Returns management software, and the returns it never sees

What returns management software is

Returns management software is the system an e-commerce retailer runs a return on: a portal that takes the request, a rules engine that approves or declines it, label generation and carrier routing for the trip back, and the refund, exchange or store credit at the end. Products sold under this name split two ways: a customer-facing portal, and a warehouse receiving module for the other half of the same job. This software is very good at running a return you have already lost, and structurally incapable of preventing the one you were about to cause. I once looked at a single bundle line: 157 units sold, 156 returned. Nobody changed their mind. A bundling app was splitting orders into separate lines for the warehouse, so what arrived was never what was bought. The portal processed all 156 perfectly.

Returns are a modelled supply-chain process, not a support function: ASCM's SCOR Digital Standard carries Return as a top-level process, split into Source Return and Deliver Return. The shape is defined, and it's wider than the portal.

What the software actually does, end to end

Seven steps, whatever the product is called. Each closes on the way it fails, the half the feature lists leave out.

1. Request: a portal takes the return without the customer writing to anyone, so the first record of the problem is a form.
2. Eligibility: a rules engine checks window, product class, condition and order value, then issues the RMA authorization knowing only the policy, never what happened to the parcel.
3. Reason capture: a structured reason code is attached, the only returns data most brands collect, self-reported by the person who wants the refund.
4. Label and routing: a prepaid label is generated and a carrier chosen, at a cost the brand absorbs and never charges back to the order that caused it.
5. Transit: the parcel travels back, the longest leg and the least watched: nobody monitors a return the way they monitor an outbound shipment.
6. Receipt and disposition: the item is scanned in, inspected and sent to restock, refurbish, liquidate or scrap. Resale value drains in the gap before that decision.
7. Resolution: a refund, exchange or store credit is issued, the only step the customer experiences as fast or slow.

Notice the assumption underneath the list: every step is downstream of a decision the customer has already made.

What returns cost, and which part the software touches

Total US retail returns are projected at $849.9 billion in 2025, which retailers put at 15.8% of annual sales, against 16.9% the year before. Online carries more of it: NRF estimates 19.3% of online sales will be returned.

Almost none of that money is the software. The cost is the freight back, the labor to receive and inspect, the markdown or write-off on what can't be resold at full price, and the outbound shipping you have already spent. Returns management software makes the request and the refund cheaper to administer. It doesn't make the parcel cheaper to move or the unit easier to sell again. The labor half has a published price: BLS puts the annual mean wage for a customer service representative at $46,590, and that's the person every item on the list above gets routed to when it's hand-fixed.

How to evaluate it, and what a demo will not show you

Returns processing is an operations number and returns experience is a customer number, owned by different people. A demo shows you the second. These five questions get at the first.

• Policy engine depth: whether rules can key off product class, customer history, order value and condition, or only off a date window. Ask to see a rule that declines something.
• Exchange and credit steering: whether the flow can hold the revenue instead of refunding it. Ask what share of returns it converts, and how that share is measured.
• Disposition and inventory writeback: what happens after the box is opened, and how fast a resellable unit is back on sale. Ask how long the average unit sits between scan-in and available-to-promise.
• Integration reality: whether it writes back to the store, the warehouse and your order management software, or only reads from them. Ask what happens when a writeback fails, not whether it exists.
• What it does with the reason code: whether reason data reaches merchandising and fulfillment, or only a dashboard nobody owns. Ask who reads it weekly.

Every one of those is a question about a return you already have.

The refund clock the law already started

Refund speed gets written up as a nicety. In the US part of it is a deadline. Under the FTC's Mail, Internet, or Telephone Order Merchandise Rule, a seller that can't ship inside its advertised window, or inside 30 days where it advertised none, must offer the buyer a delay to consent to or a cancellation with a prompt refund. Prompt is defined: seven working days from the date the right to a refund vests, or one billing cycle where the seller is the creditor.

Behind that sits the card clock. Regulation Z gives a cardholder 60 days from the statement to assert a billing error, and the creditor 30 days to acknowledge it and two complete billing cycles, never later than 90 days, to resolve it. A refund you don't issue comes back as a dispute with its own deadline.

