Peptide Order Contribution: Reconcile Refunds, Fees, and Reserves

Before approving another campaign, ask finance to reproduce two reconciliations for your last shipped cohort: contribution after specified costs and cash received after processor movements. In the example, 100 shipped orders and 200 units produce $1,650 of contribution, but the expected bank deposit is $7,640 and the stated current-window cash increase is $4,850: three different measures.

Use one event register to connect the order export, refunds, cost documents, processor payout, and bank receipt. Reconcile the population before calculating a rate. A reserve addition reduces current availability in this model; it is not automatically a fee. An earlier inventory payment affects this window's cash bridge but not the cohort's inventory cost.

Examples are hypothetical US dollar inputs. This runbook proposes a workflow, not an accounting standard or provider policy; apply within actual product, payment, customer-remedy, and accounting requirements. No quoted fee, business approval, customer outcome, or normal industry margin is asserted.

Keep three measures in view

Two reconciliations produce these three measures. The first follows the cohort's economics; the second follows settlement and bank cash. A cash increase can include funds needed to replenish inventory. The deposit can include money from an older cohort. Read the definitions before deciding which number belongs in a promotion review or a payment decision. None includes every company cost or cash event.

Assemble the reconciliation in a repeatable sequence

First freeze the population. Save the order identifiers and the shipment criterion used to select the 100 orders. Include the time zone and cut-off so a later export does not drift into a different population. Confirm the 200-unit total against shipment detail. If an order was only partly shipped, state how the cohort handles the shipped and unshipped portions before assigning any dollars.

Next establish a charge checkpoint. Use the original merchandise, discount, and customer-shipping fields to reproduce $9,600. Test one discounted order manually. This is where a net-sales export can cause a second discount deduction. The working file should preserve the source field name and its meaning, not just a copied column headed revenue. A reviewer needs to know whether the $1,000 has already been taken out.

Then attach later events to original orders. Build one row per refund, fee, dispute movement, replacement expense, or correction, with its own event reference. Keep the event date alongside the cohort membership date. This lets an older cohort receive a later adjustment while the cash schedule places that same event in the correct current period. It also prevents a late refund from being assigned to the newest cohort merely because it happened this week.

Assign costs through a separate document register. An invoice can feed several orders, and an order can consume several services. Record the allocation basis and verify that the total assigned does not exceed or omit the amount in scope. For the main example, the inventory cost is $20 for each of 200 shipped units. The other components retain their stated order, shipment, transaction, incident, or campaign relationship.

Only after those checks should the worksheet calculate $8,800 minus $7,150. Keep the before-acquisition result of $2,850 visible beside the after-acquisition result of $1,650. That pair shows what the $1,200 convention changes. It also gives finance a direct place to revise the analysis if the allocation is corrected, without editing the underlying customer-charge bridge.

Resolve exceptions in an event register

Give each exception a state, owner, source reference, and next evidence request. Useful states include recorded, estimated, disputed, approved credit, and applied credit. The exact labels are an operating choice; their meaning should be stable. A warehouse acknowledging a dispute does not put $100 in the bank, and a credit memo against next month's bill does not create an immediate cash receipt.

Keep a duplicate check on the financial event identifier, not just the order identifier. A single order can legitimately generate several refunds. Conversely, the same refund can appear in a storefront export and a processor export. Those two records may describe one event. Connect them rather than deducting both. Review unmatched events explicitly instead of dropping them from the model because their identifiers differ.

For disputes, preserve the sequence. A provisional debit and its reversal can net to zero principal while leaving a fee and a temporary cash constraint. Marking both as independent finalized losses would distort contribution; ignoring both in a liquidity review would hide the period without funds. The register should support both views through the same source trail.

Use a separate test for replacement costs. The main illustration's $90 is only extra shipping. Ask whether actual replacement material or warehouse work created additional costs, and assign those costs once where applicable. A zero-price replacement order is a fulfillment instruction, not proof that no resources were used. The exception cannot be closed merely because the storefront reports no additional revenue.

