Why a Great Driver Experience is Critical for Seamless Returns Experience in 2026

A returns collection is the only moment in a returns journey where the customer meets a person, and that person is doing a harder job than a delivery asks of them. They have to verify the item against a return authorization, judge whether its condition is acceptable, handle the conversation if it is not, and fit the goods into a vehicle that is filling rather than emptying. Most pay schemes and most dispatch plans treat that stop as equivalent to dropping a parcel, which means the driver absorbs the difference in time and judgment without compensation. What follows is predictable: collections get rushed, door decisions get worse, and the customer experiences the result as a returns problem rather than a driver one. Locus, the world’s first Decision-Intelligent, Agentic TMS, plans and prices collections as their own class of work against more than 250 real-world operating constraints.

Why the Driver Decides the Returns Experience: The Business Case

The returns experience now influences the sale. The National Retail Federation’s 2025 Retail Returns Landscape, produced with Happy Returns, found 82% of consumers cite free returns as a major consideration when buying, up from 76%, against returns running at 15.8% of annual retail sales, worth $849.9 billion. The collection is the only physical touchpoint in that experience.

The labor delivering it is scarce and expensive. 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, with driver-related line items among the fastest-rising. A pay scheme that quietly penalizes the work you most want done well is expensive in a market where drivers can move.

The economics that make collections viable come from density. McKinsey finds that raising drops per stop from one to five cuts labor and vehicle cost by more than 50%, and Locus data puts standalone return pickups at $4 to $6 per item against under $1.50 when the collection joins an existing route. Consolidation is what makes the program affordable, and it also puts collections onto routes designed around delivery timing.

That mix is spreading. 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 means more drivers working to more varied standards on the same customer promise.

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

What a Collection Asks of a Driver

The model below assumes a delivery takes 3.5 minutes at the door, a collection 6.5, travel between stops 2.8 minutes, and a flat $2.20 per stop. Values are illustrative; substitute your own and the structure holds.

1 A collection is a different job wearing the same uniform

A delivery is a handover with a confirmation. A collection is a verification, a condition judgment, a potential refusal conversation, and a loading problem, because the vehicle is filling rather than emptying and the last item has to come off first at the hub.

None of that is visible in a stop count. A route with fourteen deliveries and six collections looks like twenty stops on a plan and on a payslip, and feels like considerably more than twenty stops to the person doing it.

2 The extra time is real and usually unpaid

Collections in the route	Stops per hour	Driver earns per hour	Against an all-delivery route
0%	9.5	$20.95
10%	9.1	$20.00	Down 5%
20%	8.7	$19.13	Down 9%
30%	8.3	$18.33	Down 12%
40%	8.0	$17.60	Down 16%
50%	7.7	$16.92	Down 19%

At a thirty percent collection mix the driver earns 12% less per hour than a colleague running deliveries only, for work that is more demanding. Nobody designed that outcome. It falls out of paying per stop when stops are not equivalent.

The behavioral consequence is the part that reaches the customer. A rational driver responds by compressing the collection, because the compressible part of the stop is the verification rather than the driving. The pay scheme is, in effect, purchasing a faster and worse inspection.

There is a slower consequence that compounds. Drivers choose routes where they can, and the ones they avoid are the collection-heavy ones, which means the work the returns program most depends on is allocated to whoever has least choice: usually the newest people. Over time the drivers who know the returns customers, the difficult addresses and the marginal condition calls are the ones with enough seniority to be running deliveries instead, and the returns experience is delivered by the least experienced population in the fleet.

That is not a driver problem, it is a consequence of how the work was priced, and it is invisible in any report that counts stops. A TMS that models collection service time separately makes it visible immediately, because the route times stop matching the pay.

3 Parity costs about the same as the movement itself

Collection mix	Premium needed for parity	As a share of stop pay	Against a $1.50 consolidated pickup
10%	$1.05	48%	70%
20%	$1.05	48%	70%
30%	$1.05	48%	70%
40%	$1.05	48%	70%
50%	$1.05	48%	70%

The premium is invariant to the mix, because it is compensating the extra three minutes per collection rather than the proportion of them. That makes it easy to price: a collection is worth about one and a half stops.

The comparison in the last column is the one worth carrying into a budget meeting. Paying the driver properly for a collection costs roughly 70% of what moving the item costs once it is consolidated. Anyone treating driver pay as a rounding error next to transport cost has the ratio wrong by an order of magnitude.

