Every organisation designs the process of selling with care, and very few design the process of receiving goods back. Returns are handled case by case on judgement, and that is why they are slow for the customer and expensive for the operation at the same time.
What follows: why returns are handled worse, what automates well, what still requires judgement, and how value is recovered.
The asymmetry is striking. An order has clear rules, tracking and an owner at every stage. A return of the same product usually begins with a phone call and proceeds according to whoever picks it up.
The cost is paid twice: in the customer experience, which meets friction precisely when they are already dissatisfied, and in the operation, which moves goods without knowing what it will do with them.
Why they are handled worse
The first cause is that nobody owns the whole process. Customer service takes the request, logistics moves the product, finance issues the credit note and the warehouse decides its destination, without any of them seeing the whole.
The second is that the information arrives incomplete. Without knowing why a product is coming back, the decision about what to do with it gets made once it is already on the dock, which is the most expensive moment to decide it.
The third is the absence of written rules. When each case is settled on judgement, the result depends on who handles it, and the organisation can predict neither its cost nor its timescale.
What automates well
The structured request. A form capturing reason, product condition and evidence from the outset allows the destination to be decided before the goods move. It is the highest-impact change and the simplest.
Rule-based approval. Within policy — valid window, eligible product, covered reason — authorisation can be immediate. Reserving human review for what falls outside the rule speeds up the majority and concentrates judgement where it adds value.
Visible tracking. Letting the customer see where their return stands removes most follow-up enquiries, which are the ones that saturate service without adding anything.
Suggested disposition. Based on reason and declared condition, the system can propose the destination — back into stock, refurbishment, warranty with the manufacturer, disposal — and leave confirmation to the owner.
What requires judgement
The returns policy is a commercial decision. How long a window, on what conditions and with what exceptions defines the balance between customer experience and cost, and it is not settled with technology.
Cases outside policy need an explicit path. An important customer outside the window is a real situation; if the system does not contemplate it, it gets settled outside and the record stops reflecting what happens.
And abuse detection calls for balance. Unusual return patterns can be identified, and they are best treated as a signal for human review rather than an automatic block, because the cost of a false positive falls on a legitimate customer.
How value is recovered
Returned goods lose value with every day that passes without a decision. A product waiting weeks on a dock for someone to determine its destination depreciates without anyone recording it.
That is why speed of disposition matters as much as the cost of transport: deciding quickly preserves more value than negotiating a better freight rate.
And the reason for return is product information, not only logistics information. When it is captured in a structured way and reaches whoever designs or buys, it stops being a cost and becomes the most honest signal the organisation has about what it sells.
How to sequence the project
Start by measuring: how many returns, for what reason, how long they take and what final destination they had. That table usually shows that a few reasons account for most of the cases.
Automate the most frequent path first, the one clearly within policy. It covers the bulk of the volume with the simplest logic.
Then bring in the less frequent reasons, and leave until last those that require negotiation, which probably should continue to be handled by a person with context.
What has to exist first
A current written returns policy, an agreed catalogue of reasons, and an owner for the end-to-end process. The third is usually the one missing, and the one that most determines the result.
It is also worth agreeing the indicator: the time from customer request to resolved return, and the percentage of goods that go back into stock. The first measures experience; the second, value recovery.
The figure that changes the conversation
Most organisations know their return rate and very few know its composition. They are two different things and only the second supports action.
A high rate concentrated in a sizing or description reason points at the product listing. Concentrated in damage, it points at packaging or transport. Spread across many reasons, it points at the sales process.
Capturing the reason in a structured way costs little and turns a logistics cost into a commercial diagnosis. It is the project’s most valuable by-product, and it is almost always discovered afterwards.
Which part delivers a result fastest?
The structured request with rule-based approval. It allows the destination to be decided before the goods move, which is where most of the avoidable cost is generated.
Should returns be approved automatically?
Within policy, yes. Reserving human review for what falls outside the rule speeds up most cases and concentrates judgement where it genuinely adds value.
How is abuse detected without penalising a legitimate customer?
By treating an unusual pattern as a signal for human review, not as an automatic block. The cost of a false positive is paid by a customer who did nothing wrong.
Which indicator is worth following?
The time from request to resolution, and the percentage of goods returned to stock. The first measures experience; the second, value recovery.