Hand the result from finance to the operating owner

Finance should hand over the cohort definition and the bridge, not only a percentage. The receiving owner should be able to identify the 100 orders, the observation date, the $8,800 revenue base, and the included $7,150 cost scope. The $16.50 per order and $8.25 per shipped unit are useful summaries only while that population and scope remain attached.

Ask the marketing owner to confirm what the acquisition allocation can support. Ask operations to confirm the fulfillment basis and any unresolved replacement resources. Ask the financial owner to trace the $7,640 expected payout to its actual settlement window. These requests resolve specific inputs. They do not turn a managerial estimate into an audited financial statement.

After the result is accepted for its stated purpose, preserve a dated snapshot. A subsequent $100 warehouse credit produces a new bridge and an updated $1,750 contribution under the stated cost-reduction treatment. Keep its cash settlement separate. The next reviewer should see exactly why the number changed rather than receiving a rewritten historical worksheet with no explanation.

Finish with a short action record: the proposed operating choice, the supported result, the unresolved amount that matters, the owner, and the next observation. If the file still has an unexplained payout difference, the reconciliation remains open even if contribution is calculable. If the cohort is profitable under the narrow measure but cash is constrained, the next action belongs in the payment calendar. This sequence gives the founder an operational handover that can be followed, checked, and updated without inventing missing certainty.

Define the cohort and the question

Choose the population before calculating the margin. A cohort might be orders shipped in a specified week, orders placed during a campaign, or completed orders from a defined customer group. Those populations answer different questions. Use one definition and retain the order identifiers that establish membership.

For the main example, the cohort contains 100 orders, all shipped, containing 200 finished units. The original merchandise amount before discounts is $10,000. Customers are charged $600 for shipping, and merchandise discounts total $1,000. Sales taxes and similar amounts collected for remittance are excluded from the management revenue model used here; the actual accounting treatment needs appropriate review.

State what the result is intended to support. A founder may want to evaluate a promotion, a product mix, a fulfillment route, or whether a cohort is mature enough to judge. A report for choosing tomorrow's price can use estimates. A report explaining a completed cohort should distinguish actual recorded events from costs still estimated or unresolved.

Set a cut-off date for the observation. Some refunds and disputes arrive later. A cohort measured seven days after shipment may look different from the same cohort measured after more of its outcomes have developed. Keep the earlier snapshot when updating the result so the business can see how much changed and why.

Do not call the result net profit if the model excludes fixed overhead, financing, taxes, owner compensation, or other costs. Use a precise label such as modeled cohort contribution after specified variable costs. A limited but well-defined measure is more useful than an impressive percentage whose missing costs are invisible.

Build a reconciliation file that another person can use

Keep the cohort definition, order list, source exports, field meanings, cost allocation rules, formulas, and observation date together. Use transaction identifiers to connect the records. Preserve original exports separately from working calculations so a reviewer can distinguish source evidence from adjustments made in the model.

Add control totals. Order counts should match the defined population. Merchandise, discounts, shipping, refunds, and charge totals should reconcile to the relevant source reports under the same scope. Cost allocations should sum to the amount being allocated. Processor events should bridge to payouts and then to the bank observations they actually concern.

Use negative checks as well as successful examples. Confirm that a refund cannot be assigned to two cohorts, a reserve release does not create sales revenue, and a fee already netted from a deposit is not deducted again in the cash bridge. Deliberately testing these mistakes is more useful than checking only a final total that was copied from the worksheet.

Protect customer information in the working file. The financial task usually needs order and transaction references, amounts, dates, and relevant status, not every customer message or sensitive attachment. Use limited fields and appropriate access. A useful financial model should not become an unnecessary second customer database.

Record adjustments with reasons. If the warehouse invoice is partly estimated, a refund allocation is corrected, or a processor event remains unexplained, keep that status visible. The goal is not a perfectly smooth report. It is a reproducible explanation of the best supported result and the gaps that could still change it.

