DTC technology × growth systems — built to compound.
DTCUP / FIELD NOTES / OPERATE

Build a Returns Process That Resolves the Customer’s Problem and Recovers Stock

Compare your 3PL with a specialist returns center. Connect customer resolution, inventory recovery, full costs and product learning in one usable process.

Published Sep 15, 2026Updated Sep 20, 202614 min read
Jack Hu
Jack Hu

7 years of experience building and operating brand websites, specializing in website development, paid advertising and SEO. Built a brand website from launch to US$10 million in annual sales.

The Short Answer

Define the work each returned product needs before choosing its destination. Compare the existing 3PL and a specialist on demonstrated capability, unresolved cases, complete route cost, customer visibility and the movement of recovered stock. More recovered units do not automatically mean a better result after extra work and transfers.

Track the customer resolution, physical item and financial outcome separately until they reconcile. Test exceptions before switching routes, preserve ownership of returns already in transit, and turn original customer explanations and inspection evidence into specific product, packaging or website improvements.

Key takeaways

  • Choose the return destination around the work the actual product requires.
  • Compare the complete recovery route, including transfers and internal effort.
  • Receipt, saleable stock and a confirmed refund are different events.
  • Use returns and support evidence to improve the next product and customer promise.

A return can look finished in one system and remain unfinished everywhere else. The customer has posted the parcel, support has closed the conversation, the warehouse has received a box, and finance still cannot explain whether the refund went through. Meanwhile, an item that could be sold again sits on a shelf with no decision attached to it.

I would judge a returns process by three outcomes together: the customer receives the appropriate resolution, the business knows what happened to the goods and money, and someone uses the evidence to improve the next sale. A polished returns portal helps only if the work behind it reaches those outcomes.

That is also how I would choose between returning products to the existing fulfillment provider and using a specialist returns center. Start with the work each product needs. Then compare the whole route, including inspection, recovery, transfers, customer communication and exceptions. The cheapest return label or inspection fee cannot answer that decision on its own. The surrounding warehousing and fulfillment guides connect this decision to storage, order handling and the wider distribution setup.

Follow several real returns from the request to the final outcome

Before changing providers, collect a small set of completed and still-open cases. Include an uncomplicated return, an exchange, an item with missing parts, an unexplained parcel and a customer who contacted support more than once. Use actual examples from your business, with customer details limited to the people who need them.

For each case, write a short timeline. When did the customer ask? What were they told to do? When did the parcel move, arrive and get inspected? Who decided what happened next? When did the customer receive a confirmed resolution? When did inventory and finance reflect it?

The gaps usually matter more than the number of screens in the process. A parcel can arrive promptly and then wait several days because nobody knows whether an opened product is acceptable. A refund can be approved but fail to execute. A customer can select an exchange for stock that is no longer available. Changing the destination address will not solve these problems unless the decision and handoff change too.

Keep the customer journey and the product journey visible separately. The customer's outcome might be resolved before the item is recovered. Conversely, the warehouse might have finished inspecting an item while the customer still waits. Do not force both journeys into one misleading status called "complete."

End this audit with a short statement of the main failure. For example: "Opened items wait for an owner to approve their next use; support cannot see that queue." That statement gives you something useful to fix and something specific to ask a provider to demonstrate.

Define the work before deciding where products should go

Group returns by the work they need, rather than sending every item down one route. A sealed accessory, an assembled piece of furniture and a device containing customer settings are different operational tasks. Your grouping should reflect the real products you sell and the requirements your qualified product owners have established.

For each group, describe what staff must identify, inspect and record. Replace "check condition" with observable instructions: which model and variant should be present, which components belong in the package, what evidence is required, and which findings need specialist judgment. Product-specific testing or safety decisions belong with people qualified to define and perform them.

Define the possible next actions and who may authorize each one. Some items can return to the appropriate saleable inventory after approval. Others need further review, repackaging, repair, a different sales channel or an authorized end-of-life route. Do not let a warehouse employee invent those rules simply because the parcel is taking up space.