States add their own. California Civil Code 1723 makes a retailer that doesn't conspicuously post a restrictive refund policy liable to the buyer for the purchase price if the goods come back within 30 days. The law treats a refund as an obligation with a clock on it, and most brands find out they missed it when the chargeback lands.

The returns the software never sees

Split what comes back into three piles. One is a return by choice: wrong size, changed mind, didn't suit, and a portal is built for exactly those. The other two are failures wearing a return's clothing, and they cost the same money and land in the same return-rate metric, which is why the metric never improves. The guides to cutting your return rate are about sizing charts, photography and policy levers. They rarely name fulfillment accuracy or a delivery exception as a cause.

• Return by choice: the item arrived as promised and the customer changed their mind. The portal sees it first, and this is what the software was designed for.
• Return caused by a break: the wrong unit, a short shipment, damage, a delivery that missed the window. The customer sees it first, and by then the order is already lost.
• Return nobody requested: refused and undeliverable parcels that come back with no portal involved. The USPS Inspector General counted 11.6 million undeliverable Parcel Select packages scanned return-to-sender in FY2024, about $138 million in postage.

Those last two piles sit upstream of the reverse logistics load entirely, and they're the ones I would go after first. One in five orders hits an operational break after checkout, and brands hand those breaks to a helpdesk, a system for replying about problems rather than resolving them. Replying and fixing the order are different jobs, and only one keeps the promise. The second is proactive e-commerce operations.

What sits before the returns stack

Not another portal, but a layer that holds every order against what was promised at checkout. Detect the break from signals your systems already emit: the pick that came up short, the collection scan that never happened, the delivery window that has passed. Decide against the promise rather than against a ticket queue. Act: reship, re-route, refund early, or tell the customer before they ask.

The edges matter, so here they are. Keeyu isn't returns management software. We're not a returns portal, not a WMS, not a carrier, not an OMS and not a helpdesk, and we replace none of them. A brand that needs a returns portal should buy one, and the criteria above are how. We sit one leg earlier, on the break that turns a good order into a return.

If your returns stack only starts working once the order is already lost, the expensive problem is upstream of it. Every order is a promise. Keeyu keeps the promise: we detect the operational break, decide what should happen and act on it, usually before the customer knows anything went wrong, so fewer good orders ever become returns. Book a Keeyu demo and bring the returns you can't explain.

• ASCM. SCOR Digital Standard 2025, front matter. Return as a top-level supply chain process, split into Source Return and Deliver Return.
• National Retail Federation. Returns projected at $849.9 billion in 2025. Retailers estimate 15.8% of annual sales will be returned in 2025, totaling $849.9 billion, against 16.9% in 2024.
• National Retail Federation. 2025 Retail Returns Landscape. An estimated 19.3% of online sales returned in 2025, and 9% of all returns found to be fraudulent.
• US Bureau of Labor Statistics. Occupational Employment and Wage Statistics, May 2025. Annual mean wage of $46,590.
• Federal Trade Commission. Business Guide to the Mail Order Rule. The shipping window, and the seller's obligation to offer a delay or a prompt refund when it cannot be met.
• Cornell Legal Information Institute. 16 CFR 435.2. Thirty days where no shipping time was stated, and the consent-to-delay or cancel-and-refund option.
• Cornell Legal Information Institute. 16 CFR 435.1. Prompt refund defined as seven working days from the date the buyer's right to a refund vests, or one billing cycle where the seller is the creditor.
• Cornell Legal Information Institute. 12 CFR 1026.13, Regulation Z billing error resolution. Sixty days for the consumer to give notice, thirty days for the creditor to acknowledge, and two complete billing cycles or ninety days to resolve.
• California Legislative Information. California Civil Code section 1723. A restrictive refund policy must be conspicuously displayed, and a retailer that does not display one is liable to the buyer for the purchase price on or before the thirtieth day after purchase.
• US Postal Service Office of Inspector General. Protecting Revenue for Returned Parcel Select Packages, 25-053-R25. More than 11.6 million undeliverable Parcel Select packages scanned as return-to-sender in FY2024, about $138 million in postage.
• US Postal Service Office of Inspector General. Sending It Back, RISC-RI-24-005. The growth of e-commerce returns as a reverse logistics load on the delivery network.