Where The Money Leaks

Chargebacks, Refunds, and Void Audits

Chargebacks alone run 0.3 to 0.8 percent of card sales. Refunds and voids add another point of margin drift. Almost nobody runs the monthly audit that catches all three.

Kiyong (Eric) Lee·Operations Executive·August 2026·10 min read

The short version

Chargebacks, refunds, and voids leak roughly 1 to 1.5 percent of revenue when they are not audited. Combined they are one of the largest hidden margin drains in restaurant operations.
Benchmarks: chargebacks under 0.5 percent of card sales, refunds under 1.5 percent of gross, voids under 2 percent of gross.
Delivery channels chargeback at 3 to 5 times the rate of in-store. Model your rate by channel, not blended.
Fight chargebacks above $50 with a documentation package. Restaurants win 60 to 70 percent of properly-documented disputes.
Publish a monthly one-page audit ranking every unit. Sunlight fixes most of the behavior without any policy change.

Every operator knows about the big cost lines. Labor, food, rent. What they do not know, until they run the audit for the first time, is how much revenue never makes it to the deposit through the three quiet payment leaks: chargebacks disputed by cardholders, refunds issued at the unit, and voids inside the POS. Combined they typically run about 1 to 1.5 percent of gross revenue in unaudited operations. That is a margin line most operators would kill for if it appeared in a vendor negotiation, and it is available for free with a monthly workflow that takes about two hours per unit.

This is the audit I run for every multi-unit operator I engage with, and the numbers I use to know whether the current state is normal or a problem.

The three leaks, one at a time

Chargebacks

A chargeback happens when a cardholder disputes a transaction with their bank. The bank pulls the funds back from the merchant while it investigates. The merchant can either accept the loss or fight the dispute with documentation.

Chargebacks in restaurants come from four sources:

Friendly fraud: "I do not recognize this charge" from a guest who forgot they came in. Usually low dollar.
Delivery disputes: "Order was wrong, missing, or never delivered." This is the biggest source in operations with third-party delivery.
Service disputes: "Food was bad, service was terrible." Usually the guest complained onsite and did not get resolution.
True fraud: Stolen card used on a real transaction. Rare in dine-in, more common in phone-in with a card number.

Blended chargeback rates in restaurant:

Dine-in card sales:        0.10 to 0.25% chargeback rate
Online order (own site):   0.30 to 0.60%
Third-party delivery:      0.50 to 1.20%
Phone/manual entry:        0.60 to 1.50%

If a unit is running above these ranges, especially on delivery, there is a leak. If it is running below, either the audit is not happening or the packaging and handoff on delivery is exceptionally good.

Refunds

A refund is a completed transaction reversed at the unit. Different from a void, which happens before the transaction settles. Refunds are legitimate for wrong orders, guest recovery, and delivery mistakes. They are also one of the two most common ways for staff to move revenue out of the deposit.

Benchmark: refunds under 1.5 percent of gross sales in a well-run operation. Above 2.5 percent is a signal to investigate. Above 4 percent is almost never anything but a problem.

Voids

A void removes an item or a whole ticket from the check before it settles. Legitimate voids happen constantly: server hits the wrong button, guest changes their mind, kitchen misprints. Illegitimate voids are how revenue disappears: the guest pays in cash, the server voids the ticket, the cash goes in the pocket.

Benchmark: voids under 2 percent of gross sales in full-service, under 1 percent in QSR. Any unit consistently at 3 percent or higher gets a deep dive. Any unit at 5 percent has a problem, whether from training or from something worse.

Fig. 1 · Delivery and manual entry are structurally leaky. Model rates by channel, not blended.

The monthly audit workflow

The audit is a two-hour monthly exercise per unit. It happens on the second week of every month, on last month's closed data. The area director owns it, not the general manager. Independence matters here.

Step 1: Pull the three data sources

Chargeback report from the payment processor (Toast, Square, Stripe, whatever).
Refund report from the POS: date, time, amount, server, manager PIN, reason code, guest name.
Void report from the POS: same fields.

Format all three to a shared monthly audit spreadsheet, one tab per unit. The reporting is 20 minutes if the POS is set up right and 90 minutes if you have to reformat everything by hand. Fix the export once and the ongoing time falls.

