Retiring a corporate gift SKU is not a popularity contest. It is a controlled operating decision that must connect customer demand, landed economics, product quality, supply reliability, market restrictions, and replacement readiness. This guide provides a versioned scorecard and a repeatable review process so merchandising, supply chain, finance, and compliance teams can reach a traceable recommendation without pretending that a formula can replace accountable judgment.
Download the localized tools: English XLSX scorecard v1.0 · English UTF-8 CSV score export v1.0
What the scorecard decides, and what it does not
The scorecard recommends one of four actions: Keep, Improve, Localize, or Retire. Keep means the available evidence supports continued listing under the present controls. Improve means the SKU can remain, but an owner must correct a documented weakness such as defect rate, packaging, margin, or lead time. Localize means the global item should not be treated as universally deployable; a market-specific substitute, certification, label, or fulfillment path is required. Retire means demand, economics, quality, supply, restrictions, and replacement evidence together support an orderly exit.
The recommendation is not an automatic inventory write-off, supplier termination, legal conclusion, or safety clearance. It does not replace a recall process, accounting policy, contractual review, or market-access decision. A named decision owner must approve the action and record the reason. If a safety restriction or mandatory stop-sale condition exists, that control overrides the weighted score. Conversely, a high score does not authorize sale where required evidence is missing.
Use the tool at a fixed review cadence, normally quarterly for active catalogs and immediately after a serious defect, supplier disruption, regulatory change, or failed campaign. Compare a SKU against its own declared thresholds and approved alternatives, not against a fabricated industry average. Every threshold in the workbook is editable because catalog strategy, cost structure, quality tolerance, and market rules differ by organization.
The workbook separates evidence from governance. The Inputs sheet records the observed facts. The Scorecard applies the chosen weights and formulas. The Decision Log preserves the recommendation, owner override, rationale, actions, due date, and evidence link. The QA sheet tests boundaries, missing data, formula behavior, and weight totals. The Methodology sheet explains version, limits, and refresh triggers. That separation lets another reviewer reproduce the result without relying on a meeting memory.
The seven evidence fields and the weighting method
Version 1.0 uses seven input fields. Trailing demand asks whether qualified demand is continuing, not whether one exceptional order occurred. Landed margin measures the remaining contribution after acquisition, freight, duties, packaging, handling, and expected service costs under the organization's approved calculation. Return or defect rate describes observed product or fulfillment failure. Lead time captures replenishment responsiveness. Supplier capacity measures confirmed ability to meet the relevant volume and window. Market restrictions capture known certification, labeling, shipping, material, or sale constraints. Replacement readiness evaluates whether an approved substitute can carry the use case without creating a new risk.
Each field receives a normalized score from 0 to 100 and an editable weight. The default weights total 100, but they are starter assumptions, not universal benchmarks. Teams should document why a weight changes and apply the change consistently to the review population. The weighted score is the sum of each normalized score multiplied by its weight, divided by 100. Missing required inputs return a missing-data result instead of silently becoming zero.
| Evidence field | Default weight | Question answered | Required proof |
|---|---|---|---|
| Trailing demand | 20 | Is qualified demand durable enough to keep the item available? | Order history, pipeline rules, campaign context |
| Landed margin | 20 | Does the item support the approved economic objective? | Current cost build and approved margin method |
| Return or defect rate | 20 | Is observed quality within the declared tolerance? | Returns, complaints, inspection and corrective-action records |
| Lead time | 10 | Can replenishment meet the promised service window? | Confirmed production and transport lead time |
| Supplier capacity | 10 | Can supply cover the relevant demand scenario? | Supplier confirmation and capacity assumptions |
| Market restrictions | 10 | Can the item lawfully and operationally serve each target market? | Market-specific product, label, material and shipping evidence |
| Replacement readiness | 10 | Can demand migrate without an unacceptable gap? | Approved substitute, sample, cost, capacity and launch plan |
The default interpretation bands are also assumptions. A score of 75 or more recommends Keep, 60 to 74.9 recommends Improve, 40 to 59.9 recommends Localize, and below 40 recommends Retire. A market restriction marked as unresolved forces review even if the arithmetic is high. A missing required field forces Missing data. The workbook exposes these rules so an approver can inspect them; it does not hide a policy choice inside an unexplained formula.
Run the review as an auditable operating cycle
Start by defining the decision population. Freeze a list of SKU identifiers, markets, review date, data owner, and evidence cutoff. Do not mix an individual colorway, a parent product family, and a bundled kit unless the organization has explicitly chosen that review level. A retirement decision at the wrong level can remove a healthy variant or preserve a failing component.
