapshot, each pointing at a specific pile of money.
Days of cover by SKU.
Units on hand divided by average daily units sold, per SKU, not blended. The blended number hides everything. It is common to open this analysis and find bestsellers carrying three weeks of cover while dead lines carry two or three years of it. Every SKU sitting far beyond your target cover, most DTC brands land somewhere between 60 and 120 days depending on lead times, is cash doing nothing.
Dead stock cash value.
Take every SKU below a minimum sell-through threshold, multiply units by unit cost, and total it. That figure is money you already spent, sitting on shelves, incurring storage, and quietly heading toward markdown. It is usually the single most sobering number in the audit, and it is the one that funds the recovery plan: structured clearance, bundling, or channel offload, ranked by contribution margin rather than blanket discounting.
Stockout days and the margin they cost.
For each SKU, count the days out of stock in the period, multiply by the average daily units it sells when in stock, and value the lost units at gross margin, not revenue. Then be honest about recovery: some buyers wait and purchase on restock, some substitute another of your SKUs, so the true loss is a share of the gross figure. This is the mirror image of dead stock, the cost of holding too little, and putting both numbers side by side is what turns inventory from a storage question into a capital allocation question.
Cover versus lead time.
For each core SKU, compare weeks of cover against the supplier's true lead time, production plus transit plus receiving. Any SKU whose cover is shorter than its lead time is already out of stock, the system just does not know it yet. This single comparison is the origin of a working reorder trigger, and its absence is the root cause behind most chronic stockout patterns we audit.
Sell-through by collection or drop.
Units sold as a share of units bought, per launch. It tells you whether your buying is calibrated or hopeful, and after a few drops it becomes the evidence base for sizing the next one.
Inventory accuracy rate.
The bridge between the two audits: the percentage of SKUs where the physical count matched the book. Track it over time. If it is not improving, the counting layer above is theatre.
None of this requires software you do not already have. A Shopify or ERP sales export, a stock snapshot, and a spreadsheet will produce all six. What it requires is the discipline to run it, and the willingness to act on what it says.
The inventory audit checklist
Step | What you are checking | What it tells you |
|---|---|---|
Physical or cycle count | Units on hand vs system record, by SKU and location | Whether your data can be trusted at all |
Condition and location check | Sellable vs damaged, correct bin and site | Stock you hold but cannot actually sell |
Reconciliation | Root cause of every count discrepancy | Shrinkage, mis-picks, process failures |
Invoice vs goods received | Supplier invoices against actual receipts | Short shipments you paid for |
Days of cover by SKU | On-hand units vs daily run rate | Overstock and the cash inside it |
Dead stock valuation | Slow and non-moving SKUs at unit cost | The markdown problem, quantified |
Stockout cost | Days out of stock x run rate x margin | Revenue lost to under-buying |
Cover vs lead time | Weeks of cover against true replenishment time | SKUs already failing, invisibly |
Sell-through by drop | Sold vs bought, per launch | Whether your buying is calibrated |
Accuracy rate trend | Match rate over successive audits | Whether any of this is improving |
Run the top half quarterly at minimum and the bottom half monthly. The bottom half is a reporting pack, not an event: once the queries are built, it refreshes in an hour.
Doing it yourself versus bringing someone in
The counting layer is genuinely DIY, and if your team is small, it should be. The operational layer is also DIY in mechanics, the maths above is not sophisticated, but two things make an outside pass worth considering.
The first is bias: the person who bought the stock is the wrong person to declare it dead, and the person who built the reorder process will struggle to see its gaps.
The second is benchmarks: knowing your pick cost, cover targets and stockout rates are out of line requires knowing what in-line looks like across other brands at your scale.
That is the shape of our own supply chain and operations audit, which covers this inventory work alongside freight, fulfilment and planning, prices as a fixed fee by revenue band, and identifies at least three times its fee in quantified opportunity or it is free. As a reference point for what the money layer finds in practice, a single audit for a pet-accessories brand identified just over one million dollars a year in operational savings, with inventory and freight carrying much of it.
Frequently asked questions
What is the difference between an inventory audit and a stocktake? A stocktake is the count. An inventory audit is the count plus the interrogation: reconciling gaps to root cause, valuing what the stock is doing to your cash, and testing whether the processes behind the numbers work. Every audit contains a stocktake; very few stocktakes amount to an audit.
How often should we audit inventory? Cycle counts continuously, a full physical count annually, and the money metrics monthly once built. If you have never run the operational layer, run it now rather than waiting for a year-end, because dead stock and stockouts compound daily.
Cycle counting or a full count? Both, for different jobs. Cycle counting keeps accuracy honest all year without stopping operations. The annual full count catches what the cycles miss and satisfies anyone external who needs a point-in-time number.
Do we need an accountant for this? For the valuation your accounts require, possibly. For the operational audit, no. It needs your sales data, your stock snapshot, and someone with the independence to say which stock is dead and which processes caused it.