How to auto-tag risky customers in Shopify
Every Shopify merchant has looked at a customer's return history and thought "this person has returned way too much." That insight usually stays in your head. It doesn't change how their next return gets handled.
Customer tagging closes that loop. Once risk tags exist on the customer record, you can route returns, exclude the abusers from promotions, warn support before they pick up the phone, and build segments for targeted communication.
This tutorial walks through setting up auto-tagging with RefundSentry, then shows five workflows that actually use those tags.
Prerequisites
You need RefundSentry installed, returns flowing through so the scoring engine has data, and admin access to create segments and Flows. Shopify Flow and Klaviyo (or similar) are optional but come up in the workflows below.
Understanding risk zones
RefundSentry scores every return on a 0 to 100 scale and classifies the customer into a zone.
| Risk zone | Score range | Tag applied | What it means |
|---|---|---|---|
| Low | 0 to 30 | refundsentry:low-risk | Normal return behavior |
| Medium | 31 to 65 | refundsentry:medium-risk | Some concerning signals |
| High | 66 to 100 | refundsentry:high-risk | Strong fraud or abuse indicators |
Tags update on every new return. A customer who drifts from low to high over a few months will see their tag move with them.
Step 1: enable auto-tagging
Auto-tagging is on by default. To check:
- Open RefundSentry from your Shopify admin
- Go to Settings, then Tagging
- Confirm "Auto-tag customers by risk zone" is enabled
- Optionally, customize the tag names if your existing systems need a specific format
Tag format
Default tags are refundsentry:low-risk, refundsentry:medium-risk, refundsentry:high-risk. The refundsentry: prefix avoids collisions with anything else you tag, and makes it easy to search for "refundsentry" and find every risk-tagged customer at once.
Tag behavior
Each new return triggers a tag evaluation. Only one risk tag applies at a time, and the new one replaces the old. A customer with no return activity stays where they are. A GDPR deletion request clears all RefundSentry tags from that customer.
Step 2: verify tags are applied
Find a customer in Shopify admin who recently submitted a return, open their profile, and check the tags section. You should see one of the refundsentry: tags.
If no tag appears, confirm the return was actually scored by RefundSentry (check the dashboard), make sure your Shopify scopes include write_customers, and look at the RefundSentry logs for tagging errors.
Step 3: create customer segments
Shopify segments filter by tag. Three segments, one per zone.
For the high-risk segment, go to Customers > Segments > Create segment, name it "High-risk returners (RefundSentry)", and use the filter customer_tags CONTAINS 'refundsentry:high-risk'. Save. Repeat for medium and low, swapping the tag string.
The low-risk segment is useful too, not just the high one. Think of it as your VIP candidate pool.
Segments update in real time as tags change.
Step 4: build workflows with tags
Workflow 1: exclude high-risk from free-return promotions
You offer free return shipping to email subscribers. Serial abusers love free-return promos. Drop them from the list.
In Klaviyo: Lists & Segments > Create Segment, add condition "Shopify tags contains refundsentry:high-risk", save as "High-risk returners". In your free-returns campaign, add exclusion: "Not in segment: High-risk returners".
In Shopify Email: when selecting audience, pick "Specific segment" and exclude the high-risk segment.
Workflow 2: route support tickets by risk level
Your support team gets a refund request and approves it without context. Fix that by surfacing the risk level on every ticket.
In Gorgias: create a rule on "When ticket created", condition "Customer tag contains refundsentry:high-risk", actions "Add tag VIP-Review-Required" and "Assign to Fraud Review queue".
In Zendesk: create a trigger on "When ticket is created", condition "Requester tags contain refundsentry:high-risk", actions "Add tag high-risk-customer", "Set priority to High", and optionally route to a specialized group.
Workflow 3: Slack notification on high-risk returns (Shopify Flow)
Trigger: Customer tags changed. Condition: Customer tags contains refundsentry:high-risk. Action: Send Slack message to #fraud-review.
Message template:
High-risk return alert
Customer just submitted a return and was tagged high-risk.
Customer ID: {{customer.id}}
View customer: {{customer.admin_url}}
Workflow 4: require photo proof on high-risk returns
High-risk customers file "item not as described" claims without evidence. Gate the return label behind a photo upload.
This needs integration with your return management system (Loop, AfterShip, a custom portal). The general flow: on return submission, read customer tags via the Shopify API. If the high-risk tag is present, require photo upload. Don't generate the label until photos land.
Workflow 5: tier exchange incentives by risk
You offer 10% bonus credit on exchanges. Abusers treat that as free upside on their arbitrage. Tier it instead:
| Risk level | Exchange bonus |
|---|---|
| Low-risk | 15% bonus credit |
| Medium-risk | 5% bonus credit |
| High-risk | No bonus, standard exchange |
Use Shopify Flow or Scripts to check the tag and apply a different discount code: EXCHANGE15 for low-risk, EXCHANGE5 for medium, nothing for high.
Step 5: advanced segment combinations
Risk tags get more useful when combined with other customer data.
High spend plus high risk. customer_tags CONTAINS 'refundsentry:high-risk' AND total_spent > 500. These customers spent real money but show fraud signals. Often worth a personal call before you block them.
New customer plus high risk. customer_tags CONTAINS 'refundsentry:high-risk' AND orders_count < 3. First-purchase fraud is a common pattern. This segment catches it early.
Repeat buyer plus medium risk. customer_tags CONTAINS 'refundsentry:medium-risk' AND orders_count > 5. Longtime customers who started showing new concerning behavior. Investigate before you escalate.
Step 6: monitor and iterate
In the RefundSentry dashboard, track the share of customers in each zone, the movement between zones over time, and the false-positive rate (legitimate customers tagged high-risk).
A healthy distribution usually looks like 70% to 85% low, 10% to 20% medium, 5% to 15% high. If your high-risk share is above 20%, your thresholds are likely too aggressive. Below 3% and you're probably missing fraud.
When you spot a false positive, note the customer and return. The tag will move naturally as their behavior does. If you see the same false-positive pattern repeatedly, contact support so the signal weights can be reviewed.
Common questions
Will customers see these tags? No. Shopify customer tags are admin-only. They don't appear on the storefront, order confirmations, or the customer's account.
What if a customer is in multiple segments? Shopify allows multiple tags, but RefundSentry only applies one risk tag at a time (the current level). Your other tags are untouched.
Can I override the risk tag manually? You can remove a RefundSentry tag, but it will be re-applied on the next return. For a sticky override, add a separate tag like trusted-customer and build your segments to exclude it.
How quickly do tags update? Within seconds of the return being scored. Shopify Flow adds its own delay, usually under a minute.
What if a customer improves? The tag reflects the current risk calculation. A high-risk customer who makes a run of clean purchases will see their score fall, and the tag follows.
What you've built
Automatic risk tagging on every scored return. Segments for each zone. Workflows that actually use the tags. Visibility into the risk distribution of your customer base.
That shifts fraud detection from a reactive manual review to an automated system that only asks for your attention on the 5% to 15% of returns that need it. The other 85% process with zero friction.
Next steps
Wire up Slack alerts on high-risk returns (workflow 3). Connect the tags to your helpdesk so every ticket arrives with context. Review the segments weekly for trend shifts. Export the high-risk cohort quarterly for a manual sanity check.
If you need help with a custom integration, the team is in the in-app chat.