Why Returns Management is a TMS Problem, Not Just a Warehouse Problem

Returns management is usually owned by the warehouse, because that is where returned goods are received, inspected, graded and put back into stock, and it is where the visible work happens. The cost, however, splits almost evenly: in the model below, transportation carries 49% of the cost per return and the warehouse 47%. More importantly, transportation owns the one decision that determines what the warehouse costs, which is when returns arrive. A collection schedule set to minimize transport cost delivers returns in large infrequent batches, and a warehouse sized to absorb those batches is materially bigger than one receiving the same annual volume smoothly. Locus, the world’s first Decision-Intelligent, Agentic TMS, holds that timing decision inside the same plan as the forward network, solved against more than 250 real-world operating constraints.

Why the Ownership Question Matters: The Business Case

The volume is large enough that a misallocated half is a real number. The National Retail Federation’s 2025 Retail Returns Landscape, produced with Happy Returns, puts returns at 15.8% of annual retail sales in 2025, worth $849.9 billion, with online returns at 19.3%. Half of the handling cost of that flow sits outside the function that is usually accountable for it.

The transport half is also the half whose unit cost has been moving. ATRI’s 2026 analysis puts the industry-average operating cost at $2.336 per mile in 2025, up 3.4% and the highest in the report’s history, so decisions that put vehicles on the road for thin collections have got worse rather than better.

The mechanism that makes transport cost fall is density, and it is well documented on the forward side. McKinsey finds that raising drops per stop from one to five cuts labor and vehicle cost by more than 50%. Applied to returns, that is exactly the pressure that pushes toward infrequent batch collection, which is where the conflict with the warehouse begins.

Where a retailer’s own network cannot supply density, capacity has grown elsewhere. The Pitney Bowes Parcel Shipping Index records non-major carriers doubling their share of US parcel revenue from 3.4% to 7.2% in a 23.1 billion parcel market, which widens the set of options but does not change who owns the timing decision.

Also Read: AI in Reverse Logistics: Turning Returns into a Competitive Advantage

Where Returns Cost and Control Actually Sit

The model below prices a single return through six steps, and sizes a warehouse receiving an average of 500 returns a day to clear 95% of days without backlog. Values are illustrative; substitute your own and the structure holds.

1 The cost splits almost evenly, and the label does not

Step	Owner	Cost per return	Share
Collection from the customer	Transport	$3.33	32%
Line-haul to the receiving site	Transport	$1.80	17%
Receiving and put-away	Warehouse	$0.90	9%
Inspection and grading	Warehouse	$2.40	23%
Disposition and restock	Warehouse	$1.60	15%
Refund processing	Finance	$0.35	3%
Total		$10.38

Transportation carries $5.13 of the $10.38, or 49%. The warehouse carries $4.90, or 47%. The single most expensive step in the chain is collection from the customer, and it happens before the warehouse sees anything.

A program described as warehouse-owned will produce improvement plans aimed at the 47%, reported to people accountable for the 47%, and benchmarked against other warehouses. That is a reasonable way to run half a problem.

2 Transportation also owns the one decision that prices the warehouse

Cost share alone would make this a shared problem rather than a transport-led one. What makes it transport-led is that the arrival pattern is a transport output. The warehouse does not choose when returns show up; it chooses what to do with them once they have.

That distinction matters because warehouse capacity is sized to peak days rather than average ones. A receiving operation that can handle the mean will backlog on every above-average day, and a returns backlog is expensive twice, in the labor to clear it and in the refund days it adds.

3 Collecting less often makes the peak day much larger

Collection interval	Arrival variability	Peak day at 95%	Capacity needed against the mean
Daily	0.18	648	1.30x
Every 2 days	0.42	845	1.69x
Every 3 days	0.61	1,002	2.00x
Every 5 days	0.95	1,281	2.56x
Every 7 days	1.28	1,553	3.11x

Same annual volume in every row. A retailer collecting every three days has to build a receiving operation twice the size of the average day’s work, and one collecting weekly has to build three times. Nothing about the warehouse changed. A transport scheduling decision did.

This is the concrete form of the argument in the title. The warehouse cannot manage its own capital requirement, because the variable that sets it is owned by another function. It also explains a pattern most retail operations will recognize without having named it. Receiving is chronically over-resourced on some days and backlogged on others, agency labor is booked to cover peaks that arrive on a schedule nobody in the warehouse set, and every annual planning round produces a request for more receiving capacity that is approved because the peak-day evidence is real. The evidence is real. It is also downstream of a decision made elsewhere, and no amount of capacity solves a problem that is being generated upstream every week.

