Stop return abuse
before the refund.

Every return is scored for abuse risk the moment it's submitted, before any automation runs. Returner links the same shopper across emails, phones, and addresses, so wardrobing rings and serial returners get caught even when they spread activity across accounts.

Every return scored before it's approved.

The instant a return is submitted, a shared risk engine scores it 0–100 and assigns a level. Scoring runs before your automation rules, so a rule can hold or decline based on the score. Because the dashboard and the rules engine use the same engine, they never disagree.

  • 0–100 risk score computed at submit time
  • Low / medium / high risk level
  • Flags for wardrobing, return velocity, and value concentration
  • Flags for address mismatch and serial bracketing
  • Deterministic engine shared by the dashboard and rules
Dashboard / Return
Return RET-4471
High risk
86 / 100 risk score
SignalWardrobing (deadline)
SignalReturn velocity
SignalAddress mismatch

One shopper, many accounts — linked.

Returner builds a tenant-local identity cluster that links the same person across different emails, phone numbers, and addresses. At scoring time it counts how many accounts a shopper is connected to and aggregates their whole history, so spreading returns across throwaway accounts no longer hides the pattern.

  • Links emails, phones, and addresses into one identity
  • Cluster-wide 90-day return count
  • 7-day return velocity across the cluster
  • Declined count and account age
  • Counts linked accounts at scoring time
Dashboard / Identity
Cluster #C-118
4 accounts
Stockholm · +46 70 555 01
Same address · new phone
Same phone · new address
Returns (90d)21
Velocity (7d)6
Declined3

Your thresholds. Your call.

Risk controls stay in your hands. Flip the master toggle on, set the high-risk threshold, and decide what happens next: auto-flag high-risk returns for review, or — when you're ready — auto-block them outright. Risk conditions are also available inside the automation rules, and you can flag individual customers by hand.

  • Master risk toggle
  • High-risk threshold, 0–100 (default 70)
  • Auto-flag high-risk returns for manual review
  • Optional auto-block of high-risk returns (off by default)
  • Use risk conditions inside automation rules
  • Manual per-customer flags
Dashboard / Risk settings
Risk controls
Risk scoring
High-risk threshold70
Auto-flag for review
Auto-block high risk

Signals that add up.

Wardrobing detection

Returns filed right at the deadline raise a wardrobing flag — the classic pattern of wearing an item, then sending it back just in time.

Return velocity

Many returns in a short window push the score up. The 7-day cluster velocity catches bursts even when they're split across linked accounts.

Serial bracketing

Ordering many variants of the same item to keep one and return the rest is flagged as serial bracketing before the refund clears.

Address mismatch

When shipping and billing addresses don't line up, the return picks up an address-mismatch signal that feeds the overall score.

Cluster history

Identity clustering aggregates 90-day returns, declined count, and account age across every linked account, so spread-out abuse still surfaces.

Block or review

Once a return crosses your threshold, auto-flag it for manual review or auto-block it. Risk conditions also drop straight into your automation rules.

Protect every margin.

See how Returner scores a return and surfaces the signals — in real time, before the refund.

Try the demo