Providers describe different scopes. ShipBob's returns service describes processing within its fulfillment network using merchant-defined preferences, including restocking, quarantine and disposal. GXO's reverse-logistics offering describes a broader set of activities that includes inspection, refurbishment and repackaging. These descriptions illustrate different kinds of work to discuss; they do not prove that either provider will accept your product or include the required service in your quote.

Send prospective partners the same anonymized case pack. Ask them to show what their team would do with the missing-component item and the unidentified parcel, not just a pristine return. The useful answer includes the decision owner, evidence, expected handoffs and unresolved-case treatment.

Compare the existing 3PL with a specialist returns center

Keeping returns with the existing third-party logistics provider, or 3PL, can be practical when most products need straightforward inspection and the provider can perform it reliably. Approved stock may already be near the inventory used to fulfill new orders, and the team may have fewer systems and transfers to coordinate.

Those advantages disappear if returns are an unsupported extra task. Ask where the inspection queue sits, who staffs it during an outbound peak, how exceptions are surfaced and how approved stock becomes available. An existing relationship is useful evidence of familiarity, but it is not evidence of returns capability.

A specialist center becomes more interesting when the returned products need work the existing provider cannot perform consistently: detailed grading, repair coordination, parts replacement, refurbishment or handling for a separate resale channel. The specialist must demonstrate the particular capability. A company calling itself a returns expert does not establish suitability for every category.

Then follow the recovered product beyond that center. Will it ship directly to a new customer, move back to the original fulfillment warehouse, wait for a consolidated transfer or enter another channel? Who checks the transferred quantity and condition? Which system owns the inventory during the move? A good inspection process can still produce a slow, expensive recovery route.

Compare these questions side by side:

  • Can the partner carry out the required work and provide evidence that it was done?
  • How long do ordinary cases and unresolved cases actually wait, including transfers?
  • Can support see enough information to explain the next step to the customer?
  • What are the full charges, minimum commitments and exit arrangements for this scope?
  • Who owns an item that arrives without usable identification or fails the agreed checks?

One destination does not have to serve every product. A split route may make sense if simple returns go to the original warehouse and a defined group goes to a specialist. But the customer should not have to work out that internal routing. Set the correct destination before providing instructions, and test the rule when an order contains different product types.

For a small brand, I would avoid creating two operational routes merely because the second supplier looks sophisticated. Add the route when its demonstrated benefit is worth the extra reconciliation and management. Use the 3PL selection guide for the wider fulfillment decision.

Compare recovery value with the complete cost of the route

More products approved for resale does not automatically mean a better financial outcome. Ask what the recovered stock is realistically worth after the remaining work, selling costs and likely markdowns. A product sitting in storage at its original retail price is not the same as cash recovered.

For an operating comparison, use the same mix and number of returns for each option. Include return transport, receipt, inspection, additional work, storage, transfers, account charges and your own administration. Identify which costs are estimates and which the provider has actually quoted. Include startup and exit costs separately so they do not disappear inside an average.

Here is a deliberately simplified, fictional monthly example. It compares 300 returned items and excludes customer refunds because those are assumed identical in both options. If the customer outcomes differ, that assumption must change.

The existing provider approves 180 items for recovery. The planning estimate is $30 of recoverable value per approved item, after subsequent selling costs. That gives $5,400. Processing costs $8 for each of the 300 returns, plus a $300 monthly charge: $2,700 in total. Estimated recovery value less these route costs is therefore $2,700.

The specialist approves 210 items at the same assumed $30 value, giving $6,300. Its route costs $11 per return, a $600 monthly charge and $600 for transfers: $4,500. Estimated value less route costs is $1,800.

The specialist recovers more units in this example but leaves $900 less after the specified costs. These figures are invented, not provider prices, accounting profit or a prediction. The point is to ask where an apparent recovery improvement goes after extra work and movement are included.

Now change the product mix and waiting time. A specialist may be compelling for valuable repairable products and unnecessary for simple sealed returns. An apparently cheap warehouse may be expensive if delays force deep markdowns or repeated customer contacts. Compare the affected product group rather than assuming that a blended result applies to every item.

