How to Set Up Cycle Counting in an ERP System

Key Takeaways

• Cycle counting improves control by checking selected stock throughout the year.
• Prioritize high-value, fast-moving, expiry-sensitive, or high-risk items.
• Count by location and item identity, not only by a printed SKU list.
• Every adjustment needs an approval rule, reason code, and follow-up.

What is cycle counting?

Cycle counting is a planned process for counting a portion of stock at regular intervals. It is different from an annual physical inventory because the work is distributed across days or weeks. Oracle describes the purpose as keeping physical on-hand quantities aligned with the quantities shown in its inventory system, while Microsoft’s warehouse guidance separates creation of count work, physical counting, and resolution of differences.

For Egyptian retailers, pharmacies, distributors, and manufacturers, the benefit is operational visibility: errors can be found near the transaction that caused them instead of accumulating until a year-end count.

Build a risk-based count plan

Start with a simple classification. Count expensive items, fast movers, products with expiry or batch obligations, items with repeated variances, and stock in locations where access or handling creates risk. Include a small random sample of lower-risk items so the plan does not teach staff which products will always be checked.

The frequency should follow risk and evidence. A frequently sold item may need more attention than a slow mover; a low-value item with repeated shrinkage may deserve more attention than its price suggests. Document why an item is in a class and review that decision when movement or loss patterns change.

Design the count process

1. Freeze or clearly mark the counting scope and time.
2. Identify the warehouse, aisle, bin, SKU, unit, batch, and serial where relevant.
3. Use a mobile device or controlled count sheet that does not expose the expected quantity when blind counting is required.
4. Record the physical result and keep the original entry.
5. Recount material variances before approval.
6. Post an adjustment only through an authorized workflow with a reason.

Oracle’s current documentation supports scanning or tapping subinventories, locators, items, lots, and serials during cycle counts. That detail matters in regulated or expiry-sensitive operations: “ten units” is not always enough if the ten belong to different lots.

Connect counts to root-cause control

A variance is a signal, not merely an accounting adjustment. Review receiving, picking, returns, transfers, unit conversions, damaged stock, and unauthorized issues. Look for patterns by branch, location, shift, product family, and transaction type.

Set tolerances carefully. A tolerance may reduce unnecessary approvals, but it should not hide repeated small losses. Pair the count programme with barcode discipline, separation of duties, and a dashboard showing accuracy, variance value, overdue counts, and repeat causes.

A practical 30-day start

Week one: clean item, unit, location, batch, and serial data. Week two: choose one warehouse or branch and define the risk classes. Week three: run supervised counts and test variance approval. Week four: review causes, adjust frequencies, and publish a repeatable schedule. Scale only after the first location produces trustworthy results.