A corporate gift can have three defensible unit costs at the same time. Dividing assigned program cost by eligible people answers a different question from dividing it by accepted claimants or confirmed recipients. This guide and its editable workbook help finance and program teams keep those denominators separate, reconcile refunds and cross-currency costs, and explain what remains provisional. Every numerical example below is hypothetical, not a customer result or an industry benchmark.
Download the workbook and establish its boundary
Download the shared-English unit-economics workbook, the English field dictionary, and the English quick-start PDF. The workbook has four worksheets: Cohorts, Input, Definitions and QA. The dictionaries explain its twenty-six input rows; they are not transaction extracts or replacement calculation engines. The PDF explains the same method and limitations. Version 2027-v1 was planned on October 17, 2026, and produced and verified on October 20, 2026. The year identifies the planning edition, not a prediction of future prices or tax rules.
Start with a copy for one decision: whether to repeat a completed campaign, investigate expensive delivery exceptions, or compare two genuinely comparable operating groups. The file supports two cohorts. Adding another column requires extending and retesting the formulas and references; merely typing a third group beside the results does not extend the model.
The populated figures are fictional estimates. Replace them with your own roster, claim history, delivery evidence, invoices, accepted credits, allocations and dated conversion rates. The workbook does not connect to an account, determine tax recoverability, approve spending or calculate investment returns. A well-reconciled cost ratio still cannot show that a gift caused a sale, improved retention or generated incremental profit.
Use the existing participation forecast calculator when estimating future funnel volumes. Use the campaign budget calculator before committing funds. This workbook addresses observed or explicitly provisional unit costs after defining the population and observation window. Keep forecast assumptions and observed outcomes in separate versions rather than quietly replacing one with the other.
Choose the denominator before interpreting the result
An eligible person belongs to the population allowed by the frozen program policy. An accepted claimant is a distinct eligible person whose claim was accepted by the cutoff. A delivered person is a distinct claimant with delivery accepted under the documented evidence rule. These definitions use people, not invitations, clicks, orders or parcels. A kit with three items is not three recipients; a replacement shipment to the same person does not create a second delivered person.
The template expects the delivered population to sit within accepted claimants and claimants within eligible people. If your program deliberately allows proxy recipients, multiple beneficiaries or deliveries without claims, first redesign and document the population model. Do not force incompatible events into these columns simply to obtain a number.
Three denominator choices and their proper reading:
| Measure | Calculation | Decision it supports | What it does not establish |
|---|---|---|---|
| Cost per eligible person | Net assigned cost divided by eligible people | Resource intensity across the whole offered population | Successful receipt or participation |
| Cost per claiming person | The same cost divided by accepted claimants | Cost allocated across accepted demand | Delivery success or retained value |
| Cost per delivered person | The same cost divided by accepted deliveries | Cost allocated across confirmed recipients | Return on investment or causal business impact |
Use one numerator when comparing these three views of the same cohort. Changing both cost scope and denominator without disclosure makes the apparent improvement uninterpretable. If finance needs a merchandise-only measure, publish it separately with that explicit label rather than presenting it as the complete program cost.
Build a numerator that finance can reconstruct
The workbook adds products, decoration, packaging, freight and fulfillment, duties, nonrecoverable tax, platform allocation, internal labor allocation and incremental replacement cost. It then subtracts accepted refunds and credits once. Each cost belongs to one source currency and the same cohort boundary. These are user-entered amounts, not a suggested tax schedule or a universal chart of accounts.
Assign shared charges using a documented rule. An annual platform fee might be allocated by an approved activity driver; a warehouse charge may need a usage-based allocation. Procurement supplies the contract scope, operations supplies the activity evidence, and finance signs the allocation. Do not choose whichever driver makes this campaign look cheapest. Keep excluded costs and the reason for exclusion available beside the supporting records.
Separate assigned cost from cash movement. A refundable, unused deposit is not automatically consumed program cost. A pending refund is not an accepted credit. A replacement may create additional merchandise and freight charges, but those charges must not appear both in the main cost categories and the incremental replacement row. Store the reconciliation outside the aggregate file when it contains recipient-level information.
The Internal Revenue Service’s business-gift guidance distinguishes deductible expenses and supporting records. Tax deductibility is not the same as operational expenditure or an authorization to spend. Keep the actual assigned cost intact, then let the responsible tax owner maintain any separate tax reconciliation. This workbook does not apply a deduction cap, classify employee benefits or decide whether an expense is deductible.
Fix the observation window and evidence owners
Write down when eligibility was frozen, which claim events qualify, when delivery observations stop, and how later returns will affect the next version. The workbook records a period start and cutoff date; source extracts also need a timezone and extraction timestamp in the evidence record. A date alone does not resolve a midnight boundary across markets.
The program owner signs the eligible population. Program operations deduplicates accepted claims and maps cancellations without destroying their history. Logistics defines acceptable delivery evidence and identifies pending, lost, returned and replacement orders. Finance owns cost allocation, accrued estimates, invoice reconciliation and the reporting currency. The analyst joins those records using stable campaign and cohort identifiers and keeps exceptions visible.