Step 2: Compute the three ratios

Chargebacks / card sales   = chargeback ratio
Refund dollars / gross     = refund ratio
Void dollars / gross       = void ratio

Rank every unit on all three. Publish the ranking.

Step 3: Deep-dive the outliers

Any unit above the thresholds gets 30 minutes of extra work:

Concentration by server or manager PIN: Is 60 percent of the voids from one person? That is a signal.
Concentration by daypart: Are refunds happening 90 percent of the time in the last hour before close? That is a signal.
Reason codes: Is "guest dissatisfaction" being used 3x more often at one unit than others? Signal.
Payment type: Are voids concentrated on cash tickets? Very common signal.

Signals are not accusations. They are places to look. Most of the time the explanation is training or a specific broken process. Sometimes it is not. The audit is what surfaces the difference.

Step 4: Fight the chargebacks that are worth fighting

The math on chargeback disputes is not always favorable. Small chargebacks under $50 usually cost more to fight in labor than the recovery is worth. Above $50, and especially above $100, fight them.

What you need in the documentation package:

Original POS ticket with timestamp and items.
Payment authorization record from the processor.
For delivery: driver handoff proof, delivery photo if the platform captures it, order confirmation.
Any communication with the guest (text, email, call log).
A short cover letter explaining the transaction.

Restaurants win 60 to 70 percent of properly-documented disputes. On a $150 chargeback the recovery pays for the 20 minutes of documentation work. On a delivery chargeback of $18, do not bother.

Fight the disputes worth fighting. Accept the ones that are not. Documentation discipline compounds. Bad packaging documentation on delivery costs more than the food ever did.

What I have actually caught with this audit

Three examples across engagements, disguised on details:

Case 1: The Tuesday void pattern. Full-service unit, voids running at 3.8 percent. Ranking pulled up the concentration: 78 percent of voids were on Tuesday and Wednesday evenings, 91 percent were on cash tickets, and 84 percent were entered under one manager PIN. That manager was closing on Tuesday and Wednesday. Conversation with the operator produced a resignation and a separation. Voids dropped to 1.4 percent the following month with no other change.

Case 2: The delivery chargeback spiral. Fast-casual, chargebacks running at 2.1 percent of card sales. Deep dive showed almost 90 percent were from one third-party delivery platform, and almost all were "order incorrect" or "items missing." Root cause was a packaging change six weeks earlier that made the bag hard to seal, so drivers were opening it to check the order and sometimes putting items back wrong. Reverted the packaging change, added a tamper-evident sticker, chargebacks dropped to 0.6 percent within 60 days.

Case 3: The refund abuse from a server. Casual dining, refunds at 2.9 percent. Concentration audit showed one server accounted for 41 percent of all refunds despite covering about 12 percent of shifts. Interview revealed the server was refunding cash tickets and pocketing the money. Termination and PIN policy update. Refunds dropped to 1.1 percent the following month.

All three were invisible before the audit. All three were flagged by the audit within one monthly cycle. Total time to run the audit across the group in each case: under 3 hours. Combined margin recovery: several hundred thousand dollars a year.

Preventing the leak versus catching it

Catching is important. Preventing is better. Some standing policies that cut chargeback, refund, and void rates:

Manager approval required for voids over a threshold. Usually $25. Forces a second set of eyes.
Reason code required for every void and refund. Structured field, not free text. Makes pattern analysis possible.
Manager PIN unique per manager, rotated quarterly. Shared PINs destroy accountability.
Delivery packaging sealed with tamper-evident stickers. Cuts "order tampered with" disputes in half.
Photo of every delivery order at handoff. Some POS systems now do this automatically.
Standing operating procedure for guest complaint resolution onsite. Resolved complaints do not become chargebacks.

The chargeback ratio the processor will not tell you about

One number worth watching separately: your chargeback-to-transaction ratio versus your chargeback-to-dollar ratio. They are different, and they say different things.

Chargeback-to-transaction ratio is the count of chargebacks divided by the count of transactions. Card networks (Visa, Mastercard) monitor this ratio for merchant risk. If yours exceeds 1 percent, you risk being placed on a monitoring program with higher fees and increased scrutiny. Above 1.5 percent, you can lose your ability to process cards. This is existential and few operators watch it.

Chargeback-to-dollar ratio is the dollars charged back divided by dollars processed. This is the P&L number. It can look fine while the count ratio is a problem, because a lot of small delivery chargebacks add up in count without much dollar impact.

