Most merchants treat chargeback responses like a lottery ticket — submit whatever’s handy, hope the issuer sides with you. Stripe just published the receipts on why that approach loses. After analyzing evidence packets from one million disputes over a 16-week window, the company found that the difference between winning and losing a “product not received” claim often comes down to a handful of specific data points that most businesses aren’t submitting.
“Product not received” is the most common non-fraud dispute category on Stripe, and the winning formula isn’t a mystery anymore. It’s a checklist. The teams still losing these disputes are losing them because their shipping data, usage telemetry, and refund records live in different systems that don’t talk to each other.
Why Delivery Confirmation Is the Single Biggest Lever
According to Stripe’s analysis, disputes for physical goods that included delivery confirmation saw a 27 percentage point higher win rate than those without. Layer in a GPS delivery map showing where the carrier scanned the package, and win rates jumped another 15 percentage points. Add a recipient signature and you gain two more. Stacked together, that’s a 44 percentage point advantage over merchants submitting nothing.
The bottleneck isn’t legal expertise — it’s data plumbing. Dispute response stops being an art; it becomes an engineering problem. Stripe explicitly notes that shipping data and dispute workflows typically live in separate systems, making it hard to match a specific dispute to the right order at scale. If you’re running an ecommerce operation where fulfillment sits in ShipStation, orders sit in Shopify, and disputes sit in Stripe, your ops team is manually reconstructing evidence packets one at a time — and losing the ones that fall through the cracks. Clean integrations and custom API development solves this before disputes even arrive.
The prediction: within 18 months, expect processors beyond Stripe to start bundling automated carrier-data pulls into their dispute tooling. The manual evidence packet will look as dated as manually reconciling bank statements.
The Timing Trap Most Merchants Fall Into
Stripe’s data revealed something counterintuitive: when you submit matters almost as much as what you submit. Disputes with evidence submitted after delivery was confirmed had a 27 percentage point higher win rate than disputes with no confirmation. But disputes submitted while the package was still in transit? Just a two percentage point lift. Barely better than nothing.
This reframes the entire dispute response workflow. Most operations teams treat disputes like support tickets — respond fast, close the loop, move on. Stripe’s numbers suggest the opposite discipline: wait. Cardholders often file “product not received” claims before a delayed package actually arrives, and since merchants generally have 20+ days to respond, patience pays. If you must submit early, Stripe recommends including documentation of the delivery timeframe the customer agreed to at checkout.
Practical example: if you run a DTC apparel brand and your average delivery window is 5-7 days, build a rule in your dispute workflow that flags any claim filed within that window and holds submission until the tracking API confirms delivery. That rule alone can shift your win rate.
The take: any modern SaaS platform handling billing at scale needs a dispute queue that’s aware of shipping status, not just dispute deadlines.
What Digital Goods Merchants Actually Need to Prove
The digital-goods playbook has its own numbers. Stripe found that disputes including digital activity and usage logs — JSON telemetry showing a user streamed, downloaded, or accessed the specific product they purchased — had a 10 percentage point higher win rate than disputes without. Service documentation like provisioning records added an eight percentage point lift.
The underlying pattern is specificity. Proving a customer had access to your platform is weaker than proving they consumed the exact SKU they’re now disputing. If you’re running a streaming service, don’t just show the login timestamp — show the play events on the specific title. If you’re a SaaS platform, don’t just show account provisioning — show feature-level usage tied to the disputed invoice.
That changes how you should instrument your product. Analytics events that were built for growth dashboards now double as legal evidence. Product teams that haven’t logged granular consumption data at the SKU or feature level are going to find themselves rebuilding telemetry systems specifically to survive chargebacks.
The Refund Paradox That Costs Digital Merchants the Most
The most striking number in Stripe’s analysis: for digital-goods businesses, disputes that included evidence of a full refund issued through Stripe had a 63 percentage point higher win rate than those without. Refunds issued through other channels — store credit, manual bank transfers — only lifted win rates by six percentage points.
The mechanism is simple. When a refund runs through the card network, the issuing bank can verify it. When you refund via store credit or an off-network method, there’s no record the issuer can check, so from their perspective, no refund happened. This is where merchants routinely shoot themselves in the foot: they think they’ve made the customer whole, but they’ve done it in a way that’s invisible to the party deciding the dispute.
For teams building any kind of checkout — especially those investing in custom payment gateway integration — the lesson is that refund flows are as strategically important as capture flows. A refund that isn’t network-verifiable is a refund that will cost you twice.
Prediction: expect the next wave of dispute-automation tools to actively discourage or block off-network refunds when a dispute risk score crosses a threshold. The days of the “just give them store credit” customer service reflex are numbered for anyone selling digital goods.
FAQ
Q: What counts as strong evidence for a “product not received” dispute? A: For physical goods, Stripe’s data points to delivery confirmation, a GPS delivery map showing the scan location, and a recipient signature — submitted after delivery is confirmed. For digital goods, specific consumption telemetry (streams, downloads, feature usage on the exact product purchased) outperforms generic service documentation.
Q: Can a customer still dispute a charge after I’ve refunded them? A: Yes. Stripe notes that cardholders can initiate disputes even after refunds are processed, often because the refund and dispute happen in parallel, or because the issuing bank doesn’t check refund status before filing. The key is that the refund must be issued through your payment processor so the issuer can verify it on the card network.
Q: Should I always submit dispute evidence immediately? A: No. Stripe’s analysis shows evidence submitted after delivery is confirmed carries a 27 percentage point advantage, while evidence submitted while the package is still in transit adds only two points. If your dispute deadline allows it — most merchants have 20+ days — waiting for delivery confirmation is often the smarter play.
Key Takeaways
- Audit whether your dispute workflow can automatically pull delivery status, GPS scan data, and signatures from your carrier APIs — if the answer is “our ops team does that manually,” you’re leaving win rate on the table.
- Build a submission-timing rule into your chargeback process that holds evidence until delivery is confirmed, rather than defaulting to “respond immediately.”
- Instrument granular product-level usage telemetry now, before you need it as dispute evidence — retrofitting analytics into legal proof is far harder than logging it correctly upfront.
- Kill or contain off-network refund flows for digital products: if the issuer can’t verify a refund on the card network, it barely counts in dispute adjudication.
- Merchants sticking with fragmented dispute processes will see win rates diverge sharply from those adopting AI-assembled evidence packets like Stripe’s Smart Disputes — the gap will widen through 2027 as more processors ship similar automation.