A reliable apparel order starts with confirmed recipient sizes, then uses a transparent scenario only for the people whose sizes are still unknown. This calculator turns headcount, regional mix, an editable size profile, exchange allowance, uncertainty buffer, case packs, and current stock into an auditable order plan.
Download the English workbook — v2 · 2026-09-09
Download the English CSV snapshot
Read the result before changing the model
For the wider onboarding program, see new-hire welcome kits: budget, contents, sizes, and global delivery.
The sample model begins with 1,000 recipients, 35% confirmed sizes, a 3% exchange allowance, an 8% uncertainty buffer, three regional weights, and a 12-unit case pack. These are deliberately editable assumptions, not a universal distribution. The workbook separates confirmed demand from modeled demand so operators can see how much of the purchase order depends on evidence.
| Output | What it answers | Decision |
| Weighted size share | How regional scenarios combine | Review with local owners |
| Confirmed demand | Units backed by recipient data | Protect from arbitrary overrides |
| Modeled demand | Units estimated for unknown recipients | Run sensitivity cases |
| Case-rounded reorder | Purchasable units after stock and packing rules | Approve the overage explicitly |
A size curve is most useful as a planning control: it exposes assumptions, rather than hiding them behind a single “industry average.”
Collect the smallest set of inputs that changes the decision
Start with program headcount by destination or warehouse. Record the share of people whose size has been confirmed, the garment style and fit, the supplier size chart date, current on-hand stock, minimum order quantity, case pack, and unit cost. Keep exchange allowance separate from uncertainty buffer: exchanges reflect fit and preference risk, while uncertainty reflects incomplete demand information.
-
Confirm the exact garment style, cut, fabric, and decoration method.
-
Obtain recipient-level sizes with appropriate consent when operationally possible.
-
Split known and unknown demand instead of applying a model to everyone.
-
Set regional weights from the actual recipient list.
-
Enter current usable stock by size and location.
-
Record supplier minimums, case packs, and order deadlines.
-
Assign an owner for overrides and final purchase approval.
Do not collect body measurements merely because the workbook can store a size. A selected garment size is usually sufficient for fulfillment and is less intrusive than detailed measurement data.
Treat every size profile as a scenario, not a population fact
There is no defensible universal corporate-apparel size curve. Different cuts, brands, garments, cohorts, climates, and local preferences change the result. Even the same label can map to different measurements across products. Gildan’s official size charts separate products and audiences, while SanMar’s first-party fit resources explicitly direct buyers to brand-specific guidance.
A sample curve is a temporary decision aid for unknown demand. It is not evidence about an individual and should never override a confirmed size.
When a supplier or program owner proposes a default curve, document its sample, geography, garment, collection date, and exchange rate. If those details are absent, label it as an internal planning assumption with low confidence. Update it after each campaign using actual selections and exchanges, without using the data to infer protected characteristics.
Follow the calculation from people to purchasable units
The workbook uses actual confirmed counts for each size. Regional weights are applied only to recipients whose sizes remain unknown. Exchange reserve covers confirmed plus modeled demand; uncertainty reserve covers modeled demand only. Usable inventory is deducted before the final case-pack rounding.
weighted_share = Σ(region_program_mix × region_size_profile)
confirmed_units = actual confirmed recipient count for this size
unknown_headcount = headcount − Σ(all confirmed size counts)
modeled_units = unknown_headcount × weighted_share
buffer_units = (confirmed_units + modeled_units) × exchange_rate + modeled_units × uncertainty_rate
target_units = confirmed_units + modeled_units + buffer_units
net_need = max(0, target_units − usable_on_hand)
purchase_units = ceiling(net_need ÷ case_pack) × case_pack
The example contains 350 confirmed selections and 650 unknown recipients. Confirmed counts must be entered by size; changing a percentage alone cannot substitute for those selections. Keep fractional modeled demand and reserve quantities until the final purchase rounding. The scenario-comparison worksheet and CSV are dated snapshots, not live comparisons: after changing inputs, recalculate the main model and regenerate the comparison and export before approval.