Second, collect evidence through named owners. Merchandising owns demand context and customer use cases. Finance owns the landed-margin method and confirms the cost period. Quality owns defect classification and distinguishes product faults from carrier damage or customer preference returns. Supply chain owns lead time and capacity. Compliance or legal operations owns the market restriction evidence without delegating legal judgment to the spreadsheet. Product or catalog operations owns replacement readiness and the transition plan.
Third, normalize each input using thresholds approved before results are reviewed. If the team changes a threshold only after seeing an inconvenient recommendation, record the change as a methodology revision and rerun the whole population. Never tune one row to reach a preferred answer. Preserve the raw input, normalized value, source date, and source link.
Fourth, inspect exceptions before discussing the recommendation. Ask whether any mandatory stop-sale, recall, expired certification, restricted destination, unresolved safety complaint, or supplier authenticity concern applies. In the United States, the Consumer Product Safety Commission's duty-to-report guidance explains that firms have reporting responsibilities for potentially hazardous products. In the European Union, the European Commission product-safety portal provides the official starting point for product-safety rules and alerts. The scorecard records these checks; it does not decide them.
Fifth, hold a decision meeting only after the evidence packet is complete. The owner selects the action, may override the formula, and must state why. An override is valid governance when it is explicit, bounded, time-limited, and supported. An undocumented override is simply an untraceable decision.
Finally, assign execution tasks and acceptance evidence. A Keep decision still needs a next review date. Improve needs a corrective action, owner, due date, and verification. Localize needs market scope, substitute or control, and launch evidence. Retire needs last-order rules, inventory disposition, catalog removal, customer migration, supplier communication, and closure evidence.
Hypothetical decision A: keep a welcome-kit bottle
This case is fictional and demonstrates the method; it is not a Giftpack customer result. Assume a welcome-kit bottle has qualified demand and landed margin at the top of the declared bands, an acceptable but monitored quality record, a strong replenishment result, confirmed supplier capacity, and no unresolved market restriction in the reviewed destinations. A substitute exists but is only partly ready. The normalized inputs are 100 for demand, 100 for margin, 60 for quality, 91.7 for lead time, 100 for capacity, 100 for market restrictions, and 40 for replacement readiness.
Using the default weights, the weighted result is about 85.2, which recommends Keep. The team should not stop at the label. Merchandising records the campaign and reorder assumptions behind demand. Finance attaches the current cost build. Quality links inspection and complaint summaries. Supply chain records the supplier confirmation date. Compliance identifies the reviewed markets and evidence cutoff. The decision owner approves Keep and schedules the next review.
The reasonable alternative is Improve if the lead-time risk is material during a seasonal campaign. The owner might keep the item while requiring a secondary production window, higher safety stock, or an earlier order cutoff. The alternative should not be hidden; record why it was rejected or converted into a control.
Execution tasks include preserving the current listing, refreshing the cost build before the next buying cycle, monitoring defect rate monthly, and retesting capacity before a large campaign. Acceptance evidence is a signed decision-log row, linked source packet, weight-total check, formula-error scan, next review date, and evidence that the stated control was actually configured. If a new defect cluster appears, reopen the decision rather than waiting for the scheduled review.
This example also illustrates why replacement readiness is not a reason to retire a healthy item. A prepared substitute reduces continuity risk, but it should not erase the demand and quality evidence supporting the current SKU. The field measures resilience, not a preference for turnover.
Hypothetical decision B: localize a desk cable set with market constraints
This second case is also fictional. Assume a desk cable set has weak demand, acceptable landed margin, an unacceptable observed quality result that is not being treated as a safety clearance, a long replenishment lead time, partly constrained supplier capacity, unresolved fit or labeling requirements in one destination, and a localized substitute that has passed sample review. The normalized inputs are 40 for demand, 70 for margin, 0 for quality, 33.3 for lead time, 64.7 for capacity, 50 for market restrictions, and 85 for replacement readiness.
The weighted result is about 45.3, which recommends Localize. The recommendation does not mean the global SKU is defective, nor does it authorize sale in the constrained market. It means the evidence supports separating the decision by market. The owner can retain the original kit where current evidence is sufficient, block it where requirements remain unresolved, and complete validation of the local substitute.
The main alternative is Retire if the substitute cannot reach approved capacity, the low-margin condition persists, or the restricted market represents most qualified demand. Improve may be appropriate if a supplier can correct the documentation and lead-time issues within an approved window. Keep is not credible while a material restriction is unresolved. The decision log should identify the selected path and the conditions that would trigger the alternatives.