Reconcile original charges before subtracting costs

Build the revenue bridge from source records. Start with merchandise before discounts, subtract merchandise discounts, add customer-paid shipping included under the model, and then account for actual refunds and other adjustments. Keep each component separate so the totals can be checked against order and payment records.

In the hypothetical example, $10,000 minus $1,000 plus $600 equals $9,600 charged before the excluded tax amounts. This is not the same as the merchandise sales figure and not necessarily the same as the bank deposit. It is the starting customer-charge amount for the defined cohort under the stated scope.

Check whether the storefront's export already reports net merchandise sales. If it does, subtracting discounts again understates revenue. A field labeled sales can mean gross merchandise, net merchandise, total order amount, or something else. Read the current field definition and verify it against a few order lines rather than trusting the label.

Keep cancellations separate from shipped orders. If the cohort is defined by shipments, an order canceled before shipment may be outside it. Its processing costs and marketing effects may still matter to a broader campaign analysis, but they should not appear inconsistently in a shipped-order denominator. Add a separate population when needed instead of quietly changing the cohort.

Reconcile the order count and unit count as well as dollars. One hundred orders containing 200 units produce different per-order and per-unit measures. A promotion that changes basket size can improve one measure while worsening another. State the denominator beside every average.

Assign refunds to the original economic event

The main example has $800 of recorded refunds by the cut-off date: $700 relating to merchandise and $100 relating to shipping. Net modeled customer revenue is therefore $9,600 minus $800, or $8,800. The split helps explain the result and should follow the actual refund records rather than an assumed allocation.

Link each refund to its original order and, where available, its affected lines. A refund recorded this month for an earlier cohort should update that cohort's economics while also appearing in the appropriate current cash period. Cohort attribution and cash timing are two views of the same event, not competing definitions of truth.

Distinguish requested, approved, initiated, and completed refunds where the systems provide those states. A support promise is not the same observation as a processor transaction. The management report should identify which states it counts and what remains outstanding. Do not make a refund disappear merely because it has not yet reached the bank statement.

Avoid treating every returned unit as saleable recovered inventory. The material's disposition requires its own authorized assessment. In this example, no recovered inventory value is credited, so the model retains the original shipped-unit cost. If the business has a justified recovery, record the quantity, cost basis, and actual disposition separately under the appropriate accounting and release rules.

Keep the applicable customer remedy and legal obligations outside the desire to improve the cohort margin. The worksheet explains the financial result; it does not decide whether a customer is owed a refund. A processor restriction or supplier dispute should not be used as a mathematical reason to deny an otherwise required remedy.

Use a documented inventory cost basis

The hypothetical 200 shipped units carry a management inventory cost of $20 each, for a total of $4,000. State what that $20 includes. It might contain purchase cost and selected inbound costs under the business's reviewed method. It should not silently include the same testing or freight charge that is subtracted again elsewhere in the worksheet.

The IRS's small-business guidance discusses inventory and cost of goods sold with method-specific rules and exceptions. Those tax rules should not be replaced by a marketing worksheet. Use an appropriately reviewed accounting method for the books, and describe any management allocation used here so the two can be reconciled. IRS Publication 334.

If several lots were shipped, use their actual assigned costs under the selected management convention. A single convenient average can obscure a costly replacement lot or a change in inbound expense. Preserve the lot-to-order link where the business has it, while avoiding a level of allocation precision the records cannot support.

Do not change the cost denominator to units purchased if the report concerns units shipped. A purchase of 500 units and shipment of 200 units are different events. The remaining inventory can matter to cash and balance-sheet questions, but charging the entire purchase to this cohort would answer a different economic question unless that is the explicitly chosen analysis.

Record estimates separately from posted costs. If the final supplier adjustment or inbound invoice is not available, use a labeled estimate and update it when the actual document arrives. A provisional cost should not become a permanent fact because the worksheet was copied into next month's report.