Also Read: US Returns Hit $850 Billion in 2025: Why US Retailers Are Restructuring Reverse Logistics in 2026

4 Each function optimizing alone lands in the wrong place

Transport cost falls as the collection interval lengthens, because collections consolidate. Warehouse cost rises, because capacity is sized to a larger peak. Holding warehouse labor constant at $3.20 per return and letting capacity cost scale with the peak:

Collection interval	Peak against mean	Transport	Warehouse labor	Warehouse capacity	Total per return
Daily	1.30x	$5.13	$3.20	$1.70	$10.03
Every 2 days	1.69x	$4.05	$3.20	$2.22	$9.47
Every 3 days	2.00x	$3.52	$3.20	$2.63	$9.35
Every 4 days	2.28x	$3.24	$3.20	$2.99	$9.43
Every 5 days	2.56x	$3.02	$3.20	$3.36	$9.58
Every 7 days	3.11x	$2.78	$3.20	$4.07	$10.05

A transport function optimizing its own line picks the seven-day interval and produces a total of $10.05. A warehouse function optimizing its own line picks daily collection and produces $10.03. The joint optimum sits at three days and produces $9.35, roughly 7% better than either.

5 The reason the argument never resolves is that both sides are equally wrong

Look at those two isolated answers again: $10.05 and $10.03. They are within two cents of each other.

That is why this disagreement is so durable inside retailers. Transport can demonstrate that weekly collection is cheaper on its number. The warehouse can demonstrate that daily receiving is cheaper on its number. Both are telling the truth, both produce approximately the same total, and neither has access to the figure that would settle it. The argument is not a failure of collaboration, it is a failure of vantage point.

It also means the usual resolutions do not help. Splitting the difference by negotiation lands somewhere reasonable by accident rather than by calculation. Escalating to a shared boss produces a decision with no more information than either function had. Only a system that prices both lines against one objective can find the three-day answer, and it finds it by arithmetic rather than by seniority.

The shape of the curve is worth noting as well, because it is flat near the bottom. Two, three and four days all land within nine cents of each other, while the extremes cost around seventy cents more. That is encouraging rather than discouraging: an operation does not need to find the exact optimum, it needs to stop sitting at one of the ends. Most retailers are at an end, because an end is what a single function chooses when it optimizes alone.

6 The decision belongs where both variables are visible

This is not an argument that returns belong to transportation instead of the warehouse. It is an argument that the decision which sets both costs is a scheduling decision, that scheduling decisions are made in the transportation layer, and that the transportation layer is therefore where the joint optimization has to happen.

A returns platform that sits above the warehouse and hands collection to a carrier cannot make this call, because it holds neither line. A warehouse management system cannot make it, because it receives the arrival pattern as an input. A transportation management system holding collection scheduling alongside the forward network is the only one of the three that can see the variable and act on it, which is what makes returns management a TMS problem in practice rather than in category theory.

7 What to compute before reorganizing anything

The analysis is a day’s work on data most retailers already have. Take a year of returns, split every cost line by which function decides it, and total the two columns. Then take daily receiving volumes, compute the coefficient of variation, and find the 95th percentile day against the mean. That ratio is your capacity multiplier, and it is the price you are currently paying for your collection schedule.

Finally, vary the collection interval on paper and plot transport cost and warehouse capacity cost against each other. The minimum of the sum is rarely where either function would have put it, and seeing the curve is usually enough to end an argument that has been running for years. Do it jointly, with both functions in the room, because the output is only persuasive to the person whose number it contradicts if they watched it being built.

One caution on the inputs. The warehouse capacity line should be the cost of capacity the operation actually holds, including agency labor booked against peaks and space leased to absorb them, not a theoretical cost per unit processed. Understating it is the most common way this analysis gets run and produces an answer that looks like the status quo, because the peak-driven costs are usually spread across several budget lines and none of them is labeled returns.

Returns as a Warehouse Program vs a Transport-Led One: Key Differences

Dimension	Returns owned by the warehouse	Returns owned in the TMS
Cost visibility	The 47% that happens on site	The full 100% including collection
Collection schedule	Received as a given	A decision variable
Warehouse capacity	Sized to whatever arrives	Sized against a chosen arrival pattern
Improvement targets	Inspection and put-away throughput	The joint cost curve
Response to volume growth	More receiving capacity	Re-evaluate the interval first
Who can settle the trade-off	Nobody, both numbers are local	The layer holding both lines

Also Read: Reverse Logistics Explained: Process, Types and Benefits in 2026

What to Look for in Returns Management Software

Cost reporting across the whole chain, not the warehouse half. If the platform reports inspection throughput but takes collection cost from a carrier invoice, it cannot show the trade-off that matters. Ask for one cost-per-return figure with both halves in it.

Collection interval as a configurable variable. The system should let an operation model different collection frequencies and show what each does to arrival variability, not just execute the schedule it was given.

Arrival forecasting the warehouse can plan against. Receiving should know what volume is coming before it appears at the dock. That forecast is a transport output and is rarely shared.

