Year-end inventory count: the annual stocktake as an accounting event

Short answer: a year-end inventory count is the physical count that backs the inventory value in your books at the close of the financial year. Operationally it's the same act as any full stocktake — count everything, compare to what Shopify says, adjust. What makes it different is discipline around when you count relative to sales and receipts (cutoff), and documentation: the counted quantities and the corrections become the evidence behind your inventory value, so the count has to be organized, witnessed, and retained, not scribbled on the back of a pick list.

The good news: if you run cycle counts through the year, the year-end event is mostly certification, not discovery. If you don't, this is the count that finds out. Either way, here's the full procedure.

Wall-to-wall or cycle-based certification

Two legitimate shapes for the annual count:

Wall-to-wall. Every SKU, every location, counted in one event — usually a closure day, a weekend, or an overnight. You stop or pause operations, count everything, reconcile, and reopen with a clean baseline. It's the traditional shape, and it's the right choice when your cycle-count discipline has slipped, when the year saw a system migration or a big supplier change, or when your accountant or auditor needs a single dated full-count record.

Cycle-based certification. If your cycle counting covers every SKU over the year at known frequency, the year-end number is built by reconciling the accumulated cycle counts — spot-verifying high-value items, re-counting the variance-heavy SKUs, and certifying the book value from a year of evidence rather than one heroic night. Fewer dramatics, but it demands the cadence actually ran — a certification built on cycle counts that quietly stopped in August is a fiction with extra steps.

Which is right depends on your year, not on fashion. The honest test: would you defend your current book quantity, per SKU, with documented counts if someone asked? If yes, certify. If no, do the wall-to-wall. (The mechanics of the counting itself — scopes, mobile counting, applying adjustments — are the same either way and covered in the Shopify counts guide.)

Prep checklist

The count is won or lost in the week before it:

• Tidy the floor. Everything in a bin, nothing in aisles or staging. Unlocated stock is uncountable stock.
• Clear the receiving pipeline. Shipments in transit get a decision: either received and on the floor before count day, or marked and held back unopened. A half-unpacked pallet on count night is a variance generator.
• Finish returns processing. Units sitting in the returns pile are neither sold-out stock nor shelved stock. Restock what's restockable, quarantine the rest, before the count.
• Train the counters. One page: how to count (physical units, not system number), how to record (SKU-level, counted quantity, no rounding), what to flag (damaged, unlabelled, wrong item). Counters who "check the system first" will copy the system. The count exists to test the system.
• Print count sheets or set up the native count. Shopify's native stock count tool scopes a count by location and products, lets staff record quantities by scanner or keypad without touching live stock, and queues the differences for a manager to review and approve before inventory changes (Shopify Help Center, "Stock counts," accessed October 2026).
• Schedule the quiet. Count when nothing is selling: after close, before open, or a closed day. POS stores count with the till dark.

Cutoff discipline: orders in flight

Cutoff is the part beginners miss and accountants care about most. The count has to answer one question cleanly: which sales and receipts belong to the old year, and which to the new one?

Shopify's inventory states make this tractable if you respect them:

• Committed units — units in unfulfilled orders or reserved in draft orders — are still your inventory at count time. If a unit is picked and boxed on the packing bench with its order unfulfilled, it counts as on-hand stock (in the Committed state), not a sale.
• Orders fulfilled before the count left the building: their units are gone; don't count them even if the physical box is still waiting for pickup.
• Receipts in transit — POs and transfers marked in transit — are not on your shelf. If the shipment physically arrives during the count window, it's a cutoff decision, made once and written down: received-and-included, or held-sealed-and-excluded.

The practical rule: freeze the movement. Count when nothing is being picked, packed, received, or transferred, and apply the count before trading resumes — the same close-of-day logic cycle counts use, because a count applied over live sales overwrites units sold while counting (Shopify Help Center, counts documentation, accessed October 2026). Write the exact cutoff timestamp on the count record; it's the first thing any reviewer asks.