Add fulfillment and outbound delivery once

In the example, the cohort incurs $700 of fulfillment charges and $800 of outbound carrier charges. These are separate modeled costs. Confirm what the fulfillment invoice includes before adding them: some arrangements combine services, while others bill each component separately. The worksheet should follow the actual contract and invoice units.

Allocate order-level costs consistently. A pick-and-pack charge may belong to the order, while additional-item handling may depend on units or lines. A carrier charge may cover a whole shipment containing several products. Do not charge the full shipment cost to every product line and then add it again at the order level.

If the cohort contains split shipments, retain the relationship between the order and its shipment events. One order can generate two carrier charges and additional warehouse work. A per-order average based on one shipment assumption can understate the costs of an exception-heavy cohort.

Identify amounts still estimated. A carrier adjustment or warehouse credit may arrive after the first report. Keep a provisional line with its basis, then replace it with the actual charge or credit. Preserve the variance so the business can distinguish a forecast error from a change in the underlying service mix.

Check customer-paid shipping separately from carrier expense. The example includes $600 collected, $100 refunded, and $800 of carrier charges. That does not mean shipping contributes $600. The model must include both revenue adjustments and the actual cost under the stated scope. A free-shipping promotion changes the revenue side even when the carrier still invoices the business.

Separate payment fees from unavailable funds

The hypothetical cohort incurs $320 of processing fees and $40 of additional refund-related fees. These are stated scenario inputs, not a quote from a payment provider. Confirm actual fee bases, currencies, fixed charges, retained fees, and adjustment behavior from the business's agreement and transaction records.

A reserve is a different category. In the example, $1,000 of the cohort's funds is withheld in a reserve under assumed terms. The management contribution model does not automatically treat the full reserve addition as an expense. The cash model does treat it as unavailable for current payout. Recoverability and any required accounting adjustment need their own assessment.

Provider documentation illustrates why the distinction matters. Stripe describes reserve balances and release treatment in its support materials, but that is a provider-specific example rather than a universal contract. Its peptide-business policies also include conditions; this article makes no claim that a particular store is approved. Stripe reserve information, Stripe business-policy FAQ.

Do not infer fee expense from the difference between charges and the bank deposit. That difference can include refunds, reserves, delayed availability, currency conversion, prior-period adjustments, and other transactions. Reconcile the processor ledger by event type. A small deposit does not establish that the processor charged an equally large fee.

Keep disputed deductions visible. If an amount is a provisional hold, an actual fee, a disputed charge, or an unexplained adjustment, label it accordingly. The report should not improve its margin by ignoring an unresolved cost or worsen it by calling every unavailable dollar permanently lost.

Record replacements and recovery costs separately

The example adds $90 of replacement-related shipping cost. This is separate from the $800 refunded and from ordinary outbound carrier charges. Identify the order and incident behind each recovery cost so the same event is not charged twice under different labels.

A replacement can involve a new unit, additional warehouse work, another label, and customer communication. The original shipped-unit cost may remain while the replacement introduces another cost. If the model includes replacement inventory, record it explicitly. In this simplified example, the stated $90 line is only the additional shipping cost, and no unstated replacement-unit cost is assumed.

Do not use a zero-priced replacement order as evidence that the replacement cost nothing. Revenue and cost are different observations. A system may create a free order to control fulfillment while the material and service costs still affect the original cohort's economics.

If a supplier or carrier later reimburses an amount, link the recovery to the event and distinguish approved, invoiced, and received states. A promised credit is not bank cash. The economic treatment should follow the reviewed method and actual evidence, with the cash schedule showing when money or an applicable credit becomes usable.

Use the event-level detail to identify improvement opportunities. A repeated packaging issue, incorrect pick, or unclear delivery instruction may generate recovery costs that a broad monthly shipping total conceals. The worksheet can help locate the problem without making unsupported claims that every exception was preventable or that a single process change will eliminate it.

Attribute acquisition cost without inventing certainty

The main example assigns $1,200 of acquisition spending to the cohort under an explicit management rule. That produces a useful scenario only if the rule is stated. A campaign-level spend allocation, a customer-level model, and a platform's reported conversion cost may produce different results and have different limitations.

