Digital Product Business Risks: Chargebacks, Refunds and Payment Processors

Selling a course, a template, a piece of software or an ebook feels like the safest business on the internet. No stock to hold. No boxes to ship. No warehouse rent. The money lands in your account the second someone clicks buy.

Your payment processor sees it differently. To them, a digital product is one of the riskiest things you can sell. There is nothing physical to track, nothing to return, and nothing to prove the buyer got what they paid for. When a customer disputes the charge, the bank usually believes the customer first and asks questions later.

That is the real risk in a digital product business. It is not one lost sale. It is what happens when disputes pile up: held payouts, a reserve taken from your balance, and in the worst case, a closed account that follows you to the next processor. This guide explains how chargebacks, refunds and payment processors actually work for digital sellers, what each outcome costs in real dollars, and the habits that keep your account standing.

Why processors treat digital products as high risk

A processor pays you before it knows whether the sale will stick. If the buyer disputes the charge weeks later, the card network pulls the money back from the processor, and the processor takes it back from you. With a physical product there is usually a tracking number, a delivery signature and a returned parcel to argue with. With a digital product there is often only a login timestamp, and banks rarely find that persuasive.

Three things stack the risk. Delivery is instant, so an unhappy buyer already has everything before they complain. Value is subjective, so “the course was not what I expected” is hard to argue against. And a lot of digital products sell through paid ads, where impulse buys and buyer’s remorse are common.

The numbers back that up. Industry data puts chargeback rates for digital goods at around 1.8% globally, against roughly 0.5% for physical products. Mastercard research with Datos Insights, updated in June 2026, puts the global cost of chargebacks at $36.9 billion in 2026, with 286 million chargebacks expected in the year.

Chargeback vs refund: what each one actually costs you

A refund and a chargeback both end with money going back to the customer, but they are completely different events. A refund is a return you control. A chargeback is a forced reversal you do not control, and it comes with fees and a mark against your account.

Here is what each outcome costs on a $99 course sale through Stripe in the US:

Outcome	What you lose	Counts toward your ratio?
Refund before any dispute	The $99 sale (plus the original processing fee, which the processor keeps)	No
Refund through a pre-dispute alert	The $99 sale plus the alert fee	Usually no
Chargeback you accept	$99 plus a $15 dispute fee	Yes
Chargeback you fight and win	Just the $15 dispute fee (the $15 counter fee comes back)	Yes — networks count disputes received, not lost
Chargeback you fight and lose	$99 plus $30 in fees ($15 dispute fee and $15 counter fee)	Yes

Notice the second row of the table. Even a chargeback you win still counts against your ratio, because the card networks count disputes when they arrive, not when they are decided. Winning helps your wallet. It does not help your ratio.

The sale amount is only part of the damage. Industry research shows that once you add chargeback fees, lost transaction costs, the hours spent gathering evidence and any penalties, a single chargeback can cost a merchant two to three times the original transaction value. A $99 dispute can quietly cost you $200 to $300.

The numbers that get your account flagged

Card networks run monitoring programs, and they do not warn you politely before they act. Two thresholds matter most for digital sellers:

Visa VAMP: 1.5% and 1,500 disputes

Visa’s Acquirer Monitoring Program, known as VAMP, flags a merchant as excessive at a 1.5% chargeback ratio once there are at least 1,500 counted fraud reports and disputes in a month. The count includes fraud reports, not just formal chargebacks, which is why the number can climb faster than sellers expect.

Mastercard ECM: 100 chargebacks and 1.5%

Mastercard’s Excessive Chargeback Merchant program starts at 100 chargebacks and a 1.5% ratio in a month. At a 1.5% rate, 100 chargebacks means roughly 6,700 sales in the month, so smaller sellers hit the count side of that test sooner than they think.

Your own target: stay under 0.5%

Processors set their own limits too, and they usually act earlier than the card networks. A sensible working target is to keep your dispute ratio under 0.5% of transactions. Work it out once a month: divide the number of disputes by the number of sales, then multiply by 100. If you sold 1,000 copies of a template and 15 buyers disputed, that is 1.5% — already at the line where the networks start watching.

What happens when you cross the line

Processors rarely close an account over one bad week. They react to trends, and the escalation usually follows the same ladder:

1. A warning and a closer look

Your processor’s risk team starts reviewing the account and may ask for refund and chargeback figures, your sales pages and your terms. Answer fast and honestly. Silence reads as guilt.

2. A reserve

The processor starts holding back a percentage of your payouts as a cushion against future disputes. Reserves are normally reviewed after a clean track record, but “clean” means months, not days.

3. Payout holds

Money stops flowing out while the processor waits to see whether refunds and chargebacks follow a spike in volume. This is the step that kills cash flow, because the bills keep coming while the balance sits frozen.

4. Account closure

If the ratio stays above the scheme limits for consecutive months, or the business model does not match what the account was opened for, the processor can close the account. A closed account can also put you on the Mastercard MATCH list, a shared list of terminated merchants that makes opening a new account much harder.