Track both monthly. Report both to the operator. If either is trending up, take it seriously.

The compliance overlay

Two compliance items that intersect with this audit and are worth knowing.

PCI compliance requires that manager PINs and cardholder data be handled with specific controls. If you are sharing manager PINs across shift managers to "make it easier at close," you are violating your merchant agreement and every audit above catches it. Fix it. Unique PINs, rotated quarterly.

Tip credit and cash handling rules vary by state. In some states, void abuse to skim cash triggers wage-and-hour risk beyond the direct dollar loss, because it can implicate tip pool math. If you find a pattern of cash-ticket voids under one server or manager, involve HR early, not late.

What I got wrong

Two things.

First, I underweighted the delivery channel for years. In one Hana Group unit inside a Whole Foods, delivery was 22 percent of sales but 61 percent of chargebacks. I was tracking a blended chargeback rate that looked "fine" at 0.6 percent. Broken out by channel, the delivery rate was 1.6 percent, which is high. The blended average was hiding the leak. Now I always break the ratio by channel.

Second, I let voids drift for a full quarter because the general manager had a good explanation ("we had two new POS trainings"). The explanation was true and also insufficient. Voids stayed above benchmark for three months. In hindsight the deep dive should have happened month one, explanation or not, because signals do not go away because you have a story.

The point

Chargebacks, refunds, and voids are the least glamorous audit in restaurant operations. They are also one of the highest-yield. Two hours a month, per unit, protects roughly a full point of margin. Publish the ranking, deep-dive the outliers, fight the disputes worth fighting, prevent what you can with standing policy.

The audit is not about accusing anyone. It is about looking, on a schedule, at a set of numbers that quietly drift when nobody looks. Sunlight fixes most of it. What sunlight does not fix, the audit surfaces early enough that you still have options.

Not running a monthly payment audit?

I work with multi-unit operators on payment leakage, POS discipline, and chargeback dispute workflows. If your chargeback and void rates have never been benchmarked by channel, that is where the money is.

Get in touch

Questions people ask

What is a chargeback in a restaurant context?

A chargeback is when a guest disputes a credit card transaction with their card issuer and the payment gets reversed. In restaurants they come from three main sources: friendly fraud (I did not authorize this), delivery order disputes (order was wrong or missing), and true fraud (stolen card). Chargebacks typically run 0.3 to 0.8 percent of card sales in restaurants and higher for delivery-heavy operations.

How often should you audit refunds and voids?

Monthly at minimum. Weekly for high-volume operations. Voids and refunds are the two most common cover mechanisms for both accidental and intentional revenue leakage. Left unwatched they drift up quarter over quarter. The right benchmark is voids under 2 percent of sales, refunds under 1.5 percent, and chargebacks under 0.5 percent of card sales.

What is a normal void percentage?

Under 2 percent of gross sales in a well-run full-service restaurant, under 1 percent in a well-run QSR. Higher than that is a signal, not a proof of anything. Higher voids come from three places: legitimate order corrections, poor training on the POS, and intentional revenue skimming. The audit separates the three by looking at concentration, reason codes, and pattern.

Are delivery chargebacks worth fighting?

Depends on the amount. Under $50 the paperwork cost usually exceeds the recovery. Above $50 and especially above $100, absolutely fight them with a documentation package: original order, prep photo if you have one, delivery driver handoff proof, and any communication with the guest. Restaurants win 60 to 70 percent of well-documented delivery chargeback disputes.

What are common refund abuse patterns?

The three patterns to watch. First: refunds concentrated on one server or one manager PIN. Second: refunds without a matching guest name or table number when your POS captures those. Third: refunds happening late in the shift, right before close, on cash tickets. Each is a signal to look closer, not a conclusion.

How much margin does void and refund leakage actually cost?

Combined, unaudited void and refund drift typically runs 0.8 to 1.5 percent of sales. On a $2M unit that is $16k to $30k a year in revenue that never reached the deposit. In a five-unit group it is $80k to $150k. It is not a single moment of theft. It is small amounts done constantly by a small number of people who have learned nobody checks.

Who should own the monthly audit?

The area director or regional operations lead. Not the general manager, because you want independence from the operator being measured. The general manager sees the report and follows up on their own unit. The area director runs the process, ranks the units, and escalates outliers. Set it as a standing item on the monthly regional operating review.