Execution begins with a destination block in catalog and order controls, a verified list of affected configurations, and an owner for local validation. Procurement obtains final price and capacity. Quality investigates the observed failure, approves the replacement sample, and defines an inspection plan. Compliance confirms the destination evidence. Catalog operations maps customers and campaigns to the approved substitute. Finance recalculates margin with the local landed-cost build.
Acceptance evidence includes the market decision, approved substitute identifier, sample approval, capacity confirmation, final label or documentation proof, catalog rule test, and a successful order-path test in the permitted destination. If the substitute fails, recovery is to preserve the block, reopen sourcing, and communicate the service gap; it is not to reactivate the original item without evidence.
Decision rights, owner overrides, and evidence quality
The scorecard works only when responsibility is explicit. A data contributor can correct a source value but should not silently change the decision. A methodology owner can revise normalization rules but should version the change and rerun affected rows. The decision owner can override the recommendation but must identify the risk accepted, the compensating control, the expiry date, and the evidence required to close the exception.
| Role | Owns | Must not do alone | Acceptance evidence |
|---|---|---|---|
| Merchandising | Demand context, customer fit, assortment rationale | Approve safety or accounting conclusions | Qualified-demand packet and use-case notes |
| Finance | Landed-cost and margin method | Decide market eligibility | Dated cost build and method approval |
| Quality | Defect taxonomy, inspection and corrective action | Treat every return as a product defect | Issue summary and verification result |
| Supply chain | Lead time, capacity and transition feasibility | Promise unconfirmed supplier capacity | Supplier confirmation and scenario |
| Compliance or legal operations | Market evidence and escalation route | Use the score as legal advice | Official source, review scope and date |
| Decision owner | Final action, override and risk acceptance | Erase contrary evidence | Signed decision-log entry and due dates |
Evidence quality has four dimensions. Provenance asks where a value came from. Timeliness asks whether it covers the relevant period. Scope asks whether it applies to the exact SKU, configuration, supplier, and market. Reproducibility asks whether another reviewer can reach the same calculation. A precise number with weak scope is not strong evidence. A supplier email can confirm capacity, but it cannot establish a regulator's rule. A dashboard can show returns, but it may need classification before it represents defects.
Use the principles of the International Organization for Standardization's quality management guidance as a general orientation toward evidence, process, improvement, and customer focus, not as a claim that this workbook certifies conformity. For inventory accounting, IFRS Foundation IAS 2 is an official reference point for entities applying that standard, but finance must determine the organization's actual policy and jurisdictional requirements.
Failure modes and recovery paths
Open the operational failure and recovery checklist
-
Weights do not total 100. Stop the review, restore an approved weight set, document the change, and rerun every affected row.
-
A required input is missing. Return Missing data, assign an owner and due date, and do not convert the blank to zero.
-
A formula displays an error. Preserve the file, identify the broken reference, repair only the affected formula, recalculate, rerun boundary tests, and create a new version if distributed bytes change.
-
Two systems disagree. Preserve both values, identify their period and scope, select the governing source through the named owner, and record the reconciliation.
-
A safety or market restriction emerges. Apply the operational hold first, follow the official escalation route, and treat the score as secondary.
-
A substitute fails sample or capacity review. Keep the destination block, reopen sourcing, and update replacement readiness; do not force a launch to protect a target date.
-
The owner overrides the recommendation. Record reason, risk, control, expiry, and required proof. Expired overrides return to review.
-
A published asset checksum changes. Reject the unverified copy, retrieve the immutable version by URL and hash, and issue a new version rather than overwriting the receipt.
Common process failures are subtler. Teams may count requested quantities as demand even when orders were canceled, compare margins calculated by different methods, treat late carrier delivery as supplier lead time, or combine markets with different requirements. They may also retire an item before a replacement is operational, creating a service gap that the score did not intend.
Recovery begins with containment: pause the specific decision or market, not the entire catalog. Then identify the smallest affected scope, preserve the old evidence, correct the source or rule, and rerun only the defensible population. Require a second reviewer for methodology changes. The decision log should show what changed, why, who approved it, and which prior conclusions were reconsidered.
Version control matters because the workbook is a decision artifact. Version 1.0 is dated 2026-10-11. If formulas, thresholds, field definitions, or outcome bands change, increment the version and retain the old checksum. Routine row updates can remain within the version if the method is unchanged and the decision log records the review date.