Stripe vs PayPal: the rules that decide your case

The same disputed order can face two very different sets of rules depending on how the buyer paid. Stripe disputes are card chargebacks decided by the card issuer; Stripe passes your evidence on. PayPal disputes start inside PayPal and can end with PayPal’s own decision, although a card chargeback on a PayPal payment still goes to the issuer. Stripe’s own disputes documentation explains the process step by step.

	Stripe	PayPal
Who decides	The card issuer	PayPal decides claims after escalation; card chargebacks go to the issuer
Buyer’s window	Usually 120 days for most Visa and Mastercard reason codes	180 days from the payment date to open a dispute
Fees (US)	$15 per dispute, plus $15 to contest, returned if you win	$15 standard dispute fee ($30 at the high-volume tier); $20 chargeback fee when the buyer goes straight to their bank
How you respond	One submission per dispute; files up to 4.5 MB combined	Up to 20 days to settle with the buyer before it can be escalated to a claim
Best digital proof	IP address or system log showing the download or use; for Visa fraud claims, two earlier qualifying orders	A system record of the date you sent access and the email or IP it went to, or proof the buyer opened it

PayPal’s Seller Protection can cover intangible goods, but only when PayPal marks the payment eligible on the Transaction Details page. For digital goods, the US terms accept a system record as proof: it must show the date you sent the item and the email or IP address it went to. Save that record with every order, because you cannot reconstruct it after the fact. On Stripe, check the fee before you contest a low-price order: contesting a $9 template and losing costs $30 in fees on top of the $9 sale.

Seven triggers that put a new account under review

Processors do not react to one bad week. They react to patterns their monitoring flags automatically. These are the most common triggers for digital-product merchants:

1. A sudden spike in volume

A launch that does 80% of your yearly revenue in five days looks exactly like a merchant collecting money before disappearing. Tell your processor before a launch, not after.

2. A rising refund rate

Even if chargebacks stay low, a refund rate that keeps climbing tells the processor something is wrong with the offer.

3. “Product not received” or “not as described” codes

These two reason codes hurt digital sellers most, because they point straight at delivery and description problems you can actually fix.

4. Recurring billing buyers did not clearly agree to

Subscriptions are the most disputed digital product of all. If the renewal terms were buried, the chargeback is almost guaranteed.

5. High-ticket offers sold on a single call

Coaching and mastermind programs sold for thousands of dollars in one conversation draw extra scrutiny, especially when the buyer later says they felt pressured.

6. Income-claim marketing

“Make $10k a month” in your ads is a red flag for underwriters. It attracts exactly the buyers most likely to dispute when the money does not arrive.

7. A mismatch with your application

If the account was opened for “software” and the processor later finds high-ticket coaching with income claims, that gap alone can trigger a review. Describe what you actually sell, in plain words, when you apply.

Evidence that actually wins a digital dispute

There is no tracking number for a download, so you have to build the paper trail yourself. The evidence that works comes in three separate chains, and you need all three:

Purchase consent: what the buyer agreed to

Product, price, currency, renewal terms, the terms accepted at checkout, and the transaction ID. A clickwrap setup, where the buyer has to check a box or click a button to confirm they agree to your terms before paying, creates a documented record of consent. A browsewrap link buried in the footer does not.

Delivery: that access was made available

Entitlement creation, the delivery email, activation or download timestamps. Write the product description the way you would write a meta description: specific, honest, and free of hype. And keep the exact version of the sales page the buyer saw, because “the product did not match the description” is argued against the page as it stood on the day of purchase.

Usage: that the buyer actually used it

Logins, lesson access, license activations, download events tied to the purchase. Download logs prove delivery. They do not prove the cardholder approved the payment, so pair them with the consent record above.

Match the evidence to the reason code. A non-receipt claim needs delivery and access records. A cancellation claim needs the date the buyer cancelled and the terms they agreed to. An unauthorised-payment claim needs login and 3D Secure records, and a login alone does not prove the cardholder approved the payment.

Two card-network tools are built for exactly this situation. Visa Compelling Evidence 3.0 lets you fight Visa fraud claims (reason code 10.4) with the buyer’s own past orders: two earlier orders from the same card, 120 to 365 days older than the disputed one, with at least two matching data points such as IP address, device ID or account ID, one of which must be the IP or device. Keep those records for at least 365 days. Mastercard First-Party Trust helps you contest Mastercard fraud claims (reason code 4837) when the real buyer made the purchase. And 3D Secure 2, where the buyer authenticates the payment, usually shifts fraud-claim liability to the card issuer. Chargeflow’s guide to preventing chargebacks on digital goods walks through each of these tools with checklists for the evidence to keep. One last habit: keep the records per order, tied together by one order ID or buyer email, and keep them longer than feels necessary — a record you deleted in month two is a dispute you cannot defend in month four.

