Treat Shopify fraud analysis as an indicator, not proof. Open the order, read each indicator, then verify payment and status in your admin before you fulfil. Build a short checklist: payment captured, AVS and CVV results, billing and shipping match, order value against your normal range, and customer history. If signals conflict, hold and ask for verification through a neutral payment or identity check. Never label a customer a fraudster in notes or messages.
What Shopify fraud analysis actually shows you
Shopify fraud analysis is a review tool inside your Shopify admin. It flags orders that may be fraudulent so you can look at them before fulfilment. It works with online credit card orders, and it can show indicators and a recommendation depending on your plan and payment setup (Shopify Help: Reviewing orders with fraud analysis).
Two limits matter for your checklist. First, some orders do not receive a fraud recommendation at all, including test orders, free orders, orders paid in full with a gift card, Point of Sale orders, B2B orders, and subscription renewal orders. Second, if you use a third-party payment processor, check with them about fraud analysis limitations. The tool is a starting point, not a verdict.
The decision to reverse funds sits with the card-issuing bank, not Shopify, and Shopify does not cover charge reversals from banks. That is why your review should focus on evidence you can gather before shipping, not on predicting a bank outcome.
Build the checklist: what to verify before fulfilment
Use the same sequence for every flagged order so your decision is consistent. The goal is proportionate verification, not interrogation.
- Confirm payment status. Check that the payment is captured or authorised as expected in the order timeline. An order that is still pending is not ready to ship.
- Read each fraud indicator separately. Note which specific signals appear: billing and shipping address mismatch, AVS or CVV result, unusual order value, multiple attempts, or a new customer with a large basket. Do not collapse them into one score.
- Compare billing and shipping details. A mismatch is common for gifts and students. Look at whether the mismatch is plausible given the products and destination.
- Check order value against your normal range. A first order far above your typical basket is a reason to look closer, not a reason to accuse.
- Review customer history. Repeat customers with prior fulfilled orders carry different weight from a brand-new account with no history.
- Check for card testing patterns. Several small orders in a short window can suggest card testing. Shopify's guidance on responding to card testing and reducing chargeback risk is worth reading if you see this pattern.
- Decide: fulfil, hold, or cancel. Record the reason in your own internal notes using neutral language such as "verification pending" rather than "fraudster".
A worked example (hypothetical)
Imagine an order for £480 from a new customer, shipping to a different address from the billing address, with a CVV mismatch and a first-time email domain. No single item proves fraud. Together they justify a hold and a neutral verification request. By contrast, a £60 repeat order with a billing and shipping mismatch but a clean AVS result and two prior fulfilled orders may be fine to ship. These figures are illustrative only, not benchmarks.
How do you verify without accusing the customer?
Keep the message transactional. Ask the customer to confirm the delivery address or to complete a verification step through your payment provider, rather than asking them to prove they are not a criminal. A short email such as "We are confirming a few details before dispatch" is enough. If you use Shopify Flow, you can automate parts of this review.
Never write accusatory notes in customer-visible fields. Internal notes should stay factual and neutral. If you cancel, use a standard reason and refund according to your policy and the payment processor's rules.
What mistakes make fraud review worse?
- Treating a high-risk label as proof. It is an indicator. Verify before acting.
- Shipping because the payment captured. Capture does not mean the cardholder authorised the charge.
- Ignoring orders with no recommendation. Test, gift card, POS, B2B, and subscription renewal orders can still be disputed.
- Using one rule for every order. A mismatch on a gift order is not the same as a mismatch on a high-value electronics order.
- Leaving no record. Write down what you checked and why you decided to fulfil or hold. This helps if a chargeback arrives later.
Where this fits with your other Shopify work
Fraud review is an operational habit, not a one-off setup. Keep your product and collection information accurate so customers recognise what they ordered, and keep your site structure stable so order confirmations and tracking links do not break. If you are planning a redesign or a seasonal push, the same discipline applies: prepare the details before you launch. Our SEO migration checklist before redesigning your website covers the page-level checks that keep customers and search engines oriented, and the Shopify SEO guide to building a reusable seasonal collection shows how to keep a returning collection stable. Both reduce the kind of confusion that makes order verification harder.
Quick reference: your pre-ship decision table
| Signal | Look closer if | Often fine if |
|---|---|---|
| Billing/shipping mismatch | High value, new customer | Repeat customer, low value, gift note |
| AVS/CVV result | Mismatch or unavailable | Match, or known bank limitation |
| Order value | Far above your normal range | Within your usual range |
| Customer history | No prior orders | Prior fulfilled orders |
| Pattern | Multiple small orders quickly | Single order, normal timing |
Use the table to structure your thinking, not to replace it. When in doubt, hold and verify.
Follow-up questions
Does Shopify fraud analysis work for every order?
No. Some orders do not receive a fraud recommendation, including test orders, free orders, orders paid in full with a gift card, Point of Sale orders, B2B orders, and subscription renewal orders. You can still review those manually using the same checklist.
Can I test fraud analysis before relying on it?
Yes. You can simulate a high, medium, or low risk order on a test order. Test orders do not receive a real fraud recommendation, and the order total must be greater than zero. If you use another payment gateway for the test, your store must be on the Grow plan or higher.
