Define the decision boundary and reconcile crate flows

Advisory work starts with the decision, rather than the slogan ‘increase reuse’. A shortage of transport crates could mean insufficient purchases, damage on return, slow customer returns, or confirmed losses. Each explanation calls for a different intervention. Define the accounting unit as one crate over one month. Draw the boundary around serviceable stock in the warehouse, crates held by customers, returned crates quarantined for inspection, and confirmed losses. A crate still held by a customer is not waste. A damaged crate is not a confirmed loss before repairability is assessed. Record the owner, date and source of each figure, and separate physical counts from estimates.

The conservation equation is opening serviceable stock plus purchases equals ending serviceable stock plus quarantined damaged crates plus crates still with customers plus confirmed losses, provided there are no other transfers. Dispatch and return move assets between locations inside the boundary; they do not create a new unit of consumption on each move. Calculate physical serviceable stock from opening stock, purchases, dispatches and serviceable returns. Then independently reconcile every crate to an ending state. A mismatch can arise when dispatch records count orders while the warehouse counts individual crates. Do not force that mismatch into a fictitious waste category. Keep an unexplained-difference line and request transfer records or a recount.

Use a three-column evidence-gap register: what is known, what must be checked, and which finding could change the decision. The 40 confirmed losses need a reconciliation record. The 80 crates held by customers need dispatch dates and expected return dates. Repairability requires inspection results, repair cost, and a later use cycle. Do not add the damage and loss percentages into one ‘waste rate’: one uses returns and the other dispatches. A combined rate would require a shared denominator, a consistent event definition and a check for double counting. In the advisory memo, state the case is hypothetical and that its numbers are neither local statistics nor an industry benchmark.

Counterfactual decision test: if an inventory investigation establishes that the 60 customer-held crates were also lost, confirmed losses rise to 80 of 800 dispatched, or 10%. A return-tracking intervention then takes priority over studying repair of the 50 damaged crates, while safety inspection still matters. If the 60 return on time, damage or changing demand becomes a stronger explanation, and delay must not be relabelled loss. A change in investigation priority does not by itself authorize buying 100 more crates: forecast demand, procurement lead time and alternative costs remain unknown.

Dividing inventory into operational states prevents another false decision: total assets can remain constant while the ability to serve customers falls because too many crates are outside the warehouse. Fix a cut-off time for the baseline, such as the end of the final day of the month in the same ledger time zone. A crate returned the following morning must not be counted in the closed month and again in the new month. Record a transfer in transit separately with its document number. Test completeness from two sides. Total units entering the system must equal all ending states; all dispatches must equal returns, confirmed losses and outstanding crates. The first equation can balance while the second fails if a return is duplicated or not matched to a dispatch. Request evidence for the difference before calculating performance rates or attributing fault to a particular operator.

Audit the count with a repeatable sample, not only matching totals. Trace ten dispatched crate identifiers forward to customer acknowledgement or return, then trace ten physically present crates backward to their entry records. The first path finds open movements; the second finds assets absent from the ledger. Record unmatched cases and error type, without extrapolating a small easy-site sample to the whole network. A duplicated identifier can create a false arithmetic balance. Return to the commercial question: does lower serviceable stock actually prevent peak-day demand? A purchase case needs demand and lead-time evidence, not a stock decline alone.