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

4 The driver is the inspector, and both errors are expensive

Somebody has to decide whether the item at the door is the item on the authorization and whether its condition qualifies. In most operations that somebody is the driver, working from a phone screen, in the weather, with the customer watching.

A false accept sends an item into the network that should not have entered it, consuming receiving and inspection time and usually producing a disputed refund. A false reject leaves the customer holding goods they expected to be rid of, with the refund clock running and a contact guaranteed. The second is usually the more expensive of the two and the more damaging to the relationship, because the customer experiences it as being accused of something. It is also the error a rushed driver makes more often, because refusing is the safer call when you are unsure and behind schedule, and the cost of that caution lands on somebody else’s budget entirely.

5 Door time has an optimum, and it is not the shortest

Door time	Judgment error rate	Expected error cost	Labor saved against 6.5 min	Net
6.5 min	4%	$0.44
5.0 min	7%	$0.78	$0.55	+$0.22
4.0 min	11%	$1.22	$0.92	+$0.14
3.0 min	18%	$2.00	$1.28	Minus $0.27
2.0 min	28%	$3.11	$1.65	Minus $1.01

This is more useful than a blanket instruction to take care. There is a real optimum around five minutes, where the time saved still exceeds the errors introduced. Below three minutes the trade inverts and the operation is buying speed with money.

What makes the result actionable is that door time is not usually a driver’s choice. It is set by how tight the route is and how the driver is paid, which means both levers sit with the planner rather than with the person at the door. Both are TMS settings rather than training topics, which is an uncomfortable conclusion for an operation that has been running coaching sessions on collection accuracy.

It also means the error rate is forecastable before the day starts. A route built with 3.5 minutes per stop against a collection that needs six is not a route that might run late, it is a route that will produce rushed inspections, and the TMS that built it knew the numbers when it did.

6 The driver knows things the system cannot see

At the moment of collection, the driver learns whether the address was findable, whether the customer was available at the promised time, whether the item was packed as described, and whether this is the second attempt. That information predicts the next collection better than anything in the plan, and almost none of it is captured, because capturing it costs door time nobody has budgeted.

This is where driver experience and data quality turn out to be the same problem. An operation that gives a driver thirty seconds of slack to record a structured reason code gets a dataset that improves future planning. An operation that does not gets a free-text field nobody fills in, and then concludes the drivers are unreliable reporters.

The loop closes inside the TMS. Reason codes from collections feed the service-time model, the address quality record and the customer availability profile, all of which the planner uses to build the next route. A system that captures them is getting better at collections every week; one that does not is planning next month’s returns with the same assumptions it used last year, and paying for the difference in failed attempts.

7 What to measure

Three numbers, all available. First, compare service time at collections against deliveries in your own telemetry, which gives you the real gap rather than the assumed one. Second, compute effective hourly earnings by collection mix across your driver population; if the drivers with the most collections earn the least, the scheme is working against the program. Third, look at first-attempt collection success and door time together, because the correlation between them tells you whether your routes are tight enough to be buying errors.

A Delivery Stop vs a Collection Stop, From the Driver’s Side

Dimension	Delivery	Collection
Core task	Handover and confirm	Verify, judge condition, accept or refuse
Typical door time	3.5 min	6.5 min
Vehicle load	Falling	Rising
Failure mode	Item returns to depot	Item stays with the customer
Difficult conversation	Rare	Structural, whenever condition is marginal
Usual pay treatment	One stop	One stop
Compressible part	Little	The verification

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

What to Look for in Driver and Returns Software

Collections modeled with their own service time. If the planner assumes a collection takes as long as a delivery, every route containing collections runs late by construction and the driver absorbs it. Ask how service time is set per stop type.

Pay logic that can price a collection differently. The parity premium is about 48% of stop pay in this model. A system that cannot express that will keep paying drivers to rush the stop you most want done carefully.

Clear accept and reject rules at the door. Judgment error falls when the rule is unambiguous and the evidence required is specified. Ask what the driver sees on the screen at a marginal condition call, and who they can escalate to.

Structured capture with time budgeted for it. Reason codes are only as good as the seconds allowed to enter them. Confirm the plan includes that time rather than assuming it is free.

