Pre-ship order scoring: stop fraud before you fulfill
Most fraud detection happens too late. You ship the order, the customer returns it (or files a chargeback), and you eat the loss. By then the product may be worn, damaged, or never coming back.
What if you could catch the risk before the box left the warehouse?
That's what pre-ship order scoring does. Instead of waiting for a return to flag abuse, you score every order at creation and hold the risky ones for review before fulfillment starts.
Why post-return detection isn't enough
Traditional return fraud detection is reactive. Customer orders. You ship. Customer files a return or chargeback. You detect the fraud. You've already paid for shipping, burned warehouse time, and often won't see the product again in sellable condition.
For high-value orders, bracketing setups, and known abusers, this cycle is expensive. The product comes back used. Return shipping eats margin. Inspection costs add up.
The better approach: detect the risk before you ship.
How pre-ship scoring works
RefundSentry scores every new order asynchronously at checkout using 10 pre-ship signals.
Bracketing detection. A customer orders the same product in 3+ sizes or colors. They're almost certainly planning to keep one and return the rest. One of the most common and most expensive abuse patterns in fashion and apparel.
Risky first orders. A brand-new customer placing a high-value first order with expedited shipping and a different billing vs. shipping address is a classic fraud profile. The model weights multiple first-order risk factors together rather than any one in isolation.
High-abuse SKU patterns. Some products attract disproportionate abuse. If a SKU has a historically high return rate and a customer orders it alongside other high-return items, the score goes up.
Velocity ordering. Three orders from the same customer within 24 hours is unusual, especially combined with other signals.
Prior return history. Customers with a 30%+ return rate on previous orders get flagged. Past behavior predicts future behavior.
Quantity anomalies. Five units of a single item when your store average is 1 or 2 is a signal worth evaluating in context.
The fulfillment hold
Scoring alone isn't enough. You need to act on it.
When an order crosses your configured risk threshold, RefundSentry places a hold using Shopify's native Order Hold API. That pauses fulfillment without canceling the order or blocking payment, notifies your team via Slack, webhook, or email, gives reviewers the full signal breakdown for context, and enables a one-click release when the order checks out.
The key insight: holding a risky order for 30 minutes of review is dramatically cheaper than processing a fraudulent return two weeks later.
What this means by category
Fashion and apparel. Bracketing is the single biggest problem. Pre-ship scoring catches size and color gaming at checkout. Instead of shipping 3 sizes and eating 2 returns, you can review the order first.
Consumer electronics. High-value first orders with expedited shipping are a classic fraud signature. A 30-minute hold on an $800 laptop is a rounding error compared to the cost of a fraudulent return with a tampered device coming back.
Growing stores. During peak volume your team can't manually review every order. Pre-ship scoring automates triage so human review focuses where it matters.
Zero checkout impact
A fair concern: won't this slow down checkout?
No. RefundSentry runs entirely through Shopify webhooks. Scoring happens asynchronously after checkout completes. The customer sees zero latency. The hold is placed on the fulfillment side, not the payment side. The customer's payment is captured normally.
Getting started
Pre-ship order scoring is included in RefundSentry. It works alongside return scoring, customer tagging, and chargeback prediction, so you get fraud coverage from order placement through return resolution. See pricing.
Setup is automatic. Install RefundSentry, scoring begins on your next checkout, configure your hold thresholds in settings, and risky orders start getting held for review.
No migration. No workflow changes. An intelligence layer on top of the fulfillment process you already use.
Related reading
- Pre-purchase to post-return: full-lifecycle fraud intelligence. How pre-ship scores feed into return scoring and chargeback prediction for connected intelligence.
- Chargeback prevention on Shopify. The other side of pre-ship prevention: predicting chargebacks before they happen.
- How we process thousands of Shopify webhooks without dropping one. The async architecture that makes real-time scoring possible without slowing checkout.