SKECHERS Charged $29,496K of Chargebacks to Expense in FY2024. A Reserve Is a Forecast, and Yours Should Be One

SKECHERS charged $29,496K of retailer chargebacks to expense in fiscal 2024 and closed that year with a $39,937K chargeback allowance on its balance sheet. That is an accrual set before the fact, not a write-off discovered afterwards.

A chargeback here means a reduction in what the retailer pays you for goods received, not a customer reversing a card payment. Two public registrants disclose a chargeback reserve, and both treat the retailer deduction as a budgeted cost rather than an exception.

In its Form 10-K for fiscal 2024, filed 2025-02-28, SKECHERS carries the line in Schedule II, the schedule where a registrant discloses its valuation and qualifying accounts. The rollforward is three years of money.

For FY2024 alone: opening balance $33,303K, deductions and write-offs $(22,862)K, closing balance $39,937K at 2024-12-31.

Hanesbrands carries the same item in the same place in its FY2024 10-K, filed 2025-02-14, as a custom element in the filing's machine-readable data, chargebacks and other deductions, accrued as a discrete allowance alongside credit losses.

Those are two companies, not an industry, and neither tells you what to budget. Read them for what they are: evidence of accounting treatment. Both companies estimate the charge before it happens, and correct themselves at year end.

For scale, the same filing discloses $345,488K of domestic accounts receivable and $711,686K international, before allowances. Dividing the closing $39,937K allowance by that $1,057,174K is our arithmetic, not a disclosed ratio, but it gives the order of magnitude: high single digits of receivables, not noise.

A reserve is a forecast with a name on it

A remittance line is a fact that already happened. A reserve is a number you set in advance, which makes it a forecast, and the gap between the forecast and what the retailer actually took is a forecasting error somebody owns.

When the chargeback exists only as a line on an invoice, nobody owns it. The cost surfaces once, in a period that is already closed, and never becomes a variance against a number anyone committed to. A reserve converts an event into a budget line, and a budget line has a variance. That is the whole finance argument for accruing it.

What follows is mechanical: accrue monthly, compare what you accrued to what was deducted, and let the difference land somewhere a person reviews it. The largest single input is the recovery rate, the share of deducted dollars credited back inside the window, computed from your own remittance advices and covered in the recovery arithmetic. It is different at every company.

Set the reserve from last year's deductions and never true it up, and you have not made an estimate. You have copied a number.

The second deduction stream is the one that breaks the estimate

From 2022-08-01, Walmart's Collect suppliers carry a second charge that is not called a chargeback: a Collect Pickup charge, calculated monthly, department-specific, as a percent of the cost of goods received across Walmart U.S., plus a fuel surcharge, per Talk Business & Politics on 2022-07-06.

Two details matter to a reserve. The base is goods received, not sold, so it moves with shipment timing rather than with sales. And no rate for it has been published since. A model that prices Walmart exposure as 3% on on-time-in-full misses is pricing the older, smaller of two streams and calling it the whole account. What penalties cost against a purchase order is measured in 111 retailers' manuals, 63 errors.

The penalty rate is not the exposure rate

3% is a per-event rate. The average bill is much smaller.

Russell Zuppo, VP of consulting services at Uber Freight, told Forbes on 2024-02-01 that Walmart suppliers in that platform's customer base were paying 0.16% of cost of goods sold in fines to Walmart, and that across many of them the fines run well in excess of $1 million a year. One executive, one freight platform, one customer base, but the only observed figure in the public record, and an order of magnitude below the headline rate.

The wider band is softer. Dallas Counts, COO of Vendormint, told SupplyChainBrain on 2026-04-15 that depending on the vertical these deductions are generally in the 3-8% range, and that when chargebacks exceed a contractual allowance there is an auditing process usually covering the previous 12 months, often resetting the allowance in the next contract.

Those are different populations: all deductions against OTIF fines only, and cannot be quoted interchangeably. Size the reserve from your own history, not from a published percentage. Note the twelve-month reach: a reserve never trued up is an audit finding waiting for the next contract negotiation, where someone else resets it.

The same 3%, on three different bases

This is the number that decides whether your reserve is close or off by six.

Grocery Dive, on 2017-11-28 citing the Wall Street Journal, reported Walmart's fine as 3% of the value of each late or incomplete shipment. Supply Chain Dive, on 2019-03-08, reported 3% of the cost of goods sold, the figure Walmart spokesperson Michelle Malashock put on the record. 8th & Walton, on 2024-05-16, described 3% of cost of goods on all cases that did not meet the goal, working an example in which only the 15 short cases of a 100-case order were fined.

Same number, three bases, and the gap between the top and the bottom of that column is roughly six. A reserve built on the wrong base is wrong by that factor, and no amount of process quality will correct it. Walmart's own portal language is the only authority on which base applies, and it is not public.

The retailer you are accruing against publishes nothing either: a full-text search of Walmart's filings on EDGAR for OTIF returns zero hits, as does the same search for on-time-in-full. No Walmart public filing discloses an OTIF deduction rate.

The waiver nobody models

Walmart calculates and displays OTIF charges monthly but bills them quarterly, and waives any monthly OTIF penalty under $1,000, per 8th & Walton on 2024-05-16: a 3PL's explainer rather than Walmart's own document, and the only published statement of the mechanic.

Build the reserve line by line from remittance advices and you over-predict: you are adding monthly penalties that never reach an invoice. Build it from what was billed after the waiver and you under-predict, a month that ran $950 leaves no trace, and the trend underneath the floor is the one that tells you the account is deteriorating. The waiver hides the signal at exactly the volume where the signal is most useful.

What the record cannot tell you

The most-quoted number in this category is not a measurement. Jon Allen, CEO of the Woodridge Group, told Talk Business & Politics on 2026-01-04 that many CPG brands find 10-20% of deductions invalid, that extra trucks, promos and online orders create deductions that will hurt holiday profits by up to 15% of gross sales, and that 20% of deductions may be preventable. All three come from one executive at one consultancy in one interview, with no primary measurement behind any of them. None is a planning input.

The 15% figure is also scoped to the holiday quarter, which is the point worth keeping: a single quarter's recovery rate is not a run rate, and a reserve annualised from Q4 is inflated by construction. There is also no verified recovery rate anywhere in the public record, which is why the ratio you compute from your own remittance advices, through the detect-match-classify-file loop in the dispute workflow, is the only version that means anything to a budget.

Which returns you to the disclosure that started this. SKECHERS closed fiscal 2024 with a $39,937K chargeback allowance, filed 2025-02-28, after $29,496K went through expense and $(22,862)K came back out. That true-up is the whole argument. A number nobody trues up is not an estimate; it is a habit, and the reserve is the only place a chargeback becomes something you own.