Rounding occurs late so the model does not repeatedly compound overage. Review very small and extended sizes one by one: a 12-unit case pack can create a large percentage overage when net need is only one or two units. If the supplier permits mixed-size cases, change the packing rule rather than forcing a full case per size.
Adjust regional mix without turning geography into a stereotype
Regional scenarios should come from actual campaign data, supplier fit feedback, or a documented pilot. Geography alone is not a reliable proxy for body size. The workbook uses regional weights to support warehouse planning and different product assortments, not to assign sizes to individuals.
For a global program, consider one curve for each distinct garment and cut. A unisex hoodie, women’s fitted tee, performance polo, and safety jacket should not share one profile. If a market uses different label conventions, map each local supplier’s measurements to the offered product rather than assuming that an alphabetic label is equivalent everywhere.
The safest operational design gives recipients a choice before production whenever lead time permits. If the program must pre-buy, use a conservative pilot, reserve undecorated blanks where possible, and keep the model’s unknown share visible to approvers.
Separate buffer policy from supplier constraints
Buffer stock protects service levels, but it also creates cost and waste. The base workbook exposes exchange allowance and uncertainty as separate inputs. Operators can reduce the uncertainty component as confirmed-size coverage increases, while keeping a smaller exchange reserve for fit issues.
| Scenario | Use when | Review focus |
| Lean | Most sizes are confirmed and replenishment is fast | Stockout risk and deadline |
| Base | Some sizes are unknown and exchanges are manageable | Balanced service and overage |
| High uncertainty | New garment, weak data, long lead time | Cost cap and post-event reuse |
Minimum order quantity belongs at the style or supplier level, whereas case pack often applies at the size or color level. Confirm the contract before interpreting either input. A model that meets a total minimum may still be impossible to order if each size has a separate pack rule.
Use scenario comparison to approve risk, not just quantity
Review at least three cases before issuing the purchase order. Change one assumption family at a time: confirmed-size share, regional mix, size profile, exchange allowance, uncertainty buffer, or pack rule. Compare total planned units, case-pack overage, extended-size coverage, estimated merchandise cost, and units that could be reused in a later program.
Ask approvers to choose a service-risk posture. A low buffer can be correct for optional event merchandise with rapid reorders. A higher buffer can be justified for fixed-date onboarding waves or remote markets with long lead times. The workbook is not designed to “optimize” one answer automatically; it makes the tradeoff reviewable.
The CSV export is a locale-matched, flat output for warehouse, purchasing, or decoration handoff. The XLSX remains the source for assumptions and formulas. Do not edit the CSV and then treat it as the model of record.
Govern the workbook like a purchasing control
Name an owner, a version, and an approval date. Lock or protect formula cells in the purchasing copy if your process requires it, while preserving an editable planning copy. Archive the approved input snapshot with the purchase order and supplier quote.
-
The size profile totals 100%.
-
Regional mix totals 100%.
-
No input or inventory value is negative.
-
Confirmed and modeled demand are visibly separated.
-
Every formula error scan is clear.
-
Every worksheet has been visually reviewed.
-
Supplier chart, fit, minimums, and case packs are current.
-
Case-rounded units are not below net need.
-
Cost and overage are approved.
-
Version and change log are complete.
For recurring programs, compare forecast with selections, shipments, exchanges, unused units, and stockouts. Refresh quarterly or when the garment, supplier, cut, region, or program audience changes materially.
Common questions and exceptions
Should we use an industry-average size curve?
Use one only as a clearly labeled temporary scenario when recipient data is unavailable. Do not present it as a fact. Record the source and test a range around it.
What if employees do not want to share a size?
Offer a recipient-choice workflow, a redeem-later option, or a non-sized alternative. Participation should not depend on disclosing more personal information than fulfillment needs.
Can one curve cover all apparel?
No. Separate curves by product, cut, supplier, and meaningful fit difference. A tee and outerwear item may require different allowances even for the same audience.
How should we handle extended sizes?