From count to value: what the number is for

The counted quantities exist so your inventory value can be defended. The arithmetic is simple: quantity on hand × cost per unit, summed. The discipline is not:

• Cost per unit must be real. Shopify's cost per item field is the natural source — but it's a static field that only reflects what you last entered, and it excludes freight and duty by design (Shopify Help Center, "Profit reports," accessed October 2026). If your true landed costs differ from the field, your year-end inventory value inherits the difference. Getting the cost fields honest before the count is part of prep — the inventory cost guide covers the tracking options.
• The count is not the valuation method. Counting establishes quantities; how those units are costed (and how write-downs for damaged or obsolete stock are handled in the books) is an accounting question. Record what you found — including damaged and dead stock found during the count — and let the books handle treatment.
• Damaged and unsellable stock gets recorded as found. The count that finds 40 damaged units should record both the 40 damaged and the reduction in sellable stock. Pretending damaged units are good stock inflates the year-end value by exactly their cost — e.g., 72 sellable + 8 damaged units at $16.10 each is $1,159.20 of sellable value, not the $1,288.00 you'd claim if the damaged 8 were counted as good stock.

One educational note, because this article is not tax advice: inventory value at year end feeds into the cost-of-goods figure on your financial statements, and write-downs and count adjustments have tax implications that vary by jurisdiction and accounting method. Confirm treatment with your accountant before the count, not after — they may want specific documentation from the event (dated sheets, adjustment logs, photos of damaged stock), and gathering it live is free.

Reconciling variances

Counted ≠ system. Now the work, in order:

1. Approve adjustments so the system matches the shelf. In the native count flow, a manager reviews counted vs expected and applies the corrections; each change lands in the adjustment history (as of October 2026).
2. Recount the big variances before accepting them. A 100-unit miss is usually a process event — a mis-received PO, an unprocessed transfer, a location mix-up — not shrinkage. Recount, check the SKU's recent trail, then accept.
3. Categorize every variance. Damage, theft or loss, process error (receiving, counting, transfers), timing error (cutoff slip). The native adjustment reasons map to this, and the pattern across a year of categorized variances is where next year's fixes come from — the reporting surface for this is covered in Shopify inventory reports.
4. File the record. Count sheets, adjustment history export, variance log, damaged-stock photos, cutoff note. Retained with the year's books. That file is the difference between "we counted" and "we have evidence."

Then the forward step: every variance category becomes a cycle-count priority for the new year. The count isn't the end of counting — it's the calibration.

Common questions

What is a year-end inventory count? The full physical count of every SKU taken at the close of the financial year, used to establish and defend the inventory value in your books. It's operationally the same as any full stocktake — count everything, reconcile against Shopify, adjust — with added discipline around order cutoff and documentation, because the result backs your financial statements.

Does my business need a year-end inventory count? If you carry inventory as an asset, the year-end value needs a defensible basis, and a physical count is the standard evidence. Whether it must be a single wall-to-wall event or can be a certification built from documented cycle counts depends on your accounting approach and who reviews your books — worth asking your accountant before choosing.

How do I handle orders in flight during the count? Freeze movement and use Shopify's inventory states: committed units in unfulfilled orders still count as on-hand inventory, fulfilled orders have left the building, and in-transit receipts are either received-and-included or held-sealed-and-excluded — one decision, written down with a timestamp. Count when nothing is being picked or received.

What do I do with variances found in the count? Apply the correction so the system matches the shelf, recount large misses before accepting them (most trace to a process event, not shrinkage), categorize each variance — damage, loss, process error, timing — and file the record with the year's books. Each category becomes a cycle-count priority for the new year.

How does the count affect inventory value? Year-end value is counted quantity × cost per unit, summed. The count fixes the quantity side; the cost side comes from your cost records — and Shopify's cost per item field is static and excludes freight and duty, so true landed costs may differ. Damaged stock found in the count reduces sellable value; treatment in the books is an accountant question.

Next step: two weeks before your count date, run the prep checklist as an audit of the pipeline — returns pile, in-transit shipments, staging areas, cost fields. Whatever the audit finds broken, fixing it takes a day now and ruins the count if left for count night.