Cycle counting without losing lot and status information

The inventory report says 100 packs. The counter finds 100 packs. The count is marked accurate, even though ten packs belong to a different lot from the one recorded. When an order needs a specific remaining shelf life or a recall drill starts, the apparently accurate stock position breaks down.

Cycle counting should test the record at the level needed for real decisions. An item total can be useful for financial reconciliation, but pharmaceutical operations also depend on lot identity, location, quantity, unit and the applicable status. The count must preserve enough detail to expose a wrong composition within a correct total.

WHO distribution guidance calls for stock reconciliation and investigation of discrepancies. The process below translates that objective into a proposed counting workflow. WHO TRS 1025 Annex 7, section 13.

Choose a counting boundary people can understand

Define whether the count covers a location, a material family, specific handling units or another controlled population. Then identify which records and physical areas belong inside that boundary. Open picks, receiving staging, samples and transfers can otherwise fall between teams.

Choose frequency and selection criteria using the site's approved approach and actual discrepancy history. High movement, recurring errors, difficult units and complex handling may justify closer attention. A value-only ranking can miss low-cost stock whose identity matters to an operating decision.

Before counting, agree what the operator must observe independently and what information the application displays. A blind quantity count can reduce copying of expected balances, but the counter still needs a reliable way to identify the physical object. Do not hide essential identification information merely to make the exercise appear more independent.

Reconcile the correct total with the wrong composition

In an illustrative location, the system records 70 packs of Lot A and 30 packs of Lot B. The physical count finds 60 packs of A and 40 packs of B. The total is 100 in both cases.

The zero net difference does not establish accuracy. Investigators need to determine whether the cause is a wrong-lot pick, a misidentified receipt, a location transfer error or another event. Each explanation can affect different historical records and downstream decisions.

Do not automatically post a negative adjustment to A and a positive adjustment to B simply because those entries make the screen balance. First establish the observed identities and investigate the relevant transactions under the approved procedure. The correction should explain what is known and preserve uncertainty that remains unresolved.

If the lots have different quality states, the operational effect can extend beyond stock accuracy. The quality owner should assess the relevant scope rather than allowing a counting clerk to change disposition as part of a quantity adjustment.

Establish a reliable movement cutoff

A count compares physical stock at a point in time with a corresponding record. If movements continue, the team needs either a controlled pause or a defined way to reconcile those movements to the count.

For a paused location, document when the boundary became effective and how exceptions are managed. Material can still arrive unexpectedly or a staged pick can be collected unless the physical operation follows the same boundary as the software.

For a live count, preserve count time, movement time and posting time separately where needed. A receipt physically completed before the count but posted afterwards can create an apparent shortage or excess. Reconcile the event sequence before concluding that stock was lost.

Choose the method that fits the operation and can be demonstrated reliably. An elaborate live-count algorithm may be less useful than a short, well-controlled pause in a low-volume area. In a busy distribution centre, a broad freeze may be impractical, so the transaction model needs more careful design.

Record observations before explanations

Capture what the counter found: object identity, location, observed quantity and unit, container condition relevant to the count, and the count time. Record an unreadable label or inaccessible container as an exception rather than filling in the expected value.

Keep the initial count when a recount occurs. The second count can resolve a counting error, but overwriting the first observation removes evidence about why the discrepancy appeared. Define who performs the recount and how independence is handled for the site's process.

Separate observation, investigation and adjustment approval in the record. These can be different stages even where a small team performs them close together. The distinction helps prevent an explanation written after the adjustment from appearing to have supported the original decision.

Follow the transaction before changing the balance

Start an investigation with the last reliable count or reconciliation point. Inspect subsequent receipts, picks, transfers, sampling withdrawals, returns and corrections. Check units and pack-size conversions as well as quantities.

A supplier carton can contain several saleable packs; an opened carton may contain fewer than its nominal pack count. If the application assumes every scanned carton is full, recurring differences are predictable. The investigation should address that process rather than treating each count as an isolated operator mistake.

Sampling can also explain a physical reduction without a customer shipment. The sample inventory reconciliation guide develops that movement. For duplicate or delayed receiving events, use the offline barcode guide.

Document both the correction and any wider impact assessment. Finding a wrong-lot movement may require checking affected orders or records. Correcting the current balance alone may leave the historical error unresolved.

Keep approval proportional to the decision

Define who may approve quantity adjustments, identity corrections and quality-related changes. A financial-value threshold may help route an adjustment, but it does not establish that a low-value identity change is operationally unimportant.

The approval view should show the original count, recount where applicable, relevant transactions, proposed correction and investigation conclusion. Provide a separate unresolved state when evidence is insufficient. Do not force the reviewer to choose a favourable explanation merely to close the counting period.

Connect the approved correction to its accounting effect without making accounting the sole definition of completion. The batch-costing guide addresses the financial view. Warehouse and quality records still need to remain coherent at their own decision scope.

Pilot counts that challenge the model

Include a correct item total with a wrong lot split, a pending movement across the cutoff, a partial carton, a sample withdrawal and a stock-status mismatch. Verify that the count captures each difference and that the proper owner can resolve it.

Measure accuracy at more than one level. Item-total agreement, lot-location agreement and unresolved discrepancy age reveal different problems. Publish the denominator and exclusions for each measure so a change in counting scope does not look like improved performance.

Review recurring causes by workflow. If discrepancies repeatedly originate at a transfer handover, improving the transfer record may be more valuable than increasing count frequency. A cycle-count programme should help repair the process that creates the balance, not become a permanent substitute for it.

NextEdge Labs can help connect count observations, movement records and approval tasks into a traceable workflow. A useful initial scope is one area where the team can compare the digital result with a manually reconstructed transaction history.

Summary and next actions

A matching item total does not prove accurate pharmaceutical inventory. Count and reconcile the lot, location, unit and status dimensions that the operation depends on, using a clear movement boundary.

• Select a counting scope and agree how movements across its cutoff will be handled.
• Capture independent observations before recording investigation conclusions or adjustments.
• Test a correct total with an incorrect lot split and follow its historical transactions.
• Use recurring causes to improve the originating workflow, with named owners and follow-up checks.