Validate availability and measurements before launch, avoid hiding extended sizes in an “other” bucket, and review pack-rounding overage explicitly. Never infer demand from stereotypes.
Worked example: preserve confirmed choices in a one-hundred-person order
Consider a hypothetical program with one hundred recipients and three offered sizes for a simplified demonstration. Ten people have selected small, twenty medium, and ten large. Forty choices are therefore confirmed and sixty remain unknown. The planning profile for those unknown people is twenty percent small, fifty percent medium, and thirty percent large. These are invented scenario inputs, not a population distribution or a Giftpack customer result.
The unknown portion becomes twelve, thirty, and eighteen units. Adding the actual confirmed counts gives twenty-two small, fifty medium, and twenty-eight large. Notice that the model never replaces the ten confirmed small selections with twenty percent of the confirmed population. Those selections are evidence. The profile applies only to the sixty people who have not selected a size.
Now assume a three-percent exchange reserve and an eight-percent uncertainty reserve. Small receives 0.66 exchange units plus 0.96 uncertainty units, for 1.62 reserve units. Medium receives 1.5 plus 2.4, for 3.9. Large receives 0.84 plus 1.44, for 2.28. Fractional units represent planning expectations; they are not instructions to ship fractions of garments.
Before inventory, the targets are 23.62 small, 53.9 medium, and 30.28 large. Suppose there are twelve usable medium garments and no usable stock in the other two sizes. Net requirements become 23.62, 41.9, and 30.28. With twelve-unit packs purchased separately by size, the order becomes twenty-four small, forty-eight medium, and thirty-six large: 108 purchased garments.
At an illustrative merchandise cost of twenty currency units per garment, that purchase costs 2,160, excluding decoration, freight, taxes, handling, and other charges. The existing twelve medium garments are stock consumed by the plan, not additional units purchased in this order. Keep those two quantities separate when presenting both purchasing cost and available supply.
There is no contradiction in purchasing more than the unknown headcount. The purchase covers confirmed demand as well, then accounts for reserves, inventory, and pack rounding. The useful approval question is whether each component is justified. A buyer should be able to trace the 108 units back to the selected sizes and assumptions instead of accepting the total because it looks plausible.
Before approving, compare the model with the actual supplier offer. If small and large cannot be bought in twelve-unit packs, this example is not directly orderable. If existing medium stock belongs to another campaign or has a different cut, it is not usable inventory for this calculation. Correct those facts and recalculate; do not change the formula output manually to match an attractive purchase total.
Worked example: a rare size creates a pack decision
For a second hypothetical decision, assume the model shows a net requirement of two units in an extended size after the approved reserve and usable stock have already been considered. A supplier offers a twelve-unit case at twenty currency units per garment. Purchasing the case costs 240 and creates ten units above the modeled net requirement.
Suppose another reviewed supplier can deliver two equivalent garments individually at twenty-eight units each, with forty units of additional handling. That alternative totals ninety-six before other excluded charges. On those stated assumptions, the smaller order costs less despite its higher unit price. This does not establish that either supplier is available or suitable; it demonstrates why unit price alone can mislead.
The buyer still checks garment equivalence, decoration quality, size measurements, arrival date, return terms, and any separate order minimum. If the alternative product has a different fit, the confirmed recipient should be offered its actual measurements or another suitable choice. A size label copied from the original garment is not enough to establish equivalence.
There may also be a third option: hold unprinted stock already owned by the company and decorate only the quantity needed. That works only if the stock is genuinely available and the decoration schedule fits. Do not count goods allocated elsewhere as a free reserve. Record who releases the stock and which campaign bears the cost.
The decision owner compares total landed cost and service risk, then records the selected option. If the twelve-unit case is chosen because the remaining ten units can support a recurring program, document that reuse assumption and an owner for the stock. Do not label the excess as guaranteed savings or assume every unused garment will eventually be worn.
This case illustrates a model boundary. The workbook applies its configured pack rule; it does not negotiate suppliers, optimize mixed cases, or guarantee extended-size availability. A manual exception belongs in the purchasing decision record with its rationale. Preserve the model output alongside the final order so the difference remains explainable.