Record the cost ledger at a useful grain before entering its aggregate. A charge reference should identify the supplier invoice or credit, campaign, cohort, currency, service period, assigned category, allocation rule and reviewer. If one invoice covers two campaigns, the assigned portions should reconcile to the invoice once, not independently claim the full amount in both files. Keep an allocation remainder visible until its owner resolves it.
For labor, distinguish a measured time record from a planning allowance. A retrospective estimate may be acceptable for a provisional management review, but it must not become a measured efficiency claim. Record which activities were included, which rate finance authorized and whether the same scope was used in the comparison cohort. Otherwise a vendor-supported program can appear more expensive simply because its invoice exposes work that another team leaves outside the model.
Do not join a current delivery extract to an old roster without investigating the difference. A person transferred between groups may otherwise appear twice, while a deleted record may remove genuine cost from the denominator history. Preserve the earlier snapshot and produce a bridge that explains additions, removals and corrected identifiers. An unresolved difference is a reason to keep a result provisional, not to select the more convenient extract.
Close the observation window deliberately. A campaign with many parcels still in transit should not be declared inefficient solely because its early cost per delivery is high. Compare cohorts at the same maturity or show each observation age. Conversely, leaving a campaign open indefinitely can conceal unaddressed exceptions. Name the next review date and owner, while keeping the current cutoff unchanged for reproducibility.
Apply currency and tax conventions consistently
Each input column uses one source currency. The conversion rate means reporting-currency units for one source-currency unit; source net cost is multiplied by that rate. If source and reporting currencies match, the rate must be one. A missing, zero or negative rate blocks conversion. The sample rates are fictional and must never be used as current exchange quotes.
The IFRS Foundation’s overview of IAS 21 identifies currency and exchange-rate reporting as accounting questions. This operational template is not a financial-statement translation model. Finance must select the actual conversion source, date, purpose and accounting treatment. Preserve settlement differences separately when a management comparison uses a fixed analytical rate; do not claim that the analytical result equals the ledger.
Tax conventions also change comparability. One group may include a nonrecoverable charge while another reports a recoverable amount outside cost. Obtain an explicit finance decision before treating them as equivalent. The cost-basis label documents that decision; matching text is not evidence that the underlying invoices follow it. If a tax review is pending, retain the estimate and its limitation rather than setting the tax cell to zero.
Rounding belongs at presentation or the required transaction boundary, not repeatedly inside intermediate calculations. Keep sufficient conversion precision and show the display precision. A displayed two-decimal unit cost multiplied back by hundreds of people may not reproduce the unrounded total exactly. Explain that rounding difference instead of entering an unexplained balancing amount.
Worked decision one: reconcile 800 eligible people
Consider a hypothetical program with 800 eligible people, 500 accepted claimants and 450 accepted delivered recipients at the fixed cutoff. The team has 18,000 dollars of product cost, 1,000 of decoration, 1,500 of packaging, 4,500 of freight and fulfillment, 600 of duties, 400 of nonrecoverable tax, 1,000 of platform allocation and 2,000 of labor allocation. Incremental replacements add 1,000; accepted credits subtract 1,000. Net assigned cost is therefore 29,000 dollars.
The three results are 36.25 dollars per eligible person, 58 dollars per accepted claimant, and approximately 64.44 dollars per delivered person. None is the unique correct unit cost. Together they explain the cost boundary and the population reached. Reporting only 36.25 as the cost of a successful gift would confuse offered eligibility with confirmed delivery.
The analyst’s inputs are a versioned roster, deduplicated claims, original-to-replacement order links, carrier or recipient evidence, invoices and a signed allocation schedule. Operations confirms that the replacement cost is incremental and that replacement recipients remain counted once. Finance independently enters 29,000 as the invoice/control total, rather than copying the workbook’s result into its own check. A zero difference then supports a genuine tie-out.
The team considers three alternatives: report only product cost, wait for every exception to close, or issue a provisional full-cost report with named unresolved items. The first hides delivery and operating charges. The second delays a useful review. The chosen provisional report preserves the full cost, all three denominators and the maturity date. It does not describe estimates as paid invoices.
Suppose a missing freight adjustment appears during review. Leave its cost input unknown until logistics supplies a supported value; the workbook should stop the affected result rather than silently substitute zero. After correction, preserve the old export, reconcile the difference and issue a new dated activity version. Acceptance requires matching populations, an independently checked numerator, replacement deduplication, explicit unresolved items and a reviewer able to reproduce the three divisions. It does not require an artificially low unit cost.
Worked decision two: compare currencies without manufacturing savings
Now add a hypothetical second cohort with 400 eligible people, 250 accepted claimants and 225 delivered recipients. Its assigned source-currency costs total 480,000 Taiwan dollars before accepted credits of 16,000, leaving 464,000. Using the fictional rate of 0.03125 reporting dollars per Taiwan dollar gives 14,500 reporting dollars. Its three ratios happen to match the first cohort: 36.25, 58 and approximately 64.44.