Record the source period and included spending. If the cohort consists of orders shipped this week but the advertising occurred earlier, the attribution should explain that connection. Do not subtract only the spending that happens to appear in the current bank week and call it the cohort's full acquisition cost.

Keep repeat-customer and new-customer views distinct where the data supports them. Dividing all marketing spend by only new customers can answer one question; dividing it by all orders answers another. Neither is automatically wrong, but an unlabeled denominator can make comparisons misleading.

Avoid treating a platform's attribution output as proof of causation. It may assign credit under its model without establishing that the order would not have occurred otherwise. For this worksheet, describe the amount as attributed acquisition cost under the chosen rule. A stronger causal claim requires a different evaluation design and evidence.

Where attribution is weak, show a range or a broader campaign contribution view rather than fabricating order-level precision. The purpose is to understand the decision's sensitivity to acquisition spending. A report with an explicit uncertain input is more useful than an exact-looking margin whose most important cost was assigned arbitrarily.

Calculate contribution from the defined components

The example's net modeled revenue is $8,800. The included costs are $4,000 of shipped inventory, $700 of fulfillment, $800 of ordinary outbound delivery, $320 of processing fees, $40 of refund-related fees, $90 of replacement-related shipping, and $1,200 of attributed acquisition spending. Adding those costs gives $7,150.

The resulting contribution is $8,800 minus $7,150, or $1,650. Before the attributed acquisition cost, contribution is $2,850. Dividing the final $1,650 by 100 orders gives $16.50 per order. Dividing by 200 shipped units gives $8.25 per shipped unit. These are different denominators applied to the same defined contribution.

The contribution rate on this model's net revenue is $1,650 divided by $8,800, or 18.75 percent. It should not be compared directly with a product gross margin calculated before fulfillment, refunds, payment fees, and acquisition costs. State the cost scope and revenue base beside the percentage whenever it is reported.

The SBA's break-even materials distinguish fixed costs and unit contribution in a simplifying model. That is useful context for keeping contribution separate from the full business result. A cohort contribution figure does not by itself establish that the operation covers all fixed costs or produces net profit. SBA break-even guidance.

Preserve the component bridge, not just the final answer. If next month's result changes, the business needs to know whether refunds, basket size, product cost, service mix, or acquisition spending moved. A final percentage without its inputs makes that explanation difficult and encourages comparisons between measures that do not share a definition.

Reconcile the same charges to the processor payout

Now examine cash movement using the same hypothetical charges. Begin with $9,600 of customer charges, subtract $800 of refunds and $360 of combined processing and refund-related fees, and subtract the $1,000 new reserve addition. The amount from these cohort events available for payout in the simplified scenario is $7,440.

Suppose the processor also releases $200 from an older cohort's reserve during the same bank period. The expected bank deposit becomes $7,640, assuming no other timing or adjustment differences. The older release is not new revenue from the current cohort. It is a cash movement from a previously restricted balance under the example's assumptions.

The equation is $9,600 minus $800 minus $360 minus $1,000 plus $200 equals $7,640. Compare that expected amount with the actual payout and bank transaction. If they differ, inspect the ledger for other events or timing. Do not force the difference into fees merely to make the reconciliation end at zero.

Keep the current cohort's remaining reserve balance visible. A later release changes cash availability, not the original sale amount. If the recoverability or accounting treatment changes, that requires a separate recorded assessment. The worksheet should not assume that every reserve will be released on a generic timetable.

This bridge depends on the example's simplifying assumption that the stated events fall into the same payout window. Real processor schedules can divide them across dates and currencies. Reconcile actual transaction identifiers and availability dates. A monthly sales total cannot automatically be matched to one bank deposit just because the amounts look close.

Explain why bank cash differs from contribution

The example's $4,000 inventory purchase was paid before the current bank-reporting window. During the current window, the business pays $700 of fulfillment, $800 of ordinary outbound delivery, $90 of replacement-related shipping, and $1,200 of acquisition spending. Those non-processor cash payments total $2,790.

