Stakeholders and evidence ownership

A resource balance is not complete merely because warehouse figures add up. Material may pass to a packaging supplier, carrier, customer or end-of-life processor, and those parties have different incentives when recording losses. Map actors around a specific decision: who owns the asset, controls its use, bears the loss cost, measures the outcome, and can change the design? The signatory is not necessarily the data owner. Procurement may approve replacement crates, but the return rate cannot be tested without carrier and customer handover records. An external adviser does not gain authority to certify a fact merely by compiling numbers supplied by others.

Start with an actor–evidence–incentive matrix. Include the client operator, warehouse, carrier, asset supplier and receiving processor. For each, name the information generated, an incentive to overstate or understate it, and an independent check. A warehouse may count crates at dispatch while a customer acknowledges receipt the next morning; the time difference is not automatically loss. A supplier may quote design life from an internal test, useful as a hypothesis but not a substitute for operating life in the client network. Use common event definitions: dispatched, received, returned, inspected, declared serviceable and transferred for recycling. Each event needs a date, asset identifier and unit of measure.

Record evidence strength, rather than displaying false numeric precision. A dated physical count supports stock at a specified site. A carrier document without receiver confirmation is weaker evidence of receipt. An approximate email is not enough to settle a financial dispute. These ranks do not mean one source is always truthful: check duplicate identifiers, cut-off rules, signatory authority and whether a ‘crate’ was confused with a ‘shipment’. When evidence conflicts, document competing explanations and how each changes the decision, then design a short check capable of distinguishing them. Urgent containment of a safety or leakage issue can proceed, but a causal claim or saving estimate should await support.

Actors also hold different decision rights. The client sets an acceptable service level; a safety lead sets conditions for reusing a food-contact container; procurement determines contract procedures; the carrier explains feasible scans and their cost; and the supplier specifies warranty and repair limits. These are not interchangeable votes. A safety gate cannot be averaged away by material savings. If management needs a prompt recommendation, propose a reversible, time-limited joint data collection step with explicit stop conditions instead of a long contract based on conflicting loss definitions. Give each step an owner, due date and closure evidence, so ‘align on data’ becomes an executable recommendation rather than a slogan.

The advisory output is an evidence protocol, not merely a list of names. A concise memo can say: ‘Current dispatch and receipt records use incompatible cut-offs. Adopt a shared crate identifier, cut-off time and receiving acknowledgement. Suspend loss-linked penalties until two consecutive cohorts reconcile. The warehouse leads the count, the carrier supplies scans, the customer confirms receipt, and the adviser examines differences without altering original records.’ This separates data creation from verification. If identifiers cannot be shared for privacy or commercial reasons, propose a restricted-purpose token or aggregated report rather than presuming access to all customer data. The remaining choice of verification design belongs to the parties who control the records and contract.

Make evidence-sharing part of the decision design. If a carrier cannot release raw customer records, request pseudonymous asset identifiers, timestamps and signed handovers, with an agreed way for a second party to recheck a sample against the owner’s originals. Define who can correct an entry and retain the original and correction trail. Compare sites using dispatched cohorts and waiting windows, not unmatched counts alone. A rapid penalty for a cut-off discrepancy can discourage honest reporting; an auditable protocol should reduce incentives to hide errors.