111 Retailers' Manuals, 63 Errors, and a Penalty That Mises the Real Cost by 38% to 75%

Sixty-three distinct ways to get an invoice short-paid, from the supplier compliance manuals of 111 US retailers. Fifty-four of them are not about price, not about quantity, not about product quality, and not about delivery timing.

The list is Craig, DeHoratius and Klabjan, "Execution quality and chargeback penalties in retail supply chains," Journal of Operations Management 67:31-48, 2020. It is the only peer-reviewed measurement of what retailers actually charge, and it is six years old. The suppliers losing money to chargebacks are mostly losing it to execution, packaging, carton labels, barcodes, advance ship notices, ticketing, pallet type, pack-by-store, and the industry files all of it under logistics.

The researchers call that second group execution quality. One caveat first: these are penalties as documented at that paper's vintage, averaged across 111 retailers: not Walmart's numbers, not Kroger's, and not checked by us against any retailer's current guide.

Execution quality is most of the money

Craig, DeHoratius, Jiang and Klabjan, Journal of Operations Management 38:25-40, 2015, put a distribution centre under a microscope, and the 2020 study reports two numbers from it: execution-quality chargebacks were 60% of the dollar value of all chargebacks that retailer issued, and fulfilment errors were associated with nearly 7% of its purchase orders.

Sixty percent of the money, out of one retailer's books. That is the entire published evidence that the largest slice of your deductions has nothing to do with price or delivery, one company, one distribution centre, and nobody has repeated the measurement.

The penalty does not match the harm

Craig, DeHoratius and Klabjan modelled 16,380 inventory scenarios to price an execution-quality error properly: what it actually costs the supply chain once downstream damage is counted, and set that against what the manual charges. The penalty deviates from the true cost by 38% to 75% on average, depending on which of the three penalty types applies.

The average hides a number you can use. Across fixed, per-unit and value-proportion penalties together, the mean absolute gap is $11.34. Under the fixed-charge-only policy most retailers actually run, it is $22.62. The policy nearly everyone uses is the one that gets the amount most wrong.

The structural result comes from the same paper. Using chargebacks at all raises total supply-chain cost 18.95% above a coordinated optimum. Having no chargeback mechanism at all raises it 38.49%.

So the answer to "are chargebacks justified" is neither yes nor no. They are directionally right and quantitatively wrong, an argument for disputing the large ones carefully, not for disputing everything and expecting a different number.

What those 111 manuals actually charge

Per-unit charges typically run $0.12-$0.97. Rarer value-proportion charges average 8.33% of purchase-order cost for carton labels (standard deviation 4.51) and 5.00% for UPCs.

The detail that ends the argument before it starts: only 8 of the 111 retailers vary the size of a penalty by supplier or product at all. The other 103 charge the same flat number whoever you are. The identical $152.50 carton-label charge lands on a seasonal novelty and a core line, and the manual offers no argument with which to tell them apart.

Where the four document families sit

The same paper published the taxonomy itself, by distinct error type, with the execution-quality subset in brackets.

Add the rows a document packet is built from, packing lists and advance ship notices (7), carton packing and labeling (6), pack-by-store (5), pre-ticketing (4) and invoicing (3), and that is 25 of the 63 error types, and 24 of the 54 execution-quality types. More than a third of the error surface by type count, before a single dollar is counted. The advance ship notice is the largest of those rows and the one most often wrong: the message the retailer checks against what physically arrived.

The benchmarks older than the paper

Craig, DeHoratius and Klabjan state that chargebacks reduce supplier revenue by 2% to 10% annually. Their source for that band is not their data: it is a 2003 trade article, Zieger 2003. It is the only revenue-magnitude figure in the peer-reviewed literature: a 2003 number inside a 2020 paper.

The error rates are older still, all cited in the 2020 study's literature review: 13% of inbound shipments carried an inaccurate advance ship notice, from a survey of 42 retailers by Retail VCF, Auburn University and Compliance Networks, ASN Exposed, 2010; 5% of orders were subject to execution-quality errors across 45 pharmaceutical distribution centres, Speh at WERC 2001; and shipping discrepancies hit 12% of supplier shipments on average in automotive just-in-time, Srinivasan, Kekre and Mukhopadhyay, Management Science 40(10), 1994.

The most recent of those three is from 2010 and none has been replaced. Treat any error rate quoted to you as current as a citation to this paragraph.

The number nobody can source

There is no primary measurement anywhere of how much of what you are deducted is wrong. What circulates instead is two estimates from executives who sell deduction-recovery services.

Akash Raju, CEO of Glimpse, says about 90% of invoice deductions are legitimate. Jon Allen, CEO of Woodridge Group, says many CPG brands find 10% to 20% of deductions invalid but untouchable, and separately that about 20% may be preventable. Two vendor executives, no published methodology, and a disagreement of a factor of two. If you need a recovery rate for a board paper, the honest state of the record is that there isn't one.

The gap between them may not be a real conflict: a deduction can be legitimate and still recoverable if you failed to file the evidence for it. Invalid and unchallenged are different problems.

One company's numbers, and they are not a benchmark

In January 2021 a toy supplier, Think3Fold, subscribed to SupplyPike's Deductions Navigator with more than 1,000 outstanding Walmart deductions. Its own case study reports that under the manual process it used before, 90.73% of disputes were cancelled and 2.71% approved; after, the company says it won 92.91% of the claims it disputed and recovered 99.57% of the money deducted, at 2 to 27 days to resolution.

One supplier, one retailer, one tool, no control group, published by the vendor. It is not a recovery benchmark and should not be quoted as one. Worth knowing who published it: SupplyPike now redirects to SPS Commerce's Revenue Recovery page, so this is the incumbent's own case study, not a challenger's. That matters twice, the result is the best kind available, and it is still one customer. The same vendor's 2022 release says an Amazon supplier spends about 30 minutes of manual labour researching and disputing each individual deduction, which is the cost on your side of the same trade and also a vendor statement rather than a survey.

The number you can defend is the one computed from your own remittance advices, and the arithmetic is in the recovery rate is the part nobody publishes.

The part that decides it

Two findings sit at the end of this: the penalty misses the real cost of the error by 38% to 75%, and 103 of 111 retailers charge a flat rate that does not care what happened, who you are, or what the goods cost them.

Every window in the dispute-window table is shorter than the habit producing the deduction. The mean gap under a fixed-charge policy is $22.62 per chargeback, and Walmart's OTIF clock runs about 30 days from the invoice.