Starting with the $7,640 bank deposit and subtracting $2,790 gives a $4,850 cash increase from the stated events in the current window. That is larger than the $1,650 contribution because the inventory cash was paid earlier, while the current payout also reflects a new reserve hold and an older reserve release.

The bridge is contribution of $1,650, plus the $4,000 inventory cost paid before the window, minus $1,000 newly unavailable in reserve, plus $200 released from an older reserve, equals $4,850. This arithmetic explains the difference under the stated assumptions. It does not make the extra cash free profit or available for owner distribution.

If the business subtracts the $360 processor fees again from the bank deposit, it double counts them because they were already deducted in the payout bridge. The same risk applies to refunds. A cash worksheet needs to distinguish gross customer receipts from net processor deposits and use a consistent method for the deductions between them.

Keep other business cash events outside this narrow bridge until they are explicitly added. Payroll, rent, taxes, debt payments, other inventory purchases, and owner transfers can change the actual bank balance. A reconciled cohort bridge is one component of the company's cash picture, not a substitute for a complete bank reconciliation and cash forecast.

Allocate order discounts before partial refunds

A partial refund should be connected to the actual charged amount and the applicable remedy, terms, and law. For management analysis, preserve how an order-level discount was allocated across lines. Otherwise, a returned line can be valued at its undiscounted price while the remaining order keeps the entire discount, distorting both revenue and the explanation to the customer.

Consider a hypothetical order with two merchandise lines priced at $70 and $30 before a $10 order discount. Under a proportional allocation used only for this example, the first line receives $7 of the discount and the second receives $3. Their net merchandise amounts are $63 and $27. The order's net merchandise total is $90.

If an applicable refund decision returns the first line's net merchandise charge, the model records $63 against that line. This illustration does not determine the customer's legal entitlement, shipping treatment, tax adjustment, or other remedies. It shows why the worksheet should retain the actual allocation instead of reconstructing it from a current list price.

Check whether the commerce platform already supplies line-level discount and refund allocations. Use those records where appropriate and reconcile them to the payment event. If a manual adjustment differs, record the reason and approver. A later analyst should be able to understand the difference without guessing which version of the promotion applied.

Use the same approach for bundles, account pricing, and shipping promotions when their structure requires it. The method should be explicit and reviewed for the business's actual arrangements. A consistent allocation helps explain the cohort, but consistency alone does not make an inappropriate method correct for every transaction or obligation.

Avoid double counting disputes and refunds

A disputed payment can create several ledger events: a provisional debit, a fee, a reversal, a refund, and a later final outcome. Those events need to be linked to the same underlying transaction. Summing every negative entry as a separate permanent loss can overstate the cost, while ignoring unresolved exposure can understate it.

For a hypothetical $100 transaction, suppose the processor records a $100 provisional dispute debit and later reverses that debit after resolution. The two principal movements net to zero in the simplified example, while any actual fee needs separate treatment. The cash path still matters because the business temporarily lacked the $100.

If a refund and dispute concern the same payment, investigate their relationship. Do not automatically deduct $100 of refund and another $100 of chargeback principal without checking what happened. The actual ledger and case status determine whether both amounts represent distinct outcomes, a temporary overlap, or an adjustment that will reverse.

Separate actual finalized outcomes from estimates and pending cases. The cohort report can show confirmed contribution and a sensitivity for unresolved exposure. That is more informative than presenting every open case as either a guaranteed recovery or a complete loss. Have the accounting treatment assessed appropriately rather than making it follow the report's preferred result.

Track the identifiers across support, commerce, and processor systems. A support ticket number may not match the dispute reference or original order identifier. Build a small cross-reference so the same event can be reconciled. This improves both the financial report and the team's ability to respond accurately through the processor's authorized process.

