The 4,200-SKU Warehouse That Cannot Trust Its Own Stock Report

Sequence matters: contain, verify, then correct.

1. Freeze the SKU records so no dispatch or receipt posts against them until the count is verified.
2. Blind re-count the brake pads by a second person, then reconcile against the duplicate goods receipt. If the duplicated invoice accounts for exactly 72 units, reverse the duplicate receipt; if not, investigate further before adjusting.
3. Quarantine the expired sealant. Red-tag it, move it off the pick face, block it in the system, and check whether any of batch B-2214 has already been issued. Record a write-off of Rs 18,400 with the reason code and manager approval.
4. Segregate the water-damaged filters, photograph them, raise a claim with the insurer or transporter, inspect stock stored behind them, and repair the roller shutter before the next rain.
5. Fix the stock-out commercially: transfer from another location or make an emergency purchase, and tell affected customers.

Recording the adjustment: use a stock adjustment document showing SKU, batch, system quantity, counted quantity, variance, root cause and approver signature. Never edit the quantity field directly, because the audit trail is what protects staff when the owner asks who changed 1,912 to 1,840.

Root causes, not symptoms:

• Duplicate goods receipt: no three-way match of purchase order, invoice and GRN, and one person both receives and posts.
• Annual-only counting: errors survive up to twelve months and the system keeps feeding wrong pick quantities.
• No expiry control: no FIFO enforcement or batch expiry report, so September stock sits on the pick face.
• Building maintenance gap: the shutter leak was known but not fixed, so the same loss will recur.
• Unrecorded issues: the stock-out against a positive book balance suggests picks or returns not being posted.

Controls: three-way match before posting; separate the receiving and posting roles; barcode-scan GRNs and picks; add an expiry rule that blocks issue of stock within 30 days of expiry; add a monthly maintenance checklist for shutters, drains and roof.

Cycle-count plan within capacity (60 SKUs/day):

• A-class: 340 SKUs counted fortnightly = 34 per working day, about 57% of capacity.
• B-class: quarterly, planned in the remaining capacity.
• C-class: annual, spread over the year, plus a full count at year-end.
• Count high-value and high-movement SKUs first each morning, before dispatch starts, and recount any variance above 2% the same day.

Choose a small set that shows both accuracy and financial loss:

1. Inventory record accuracy = correct SKUs / SKUs counted. Target at least 98% for A-class and 95% overall; the brake-pad case is a direct failure of this measure.
2. Shrinkage value = unexplained adjustment value as a percentage of stock value. Target under 0.5% per month, with every adjustment above a set threshold investigated.
3. Expiry write-off value. Target zero avoidable write-offs; report near-expiry stock (within 60 days) separately so it can be sold or returned.
4. Fill rate / stock-out count. Target 97% plus; the brake-pad stock-out while the book showed stock is the evidence the owner needs.
5. Cycle-count completion against plan, because an unexecuted plan is the usual reason accuracy decays.

Report these as a one-page dashboard with trend, not a single number, and link each variance to an owner and an action date. When throughput doubles at festive season, accuracy should improve, not fall, because scanning and counting discipline is tested hardest during peak.