Test the model before using a real recipient list
Start with a synthetic copy. Set all recipients as confirmed, with counts that sum to headcount. Unknown demand should become zero, and uncertainty reserve should disappear because there is no modeled unknown demand left. Exchange reserve can remain. If the final purchase total does not fall, inspect pack rounding before assuming a calculation error: reserve changes can stay within the same pack boundary.
Next test usable stock above the target for one size. Its net need should stop at zero rather than producing a negative purchase. Do not use surplus small garments to offset a large-size shortage. Inventory subtraction is size-specific and also assumes the same garment, color, location, quality, and availability.
Then test invalid inputs deliberately in the synthetic copy. Confirmed counts above headcount and regional weights that do not total one should prevent normal model output. The workbook includes basic checks, but these are not comprehensive input validation. Operators must also reject blank required fields, fractional people or pack counts, negative profile entries, and other nonsensical assumptions before approval.
Use the formula cells for calculation and the designated input cells for planning changes. Copying values over formulas can create a workbook that appears correct for one scenario but no longer responds to new inputs. After a bulk paste, compare formulas and recalculate before generating the purchasing export.
Keep a short acceptance record: the synthetic inputs, expected behavior, observed behavior, workbook version, and reviewer. The aim is to test meaningful failure modes, not to produce a long checklist that nobody can reproduce. Use only synthetic records for testing rather than circulating employee data unnecessarily.
Turn late responses into a controlled revision
A late response should move one person from unknown demand to that person's confirmed size. Do not add the person to confirmed counts while leaving them in the unknown population. Headcount stays fixed unless eligibility actually changes. Recalculate the weighted unknown allocation and reserve after the update.
If the purchase order has not been placed, the new version can replace the earlier planning copy after review. If production is committed, show the difference between the new target and committed stock instead of pretending the order can still be changed without cost. Procurement decides whether cancellation, an additional order, or an approved stock transfer is feasible.
Set a response cutoff that reflects production and delivery lead time. Communicate what happens after the cutoff: later fulfillment, available-stock choice, or another approved option. Avoid presenting a modeled allocation as though the recipient personally selected it. The recipient experience depends on that distinction as much as the spreadsheet does.
When a participant changes size after dispatch, record an exchange rather than overwriting the original selection. Capture the outgoing size, replacement size, reason where appropriate, and whether the returned garment is reusable. Those events improve the next planning profile without corrupting the evidence of what happened in this campaign.
Close purchasing with a versioned handoff
The purchasing packet should identify the exact garment, fit, color, size labels, supplier chart version, destination, quantities, decoration, price assumptions, and approval. The workbook's total is only one part of that packet. A supplier needs a clear order by purchasable item, not an unexplained aggregate.
The downloadable comparison sheet and CSV are snapshots. Editing workbook inputs does not update an already downloaded CSV or make the saved scenario comparison live. Recalculate the main model, review the selected assumptions, and create a fresh export before sending an order. Label the export with the same approved version.
Ask the receiving team to confirm the size rows and quantities they accepted. A sent file is not evidence that the supplier imported it correctly. Check for dropped extended sizes, changed labels, unexpected decimal interpretation, and a total that includes reserves twice. Resolve any mismatch before authorizing production.
After the campaign, reconcile purchased units, existing stock consumed, delivered garments, exchanges, damaged units, and remaining usable stock. Keep leftovers by actual item and location. Compare modeled unknown demand with subsequent selections separately from confirmed demand, so forecast evaluation does not credit the model for choices that were already known.
Finish with an approved, learnable order plan
The best size plan maximizes confirmed choices, limits modeling to unknown demand, and makes every buffer and rounding decision visible. Save the approved workbook version with the supplier quote, monitor actual selections and exchanges, and use the evidence to improve the next program rather than claiming a universal curve.
Once policy, privacy, supplier, and purchasing decisions are set, Giftpack can serve as the execution layer for recipient choice, branded-merchandise sourcing, storefront experiences, and global fulfillment; it does not replace those decisions.