Compare cohort maturity before ranking performance

A recent cohort has had less time to accumulate refunds and disputes than an older one. Comparing their current margins without accounting for observation age can make the newer cohort look better simply because fewer outcomes have arrived. Record the number of days since the chosen cohort event alongside each snapshot.

Use comparable windows where practical. For example, compare cohorts at the same age after shipment when the question concerns early refund experience. If you also report a later mature view, keep it separate. Do not rewrite the earlier snapshot and lose the evidence of how the result developed.

Distinguish the count of affected orders from the amount of adjustments. Two cohorts can have the same refund dollars but different numbers of customer incidents and different basket sizes. Use the measure that corresponds to the decision, with its denominator. A refund rate by orders is not interchangeable with refunded dollars divided by revenue.

Avoid declaring a cohort permanently closed merely because a calendar month ended. Establish an operating convention for the report and continue to record later material events. The business can issue a stable monthly snapshot while retaining a method for explaining subsequent changes to earlier cohorts.

Use maturity analysis to improve estimates. If actual records show a consistent pattern of later adjustments, a planning model can incorporate an explicitly supported assumption. Do not borrow an industry refund percentage without evidence and present it as your store's experience. The model should distinguish observed history, forecast adjustments, and unresolved uncertainty.

Test the decision with transparent sensitivities

Start from the example's $1,650 contribution. If additional discounts reduce revenue by $500 with all included costs unchanged, contribution becomes $1,150. If additional refunds reduce revenue by $400, contribution becomes $1,250. If attributed acquisition spending rises by $300, contribution becomes $1,350. Each calculation changes one input only.

A combined hypothetical downside includes all three changes. Contribution becomes $1,650 minus $500 minus $400 minus $300, or $450. That is not a forecast of what will happen. It shows how much room the defined cohort has under the stated changes before considering other costs or responses.

Test costs that move with the decision. A larger basket can change picking, packaging, carrier charges, and payment fees. A discount can change product mix and refunds. Do not assume unchanged variable costs merely because it makes the sensitivity easy. Use a simplifying assumption to learn the relationship, then replace it with a more suitable model when the decision requires it.

Keep the cash sensitivity separate. A reserve increase can reduce the current payout without changing modeled contribution under the base treatment. A delayed supplier payment can improve one week's bank balance without lowering the cost assigned to the cohort. These timing changes deserve their own schedule rather than being mixed into the margin sensitivity.

The useful output identifies which uncertain input could change the decision. If a modest acquisition-cost correction eliminates most contribution, the next action may be to improve attribution or pause expansion. If payout timing creates the problem while contribution remains positive, the next action concerns liquidity and commitments. The worksheet should help the founder ask the right next question.

Reconcile a late warehouse credit without rewriting history

Suppose the hypothetical warehouse later confirms that $100 of the cohort's $700 fulfillment charge was billed in error and issues an applicable credit. Under a management treatment that reduces the attributed fulfillment cost, the updated cohort contribution rises from $1,650 to $1,750. The change is linked to the original invoice and cohort, with the date and evidence of the credit recorded.

That does not mean the current bank balance immediately increases by $100. The credit may offset a future invoice rather than produce a cash refund. The cash schedule needs the actual settlement mechanism and date. A report that adds the credit to both bank receipts and a reduction in the next payment would count the same benefit twice.

Retain the original cohort snapshot and show the adjustment bridge. The earlier $1,650 result was based on the information then included; the later $1,750 result incorporates a supported correction. A reviewer should be able to explain that difference without believing the underlying orders changed or that the business quietly improved its historical results.

If the warehouse merely acknowledges your dispute but has not approved the credit, label the amount disputed or estimated according to the report's convention. Do not count it as a confirmed recovery because the team expects to win. A sensitivity can show the effect if it is accepted, while the supported base result remains separate.

Apply the same discipline to supplier allowances, carrier claims, and fee corrections. Record the underlying event, amount, status, settlement form, and attributed cohort. Different recoveries may have different accounting treatment, so use the management model to preserve evidence and obtain appropriate review rather than forcing every credit into the same category.