Product-safety, market, and accounting guardrails
A retirement scorecard can organize evidence, but it cannot collapse different governance domains into one number. Product safety, legal market access, privacy, tax, customs, employment, and accounting decisions each have their own owners and official sources. The practical rule is simple: use the score for portfolio prioritization, then route domain conclusions to the accountable function.
For product-safety concerns, preserve incident details, affected identifiers, dates, destinations, and supplier information. Escalate through the organization's reporting procedure and consult the relevant authority. Do not wait for the next quarterly scorecard. For market access, verify the exact product configuration and destination because plugs, batteries, wireless functions, food contact, children's use, materials, labeling, and language can change the applicable requirements.
For inventory economics, distinguish a commercial retirement decision from valuation, impairment, write-down, disposal, or tax treatment. The score may trigger finance review, but finance owns the accounting conclusion. Record quantities, cost basis, committed orders, return rights, salvage options, donation restrictions, and disposal controls before execution.
For suppliers, separate the SKU decision from the relationship decision. A single product may retire while the supplier remains valuable, or a product may look healthy while supplier authenticity or capacity evidence requires escalation. Contract notice, tooling ownership, intellectual property, remaining materials, and quality obligations should be reviewed before termination or transition.
These guardrails prevent two opposite errors: using a spreadsheet as a false legal clearance, and refusing to make an operating decision until every uncertainty disappears. The scorecard supports a bounded decision with explicit holds, owners, and next evidence. It makes uncertainty visible instead of disguising it.
Implementation plan and acceptance evidence
Week one establishes governance. Name the decision owner, methodology owner, evidence contributors, and escalation routes. Define the review unit, markets, cutoff date, approved margin method, defect taxonomy, and threshold set. Select a small pilot population that includes a healthy SKU, a weak SKU, a restricted-market case, and a missing-data case.
Week two validates the method. Enter source evidence, run formulas, inspect recommendation bands, and test boundaries immediately below and above each threshold. Confirm that missing required data returns Missing data. Change one input at a time to verify behavior. Confirm weights total 100 and that owner overrides do not erase the formula result. Review all six workbook sheets and the CSV export.
Week three runs the decision meeting. Present evidence before the label. Challenge scope, dates, and source ownership. Record alternatives, not just the chosen action. Assign execution tasks with owner, due date, and proof. For Localize or Retire, test catalog blocks, redirects, substitute mappings, and customer communication before removing availability.
Week four closes the loop. Collect the actual acceptance evidence, reconcile open exceptions, publish the approved decision log internally, and set the next review date. Measure process quality with completion rate, missing-data aging, override aging, corrective-action closure, and post-decision service incidents. Do not convert those operating metrics into claims that the score caused business outcomes without a suitable evaluation design.
The minimum acceptance packet contains the frozen SKU and market scope, source links and dates for all seven fields, approved weights and thresholds, formula-error scan, boundary-test results, recommendation, alternatives considered, owner decision, override details if any, execution tasks, closure proof, and next review date. The downloadable workbook's QA sheet provides the test list; the Decision Log provides the durable record.
Refresh the method annually and whenever a material defect is found in a formula, definition, threshold, official-source assumption, or workflow. Refresh individual decisions at the declared cadence and on meaningful events. A methodology refresh and a row review are different activities; record both.
Conclusion: make retirement reversible, explainable, and controlled
A strong SKU retirement process is neither a mechanical ranking nor an endless committee debate. It is an evidence packet, a transparent calculation, a named decision, a controlled execution plan, and a retained record. The seven-field model creates a common language across merchandising, finance, quality, supply chain, compliance, and catalog operations while preserving each function's decision rights.
Begin with a small population, keep thresholds editable and versioned, refuse to score missing data as zero, and treat safety or market holds as overrides to arithmetic. Require two hypothetical or historical walkthroughs before live use so owners can see how alternatives and recovery paths are documented. Most importantly, test the execution path: a correct retirement decision can still fail if catalog availability, substitute mapping, customer communication, or supplier closure is incomplete.
For organizations coordinating gift availability across markets, Giftpack can serve as an execution layer for approved catalog and fulfillment choices after the responsible teams have made the commercial, quality, accounting, and compliance decisions. It does not replace those decisions; it helps operationalize the resulting assortment and delivery rules.
Method version: 1.0 · Version date: 2026-10-11 · Source cutoff and last verified: 2026-10-11. Official sources were checked for current public availability on that date. All examples are hypothetical, and every threshold is a user-editable assumption rather than an industry benchmark.