# Return Velocity Limits: Slowing Serial Returns Without Banning Good Buyers

Source: https://returnwarden.com/blog/2026-11-03-return-velocity-limits-account-rules.html

November 3, 2026

Per order return rules have a blind spot shaped exactly like your worst returner. Each individual return is unremarkable: in window, tags on, reason plausible. It is the eleventh one in ninety days that tells the story, and no rule that reads returns one at a time will ever see it.

Velocity limits read the story instead of the page. Count returns, return value, or keep it claims per account over a rolling window, and let the count, not a support agent's memory, decide when an account gets a warning, a friction step, or a restriction.

## Set tiers on your own data first

Pull twelve months of returns by account before choosing a single number. In most apparel and general merchandise data the honest bulk of customers returns rarely, a fit driven minority returns often and keeps plenty, and a thin tail returns nearly everything. Your tiers should sit in the gaps of your distribution, not at numbers copied from a conference talk.

A workable shape: a soft tier that triggers a polite notice, a middle tier that removes free labels and shortens the window for that account, and a hard tier that requires review before further orders ship. Appriss and NRF returns research is useful context for what the tail looks like industry wide, but your own histogram decides where your lines go. See also [Refund to Original Payment Method: Why the Rule Exists and Where It Bends](https://returnwarden.com/blog/2026-11-02-refund-original-payment-method-guide.md).

## Count the right thing

Raw return count punishes your best customers, because a household that orders forty times a year and returns ten is a profit center wearing a suspicious costume. Count return rate alongside count: returns divided by orders, weighted by value kept. The account to restrict is high count, high rate, low kept value. The account to leave alone is high count, moderate rate, high kept value.

Exempt the signals that mean your store failed, not the customer: damaged arrivals, wrong items shipped, size chart errors you have already acknowledged. Velocity systems that count your mistakes against the customer produce righteous, screenshot ready complaints, and they are correct.

## Warn before you restrict, every time

A restriction that arrives without warning reads as an accusation, and some share of the accounts it hits will be edge cases you got wrong: a stylist buying for clients, a new parent in a sizing storm, a reseller you might actually want as a wholesale lead. The warning is not softness. It is the step that separates a policy from a trap, and it gives the genuinely abusive account a documented chance to stop, which strengthens every action after it.

Keep an appeal path with a human at the end of it. Velocity systems are counting machines, and counting machines deserve an override with a name attached.

## Frequently asked questions

### What is a return velocity limit?

It is a rule that counts an account's returns over a rolling period, such as 30 or 90 days, and applies graduated responses as the count grows: a notice, added friction like paid labels, then restriction or review. It catches patterns that per order rules cannot see.

### How many returns is too many?

It depends on the account's order volume and kept value, which is why velocity should be measured as rate plus count. An account returning 80 percent of forty orders is a different case from one returning 20 percent of ten. Set tiers from your own twelve month distribution.

### Should high value customers be exempt from velocity rules?

Not exempt, but measured on kept value. A customer who returns often and keeps thousands of dollars of merchandise a year is profitable behavior, and the system should treat them accordingly. Velocity tiers keyed to kept value handle this without a manual VIP list.

### Do I have to warn customers before restricting returns?

Legally it depends on your terms and jurisdiction, but operationally you should. A warning creates a record that the account knew, resolves the edge cases you misclassified, and makes later restrictions defensible in disputes and reviews. Surprise bans are where velocity programs earn their bad press.