The matching ratios do not prove equal service, satisfaction, tax treatment or commercial performance. The analyst first compares purpose, policy, recipient definition, observation period, evidence maturity and cost scope. Finance checks whether tax inclusions and allocation rules are compatible. Logistics checks that “delivered” uses equivalent evidence. Until those decisions are made, the two rows remain useful independently, but the team should not publish a combined performance claim.
If finance confirms a matched basis, aggregate the converted costs and matching people: 43,500 reporting dollars divided by 675 delivered people is approximately 64.44. Do not average regional unit costs unless the weighting happens to be identical; a ratio of sums is the repeatable method. The template suppresses its combined result when reporting currency, cost-basis label, comparison group, period or evidence-state declarations differ.
The team compares using current settlement rates, a fixed management rate, or no combined comparison. Settlement rates help explain actual accounting exposure but can confound an operating comparison. A fixed analytical rate can isolate one difference but needs a separate settlement bridge. Keeping separate results is the right answer when definitions cannot be reconciled. The workbook does not choose that accounting policy for the organization.
If a rate was accidentally entered in the opposite direction, stop the comparison, recover the original rate evidence and correct the input. A positive number alone cannot prove direction. Independently multiply one known source amount and verify the expected order of magnitude. Acceptance evidence includes the approved rate direction and date, compatible cost definitions, separate local-currency totals, a reproducible weighted calculation and a signed comparison decision. Equal-looking outputs are not acceptance evidence by themselves.
Recover refunds, missing values and misleading averages
When credits exceed the current cohort’s cost
Keep the negative result visible and investigate the credit’s source period. It may represent an older campaign, a supplier correction or a genuine net recovery. Do not clamp it to zero, invent extra recipients or call it a profitable gift program. Finance decides whether to move the credit into a separate reconciliation and issues a dated corrected version when needed.
A known zero and an unknown value require different actions. If no duty applies under the approved treatment, enter numeric zero and retain the evidence. If the duty invoice is missing, leave the input blank and assign the investigation. A blank denominator after zero deliveries means that the ratio is unavailable; it does not mean delivery cost nothing. A failed lookup, a text value in a cost cell or a missing owner should not become a healthy-looking zero.
Cancellation and return conventions must remain stable. An accepted claim later cancelled may remain part of a cumulative accepted-claim denominator, while a current-active-claims report would exclude it. Both can be useful if labeled, but they cannot share one unexplained column. Similarly, decide whether a later return reverses an accepted delivery, then apply that rule to every comparable cohort and disclose the revision window.
Avoid averaging averages across small and large programs. Even after currency conversion, a small expensive group and a large inexpensive group need their actual weights. Also inspect the groups individually: a stable total can hide a worsening route or a supplier problem. Investigate the underlying exception records before inferring poor recipient engagement or a vendor defect from a single aggregate ratio.
Run the workbook as a controlled review process
Use a spreadsheet when two cohorts, transparent formulas and a small number of reviewers are sufficient. A governed analytical model becomes more appropriate when many campaigns require repeatable joins, access control and refresh schedules. Do not maintain a manual blended metric if the populations and accounting bases cannot be aligned. These are operating choices, not a ranked platform comparison.
Before circulating the result:
-
Freeze the cohort identifiers, population definition, observation date and source versions.
-
Reconcile distinct eligible, accepted-claim and delivered people without counting replacement parcels twice.
-
Populate every cost category, distinguishing known zero from unknown and accepted credits from pending refunds.
-
Confirm tax basis, allocation owners, conversion direction and the rate’s dated evidence.
-
Recalculate the file, inspect missing-input warnings, and independently recompute one cohort.
-
Tie the numerator to an independent control total and document every nonzero difference.
-
Review the comparability decision before using the combined ratio and retain the prior export.
The file was checked in its authoring engine and after export and reopening. Recalculate and test representative changes in your own spreadsheet application before operational use. Its checks do not certify invoice completeness or replace human review. Refresh this annual template after formula corrections and whenever definitions change; preserve activity-specific versions so a future reviewer can reproduce the decision.
Turn a cost ratio into a clear next decision
A useful unit-economics review ends with a concrete action: obtain a missing invoice, revise a replacement process, keep two markets separate, or repeat a program on a verified cost basis. Keep all three denominator views available and explain which decision each supports. Pair cost efficiency with recipient experience and separately designed outcome measurement; a lower number alone does not establish a better gift or a successful relationship.
Giftpack’s official FAQ describes invitation, claim, spend and delivery metrics while noting that there is no single prescribed return-on-investment formula. Where the contracted scope supports the required reporting and fulfillment evidence, Giftpack can help execute the approved gifting program. Finance, tax, legal, privacy and program owners remain responsible for eligibility, accounting, data use and approval; the platform and this workbook do not replace those decisions.