Keep the assumptions connected to evidence from a trial and later invoices. For the broader order-cost view, see the real cost of fulfilling an online order.

Connect the return record, the physical item and the customer's money

Give each authorized return a reference that support and the receiving team can find. Link it to the order and expected items, but allow staff to record what actually arrived. An expected quantity of two is not evidence that the parcel contained two.

The minimum useful record is a trail of decisions: request received, instructions issued, movement observed, receipt confirmed, inspection recorded, next action approved, customer resolution confirmed and stock or financial records reconciled. Add the person or system responsible for each event. Do not require every case to pass through unnecessary manual approval if a clear rule already authorizes the action.

Keep inventory states meaningful. "In the building" should not automatically mean "available to sell." An item waiting for identification or assessment needs a visible holding state with an owner and a next review. When it is approved for a particular use, record the quantity and destination so the receiving inventory system can agree with the decision.

Apply the same discipline to refunds. Approved, requested and confirmed are different events. If the payment service times out, inspect the existing transaction before submitting another refund. If a refund fails, the case remains open for the responsible team even if the warehouse work is finished.

Define refund timing and customer options using the business's approved policy and applicable obligations. Do not change them simply to match a warehouse's preferred schedule. If the customer has already been promised a response, a transfer between providers must not silently reset the clock.

For exchanges, confirm how replacement stock is reserved or rechecked. Decide what the team does if the selected item becomes unavailable, and communicate the actual options. Calling an exchange "retained revenue" before the replacement is fulfilled and any later return is understood gives a premature picture of success.

Customers need a short explanation of what has happened and what comes next. They do not need internal warehouse codes. "We received your parcel; the missing component is being reviewed by our returns team" is useful only when that review exists and someone owns the next update.

Test ordinary returns and difficult exceptions before switching the route

Run a bounded trial using a representative mix, an agreed budget and a clear owner. The goal is to verify the route, not produce a flattering average from the easiest products. Include the ordinary cases that make up most volume and the exceptions that can leave customers or inventory waiting.

Test a parcel with no readable reference, a partial return, an unexpected item, a returned replacement, a duplicate event and an item sent to the previous address. Also test what happens when a connection fails after one system has changed but another has not. Record the expected response before observing the result.

In each test, check whether the item, record and money still reconcile at the end. Can the team explain an unresolved case without searching five inboxes? Can it distinguish an inspection delay from a refund problem? Can it recover from a failed update without creating a duplicate refund or duplicate stock?

During a provider change, keep a list of returns already authorized for the old destination. Preserve ownership of those cases until they finish. Update labels, portal rules, support instructions and customer-facing addresses together. Confirm how misrouted parcels will be handled before they start arriving.

Release a limited product group when the trial supports it. Continue monitoring the previous route until its outstanding cases are resolved. If the new arrangement fails, the rollback needs somewhere real for the next parcels to go and someone responsible for every item already in transit.

Use return and support evidence to improve the next product decision

The most valuable return may be the one that exposes a problem before the next purchase order or advertising push. To make that happen, record the customer's explanation separately from the business's interpretation. "Does not fit my shelf" is an observation. "Customer ordered incorrectly" is already a conclusion.

Keep a few facts together: product and variant, relevant batch or version when known, delivery date, contact or return reason, observed condition and resolution. Connect repeated contacts about the same incident so one difficult order does not look like five independent product failures. Include support contacts that did not become returns; a confusing product can create considerable effort even when customers keep it.

Group the evidence around possible causes. Was the product unsuitable for the customer's situation? Did the website set the wrong expectation? Was a part missing, the package damaged or the setup difficult? These causes lead to different work. A better advertisement cannot fix an unreliable component, and a supplier change may do nothing for unclear dimensions on a product page.

Imagine a fictional brand launching an organizer. Eight customers contact support because it does not fit where they intended to place it. Three return it, while the other five keep it after finding another use. Counting only returns would hide most of the difficulty. Counting each follow-up email as another affected customer would exaggerate it.

