Year-End Stocktake During Peak Season: How to Count Stock Without Stopping Despatch

Yes, you can complete a year-end stocktake during peak season without closing the warehouse — but only if you stop treating it as a single event. The method that works for a Singapore SME shipping through 11.11, Black Friday on 27 November, Christmas cut-offs and the pre-Chinese New Year rush is to count in slices: freeze one bin or product group at a time, count it blind, reconcile it the same day, and release it before the next pick wave. The version that fails is the traditional wall-to-wall count on the last weekend of December, because it lands on your highest-volume despatch days and produces a closing stock number your own team does not believe by the second week of January.

Why does the traditional stocktake weekend fail in a four-peak quarter?

The shut-the-warehouse-and-count-everything approach assumes there is a quiet period to shut. In Q4 2026 there isn't one. You restock after 11.11, ship Black Friday on 27 November, work through Christmas courier cut-offs, and are simultaneously raising purchase orders against Chinese New Year factory shutdowns in February 2027. Marketplace despatch SLAs do not pause for your count, and a weekend of zero outbound during peak is real revenue foregone.

So the count gets compressed, staffed with casual labour who do not know your SKUs, and finished at 2am. Variances go unexplained, adjustments spill into January, and the closing stock figure in your FY2026 accounts is effectively an estimate. That estimate is not harmless: closing stock drives cost of sales, cost of sales drives profit, and profit drives the tax position you file. A soft number at the warehouse becomes a soft number on your tax computation.

What does cycle counting look like in a live warehouse?

Cycle counting replaces one big count with many small ones, scheduled around despatch rather than against it.

• Classify first. Rank SKUs by value and movement. A-items (roughly the top 20% by value) get counted weekly, B-items monthly, C-items once before year-end.
• Freeze the bin, not the building. Lock a specific location for 30–60 minutes, ideally before the morning pick wave starts or after the last collection.
• Count blind. The count sheet shows SKU, bin and a blank quantity field — never the expected figure. A pre-filled quantity gets confirmed, not counted.
• Reconcile the same day. Any variance above your threshold (for example 2% of quantity or S$200 of value) triggers a second count by a different person while the stock is still where it was.
• Use your own staff. People who recognise your packaging catch mis-picks that a temp cannot.

A workable December rhythm: A-items counted every Monday and Thursday morning before despatch, B-items spread across the month by zone, C-items and low-value consumables in the last week of December when order volume finally drops.

How do you stop stock moving mid-count?

Cut-off discipline is what makes a rolling count defensible. Four controls do most of the work:

1. Record the last document numbers. At each freeze, note the last goods-received note, the last delivery order and the last stock transfer. Anything after those numbers belongs to the next period, not this count.
2. Quarantine inbound. Deliveries arriving after the freeze go to a clearly labelled staging bin and are excluded until the count closes.
3. Treat picked-not-shipped as stock. Goods sitting in the outbound bay still belong to you until the delivery order is raised and the courier takes them. Count them, and be consistent about it.
4. Get third-party statements at the same date. Stock held by a 3PL, at a marketplace fulfilment centre, or on consignment with a retailer must be confirmed as at your cut-off date and reconciled to your records.

Write a one-page cut-off memo, have the warehouse supervisor sign it, and file it with the count sheets. That single page answers most of the questions an auditor or a new accountant will ask in March.

Is this the month to install a new inventory system?

No. December is the worst possible time to cut over a warehouse or accounting system, and a stocktake is not a good reason to rush one. What you actually need this quarter is modest: barcode or SKU scanning on a phone, count sheets that capture counter name, bin, date and time, an automatic variance calculation, and an adjustment journal you can export into your accounts.

Most SME accounting packages already include a stock module that does this; a disciplined shared spreadsheet with locked formulas will get you through one year-end if it must. If your conclusion from this count is that you need a proper inventory or ERP layer, the right move is to scope it in October, sign off the digitalisation roadmap while grant applications can still be prepared, and target a go-live after Chinese New Year — not on 1 January, and never during peak.

What evidence should the count leave behind?

The count has two outputs: a number, and the ability to prove the number. For a company with no finance department, the second one matters more, because it is what lets you close the books in days rather than weeks.

• Signed count sheets showing counter, bin, date and time
• The cut-off memo with last document numbers
• A variance report with written explanations for every item above threshold
• The adjustment journal, approved by someone other than the counter
• Third-party stock confirmations for 3PL, consignment and marketplace-held goods
• Your valuation basis (FIFO or weighted average) applied consistently
• Evidence supporting any write-down of obsolete or slow-moving stock to net realisable value

That last point is worth money. Stock identified as dead in October and written down on evidence is a legitimate FY2026 cost. The same stock quietly carried at full cost for another year inflates your profit, inflates your tax, and still has to be dealt with eventually.

Which numbers tell you the count actually worked?

Track five figures and you will know whether next year's count can be shorter: count accuracy by product group, total variance value as a percentage of stock value, percentage of SKUs covered before 31 December, average days from count to reconciliation, and the number of stock adjustments raised after you closed the year. The last one should be zero. If it isn't, the count didn't close — it just stopped.

What should you do before 11.11?

Between now and 9 October, classify your SKUs into A, B and C, map and label every bin, and set your variance thresholds in writing. By mid-October, pilot one A-group count and time it, so you know how long 100 SKUs actually takes your team. In late October, fix the physical problems the pilot exposes — unlabelled locations, the same SKU in three bins, stock with no home. From November, run the rolling counts. Keep the last week of December for C-items and high-value spot checks only.

Done this way, the year-end number is already substantially counted before Black Friday, and the warehouse never stops shipping to produce it.

Frequently asked questions

Is cycle counting acceptable instead of a full physical count?

Yes, provided coverage is complete over the year, counts are blind, variances are investigated and everything is documented. Auditors commonly accept cycle counting supported by good controls, and may ask to attend one count. If your company is exempt from audit, you still need a supportable closing stock figure for your accounts and tax computation — the standard of evidence is for your own protection, not just the auditor's.

Our stock sits with a 3PL or marketplace warehouse. Who counts it?

You do, in the sense that the number on your balance sheet is yours to defend. Request a stock-on-hand report as at your exact cut-off date, reconcile it line by line to your own records, and verify a sample independently — either by attending a count at their site or by testing a handful of SKUs against inbound and outbound documents. Unreconciled third-party stock is one of the most common sources of a surprise January adjustment.

What if we find a large variance we cannot explain before 31 December?

Record it, recount it, and do not adjust blind. Document what you found, what you checked and what remains open, then post the adjustment in the correct period with that explanation attached. Separately, chase the cause — unrecorded customer returns, mis-picks at the same bin, and goods received but never entered account for most large unexplained gaps, and all three will recur next quarter if nobody fixes the process behind them.