Cycle Counts: Inventory Accuracy as a Cadence

The annual stocktake is the wrong shape for an ITAD warehouse. Stock moves daily — every receiving brings new assets, every outbound takes some away, every testing failure can quietly shift status without moving a label. Once a year is too late to catch the drift. Cycle counts answer that problem with small, frequent, scoped counts that keep the database in sync with the rack.

The lifecycle

A count starts as draft: a scope (warehouse, zone or rack), an assignment and a set of positions. The new-count dialog previews the position codes in scope and pages through all of them, so a 400-position zone is not represented by its first twelve rows. Starting the count moves it to in progress; the operator scans or counts, increments quantities, marks positions as counted and reports anything that does not match. Finishing moves it to completed, and a separate approved step accepts the result after review. A count that should not have happened goes to cancelled. Completed does not mean the discrepancies were harmless — it means the walking is done. Approval is where someone says so.

Discrepancies and how they get resolved

A discrepancy is one of four things: an asset that is missing from where the system expects it, an asset found where nothing was expected, a location mismatch, or a quantity mismatch. Each is resolved with one of five actions: update the location, mark the asset missing, register the asset, correct the count, or ignore it with a reason. The original observation and the resolution are both kept, so an auditor can see what the operator saw and what the team decided. Nobody is asked to write an essay about a missing cable. They are asked for a reason.

The accuracy number

Each count shows its inventory accuracy: the positions in scope minus the discrepancies that are still open or were resolved with anything other than “ignore”, divided by the positions in scope. It is a per-count figure, computed from that count’s own positions — which is exactly the number that predicts whether your next outbound order will surprise you at the dock.

The policy lives in Settings

Under /settings/core/cycle-counts your team sets the default scope (warehouse, zone or rack), the default SLA in hours, the auto-approve threshold as an accuracy percentage, the reminder frequency in days, whether a photo is required on a discrepancy, and whether a discrepancy may be ignored without a reason. Good policy is strict enough to catch mistakes and humane enough that operators can finish the work.

Finding the count you need

The cycle-count list searches and filters across every count in the tenant before it shows a page, and the tiles at the top separate sessions, counts in progress, counts awaiting approval and unresolved discrepancies. One reconciled count per zone per quarter is easier to sustain than a perfect count once a year, and the list is built so the quarter’s counts are findable when the auditor asks.