Joint objective rather than two local ones. Ask what the platform optimizes. If the answer is transport cost, it will drift toward long intervals and quietly enlarge the warehouse. If it is warehouse throughput, it will do the reverse.

Collections planned against the forward network. Placing a collection on a vehicle already serving the street is what makes frequent collection affordable, which is how an operation escapes the trade-off rather than balancing it.

Joint Optimization in Action: Real-World Results

A leading North American retailer with several hundred stores replaced six legacy systems spanning ocean, rail and road in a multimodal automation program. Because mode, routing and scheduling decisions moved into one layer instead of being split across systems that each optimized locally, exceptions resolved in under two hours against on-time store delivery above 99%, with route compliance above 95% and manual dispatch reduced more than 80%. The program returned more than $1M in savings and broke even in year one. Store networks are also where most returns consolidate, so the same platform decision governs both directions.

A global FMCG manufacturer operating across 10 Asian countries ran distribution through more than 1,000 distributors and over 5,000 riders, with collection and delivery planned by different parties on different cycles. In the logistics automation deployment, consolidating planning into a single decisioning layer saved more than 12,000 trips per month, optimized over $4B in orders and returned 3X ROI. Trips saved is precisely the quantity that appears when two functions stop optimizing separately.

Common Returns Ownership Mistakes to Avoid

Benchmarking the returns program against other warehouses. It measures the 47% and ignores the half that sets it. Benchmark cost per return end to end, including collection.

Letting transport set the collection schedule on its own line. It will lengthen the interval, which is correct on its number and adds more to warehouse capacity than it saves.

Letting the warehouse demand daily collection. It will smooth arrivals, which is correct on its number and costs more in collection than it saves in capacity.

Escalating the disagreement rather than computing it. Both functions are right on their own figures and within cents of each other on the total, so seniority decides rather than evidence. Plot the joint curve instead.

Also Read: Top Reverse Logistics Companies and Software Solutions of 2026

How Locus Holds Both Sides of the Trade-Off

Locus, the world’s first Decision-Intelligent, Agentic TMS, treats collection scheduling as a decision with consequences on both sides rather than a transport setting. The route planning system places collections on vehicles already serving the street, which is what makes frequent collection affordable and therefore removes most of the pressure to batch, and the arrival pattern it produces is published forward so receiving can plan against a forecast rather than a surprise. Because the Capacity, Dispatch and Carrier agents hold forward and reverse flows in one plan with a single objective, the interval is chosen against the total rather than against whichever line the owning function is measured on.

Locus is recognized by Gartner for seven consecutive years, featured in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, named a Leader in TMS by QKS Group in the SPARK Matrix, and ranked #1 in Route Planning on G2’s 2026 Best Software Awards. In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently.

Returns management is filed under the warehouse because that is where the visible work happens, but transportation carries 49% of the cost per return and owns the decision that prices the other half. Collecting every three days rather than daily doubles the peak-day volume a receiving operation has to be built for, at identical annual throughput. When each function optimizes its own line, transport chooses a seven-day interval and the warehouse chooses daily, the two answers land within two cents of each other at roughly 7% above the joint optimum, and neither function holds the number that would settle it. The decision is a scheduling decision, scheduling happens in the transportation layer, and that is where the two lines have to be priced together. Request a Locus returns assessment to see where your own curve bottoms out.

Frequently Asked Questions

Is returns management a warehouse function or a transportation function?

Both, and the cost split is close to even: in this model transportation carries 49% of cost per return and the warehouse 47%. What makes it transport-led in practice is that transportation decides when returns arrive, and the arrival pattern sets how much receiving capacity the warehouse has to build.

Why does a TMS matter for returns management?

Because the collection schedule is a transportation decision that prices both sides. A transportation management system holding collection alongside the forward network is the only layer that sees transport cost and warehouse capacity cost at the same time, which is what the trade-off requires.

How does collection frequency affect warehouse cost?

Warehouse capacity is sized to peak days rather than averages. Collecting daily produces a peak of about 1.30 times the mean; collecting every three days produces 2.00 times and weekly 3.11 times, at identical annual volume. The receiving operation has to be built for the peak.

What is the right collection interval for returns?

In this model the joint optimum sits at every three days, at $9.35 per return. Transport optimizing alone picks seven days at $10.05 total, and the warehouse optimizing alone picks daily at $10.03. The right answer depends on your own cost lines, but it is rarely at either extreme.

Why do transport and warehouse teams disagree about returns?

Because both are right on the number they are measured on, and those numbers point in opposite directions. In this model the two isolated answers produce totals within two cents of each other, so neither function can win the argument on evidence and the decision usually falls to seniority instead.

Can a returns management platform solve this on its own?

Not if it sits above the warehouse and hands collection to a carrier, because it then holds neither cost line. Settling the trade-off requires a system that owns collection scheduling and can price it against receiving capacity, which in most architectures means the transportation management system.