A refund policy that turns chargebacks into refunds

The cheapest dispute is the one that never becomes a dispute. A generous, clearly written refund policy that is easy to find and easy to use turns many would-be chargebacks into refunds, and a refund never counts against your ratio.

Write it in plain words and put it where buyers can actually find it: on the sales page, at checkout, and inside your Terms of Purchase. Cover whether refunds are available and under what conditions, the time window for asking, exactly how to ask (an email address or a form, not a maze), and whether you offer alternatives like store credit. Then answer support messages fast and refund early when a customer is clearly unhappy — the sale you give back today is cheaper than the chargeback that arrives next month. Do not rely on a “no refunds” policy to protect you. It does not prevent chargebacks, and it cannot override consumer rights or card network rules.

If you sell to buyers in the EU, there is one more rule to know. Buyers usually get 14 days to cancel an online purchase. For digital content, that right ends once delivery starts, but only if the buyer agreed first. Under Article 16(m) of the EU Consumer Rights Directive, the buyer must give prior express consent and confirm they understand they lose the right to cancel. In practice that means an unticked consent box at checkout for instant access, with the exact wording, the box state and a timestamp saved with the order, and the consent repeated in the order confirmation email.

The launch trap: how a good week freezes your payouts

Launch-based businesses get hit hardest by processor risk. A course that does most of its yearly revenue in a five-day window looks to a risk model exactly like a merchant collecting money before disappearing. Processors may hold funds until they can see whether refunds and chargebacks follow.

The fix is boring and it works: tell your processor before the launch. Send a short message with the launch dates, the expected volume and average order value, the product and price, and the refund policy buyers will see. If you have not validated demand for the product yet, start with free keyword research to check that real people are actually searching for what you are about to sell. A processor that was warned is far less likely to freeze a balance than one that was surprised.

After the launch, watch the numbers daily for the first two weeks, not monthly. A spike in “product not received” disputes in the first 48 hours usually means a delivery email is landing in spam or a license key is failing to issue. Fix the delivery problem and the disputes stop on their own. And buy your own product once a quarter: check the receipt, the access email, the download link and the cancellation path. A paid order with no working access is a dispute waiting to happen.

When a merchant of record makes sense

Stripe and PayPal are payment service providers. You are the merchant, you hold the tax obligations, and you are responsible for chargebacks and compliance. A merchant of record is a different model: the provider becomes the official reseller of your product, takes on the tax obligations, and handles chargebacks and compliance as part of the service. A detailed Stripe vs PayPal comparison shows where the two processors leave gaps that a merchant of record is built to fill.

That trade is worth considering when you sell internationally, when subscriptions and tax rules are eating your time, or when you have already had a pause or a closure and need a more stable setup. The fees look higher than a payment processor’s headline rate, because they include the tax, compliance and dispute work you would otherwise do yourself or pay someone else to do. It is not the right move for everyone: a low-ticket template shop with a low dispute rate may fit standard processing perfectly well. But if you run launches or sell high-ticket programs, a dedicated merchant account underwritten for your model, or a merchant-of-record setup, is usually more stable than an aggregator account that was onboarded with light checks and is monitored hard afterwards. Whatever route you choose, keep at least one backup payment option ready — a pause on one processor should not stop your business, and you cannot set a backup up in the middle of a freeze.

Your first 48 hours after a chargeback lands

When the notification arrives, do not panic and do not ignore it. Work through these steps in order:

1. Read the reason code and the actual complaint

Fraud, non-receipt, cancellation and “not as described” need different evidence. A dispute claim alone does not prove intent, so read what the buyer actually said before you decide how to answer.

2. Check whether a refund is still possible

If no dispute has formally opened yet, a pre-dispute alert lets you refund before it becomes a counted chargeback. If a dispute is already open, check with your provider before sending a separate refund, or the buyer may be credited twice. A refund sent after a dispute opens does not remove it from the count.

3. Decide: fight or accept

Fight when you have the delivery and usage evidence that matches the reason code, and when the amount justifies the fee. Accept when the buyer has a real complaint, when the evidence is thin, or when the fees would cost more than the sale.

4. Submit once, and submit well

Stripe allows one submission per dispute. Keep the file short, relevant and readable, with a brief timeline and each exhibit labelled. Do not send the reviewer a login link to your app.

5. Save everything, then fix the cause

Keep the dispute, the evidence and the outcome together. Then ask what caused it: a confusing sales page, a broken delivery email, a renewal reminder that never went out. Disputes are feedback with a fee attached.

Digital products are still a wonderful business. The margins are real and the freedom is real. The risk is real too, but it is manageable. Keep your refund rate low, your records complete, your ratio under half a percent, check that ratio once a month, and keep your processor informed, and the payment side of the business stays boring — which is exactly how you want it.

This article is general information for digital sellers, not legal advice. Your processor agreement always governs, and a lawyer can advise on the terms for your own business.