The team should look at the relevant orders, product dimensions shown before purchase and the wording in the affected campaign. It might discover missing clearance guidance, a confusing size name, a manufacturing variation or simply a use case the product does not serve. The initial phrase "too big" cannot choose among these explanations.

My recommendation is to bring a small set of original cases to the person who can change the suspected cause. Agree one action with a date and a way to check it: clarify a measurement diagram, change the packing check, investigate a batch or revise the intended-use explanation. Preserve the evidence and reasoning so the next review is more than another discussion of return percentages.

Compare later orders with similar time to develop a problem or return. Recent orders have not had the same opportunity to return as older ones. Separate a change in product mix, promotion or delivery conditions from the change you made. A drop in returns can be encouraging without proving that one action caused it.

Route serious product concerns to the appropriate qualified owner promptly; do not wait for enough complaints to make an attractive chart. For routine issues, combine frequency, severity, customer effort and avoidable cost when choosing priorities. The objective is a better product and a clearer promise, not a lower number produced by making returns harder.

Keep one regular review focused on unresolved work and better decisions

Start the review with the oldest unresolved cases and the promises at risk. Give each one an owner and a next action. Then inspect where time is accumulating: transport, receipt, assessment, authorization, customer resolution or inventory transfer. The overall average is less useful if it hides one queue that never gets cleared.

Report counts with consistent denominators. Returned units divided by shipped units answers a different question from orders with a return divided by shipped orders. For either measure, state the cohort and observation period. Keep exchanges, partial returns and repeat returns visible rather than switching definitions between reports.

Use a few measures together: unresolved-case age, time to customer resolution, time until approved stock is available, complete route cost and the observed reasons behind preventable returns. Reconcile the operational view with finance before treating estimated recovered value as a realized business result.

End by recording the decision: keep the route, correct a specific step, expand the tested scope or stop. Assign the product-learning actions as well as the warehouse actions. A returns process earns its place when it closes today's cases and makes tomorrow's customer promise more reliable.

Questions before changing your returns operation

Should all returns go to the same warehouse?

Not necessarily. Use one route when it can perform the required work reliably. A separate route may be justified for a product group needing specialist assessment or recovery, but include transfers, system reconciliation and customer instructions in the comparison. Start with a demonstrated need rather than adding a location for its own sake.

Does an item received at the warehouse count as saleable stock?

Receipt only confirms that an item arrived. Record its identity, quantity and condition, then apply the approved decision for that product. Keep items awaiting assessment visible without treating them as ready for a new customer.

Can we judge the new process by a lower return rate?

A lower rate alone is incomplete. Check customer resolution, waiting cases, product mix, observation time and full operating cost. Use customer explanations and physical findings to understand what changed, and make sure the process has not merely made legitimate returns more difficult.

Sources

  1. www.shipbob.com — returns-management
  2. gxo.com — reverse-logistics

Still have a question?Ask the DTCUP Community ↗

FAQ

Questions related to this insight.

Clear scope, practical constraints and direct answers for the decision ahead.

Should all returns go to the same warehouse?

Not necessarily. Use one route when it can perform the required work reliably. A separate route may be justified for a product group needing specialist assessment or recovery, but include transfers, system reconciliation and customer instructions in the comparison. Start with a demonstrated need rather than adding a location for its own sake.

Does an item received at the warehouse count as saleable stock?

Receipt only confirms that an item arrived. Record its identity, quantity and condition, then apply the approved decision for that product. Keep items awaiting assessment visible without treating them as ready for a new customer.

Can we judge the new process by a lower return rate?

A lower rate alone is incomplete. Check customer resolution, waiting cases, product mix, observation time and full operating cost. Use customer explanations and physical findings to understand what changed, and make sure the process has not merely made legitimate returns more difficult.

Your workspace

Selected services

DTCUP / YOUR NEXT PERSPECTIVE

Stay curious.
Stay connected.

Ideas for the work behind growth—websites, brands and the apps that connect them.