Route tightness visible against error rates. If first-attempt success falls as routes tighten, the operation is buying speed with errors, and only a system holding both numbers can show it.

Collection history feeding the next plan. Reason codes are only worth capturing if the TMS uses them to revise service times, address quality and availability profiles. Ask what changes in next week’s route because of what a driver recorded this week.

Driver Experience in Action: Real-World Results

A Fortune 50 enterprise running a 4,500-strong driver pool across 51 sites, split between captive and third-party fleets, had plans built to site-level assumptions rather than to what drivers could actually execute. In the centralized dispatch deployment, consolidating planning into one decisioning layer lifted weekly execution from 75% to 92% and surfaced more than $14M in annualized unused capacity, $565K at one site before scaling across 25. An execution rate rising by seventeen points is a statement about plans becoming doable rather than about drivers trying harder.

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, moving planning into a single decisioning layer saved more than 12,000 trips per month, optimized over $4B in orders and returned 3X ROI. Trips removed are stops a rider no longer has to make, which is a driver-experience result before it is a cost one.

Common Driver and Returns Mistakes to Avoid

Counting a collection as one stop. It takes roughly twice the door time of a delivery and carries a judgment call. Paying and planning it as equivalent guarantees it gets compressed.

Tightening routes and expecting inspection quality to hold. Below about three minutes at the door, the errors introduced cost more than the time saved. Route tightness is an inspection-quality decision whether or not anyone frames it that way.

Blaming door decisions on driver diligence. Error rates track available time far more closely than they track individual care, and available time is set by the plan.

Asking for structured reporting without budgeting the seconds. Reason codes entered in a hurry are worse than none, because they look like data.

Allocating collection-heavy routes by default rather than by design. Where drivers have any choice, they avoid the routes that pay worse per hour, so the returns program ends up staffed by the least experienced people unless the allocation is deliberate.

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

How Locus Supports the Driver Side of Returns

Locus, the world’s first Decision-Intelligent, Agentic TMS, plans a collection as its own class of work rather than as a delivery in reverse. Service time is modeled per stop type in the route planning system, so a route carrying collections is built with the extra minutes in it rather than discovering them during execution, and rider and driver management holds the allocation, eligibility and performance view that lets a collection go to someone equipped to handle it. Because the Dispatch and Customer agents generate the attempt event, a marginal condition call or a refusal is recorded with a timestamp and a reason rather than reconstructed later from a disagreement, which is what makes both the customer conversation and the driver’s position defensible.

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.

A seamless returns experience depends on a driver doing a harder job than a delivery asks, usually for the same money. A collection runs about 6.5 minutes against a delivery’s 3.5, which costs a driver on a thirty percent collection mix roughly 12% of hourly earnings, and the rational response is to compress the verification, which is exactly the part the customer experience depends on. Restoring parity costs around $1.05 per collection, about 70% of what the consolidated movement itself costs, and door time has a genuine optimum near five minutes below which errors outrun the labor saved. All three levers, service time, pay treatment and route tightness, sit with the planner rather than the driver. Request a Locus returns and driver assessment to see what your own collection stops actually cost.

Frequently Asked Questions

Why does a returns collection take longer than a delivery?

Because it involves verification rather than handover. The driver has to match the item against a return authorization, judge whether its condition qualifies, handle a refusal conversation where it does not, and load goods into a vehicle that is filling. In this model that runs about 6.5 minutes against 3.5 for a delivery.

How does per-stop pay affect returns quality?

It penalizes collections. Paying the same for both stop types means a driver on a thirty percent collection mix earns roughly 12% less per hour than one running deliveries only, and the only compressible part of a collection is the verification, so the scheme effectively purchases a faster and less accurate inspection.

What premium makes a collection worth the same as a delivery?

About $1.05 per collection in this model, or roughly 48% of stop pay, and it does not vary with the collection mix because it compensates the extra minutes per stop rather than their proportion. For context that is around 70% of what a consolidated pickup movement itself costs.

How much time should a driver spend at a collection?

Around five minutes on these figures. Judgment error rises from 4% at 6.5 minutes to 18% at three, and below three minutes the expected cost of accept and reject errors exceeds the labor saved. The optimum is a real point rather than a direction.

Who is responsible for door decision errors?

The plan more than the person. Error rates track available time closely, and available time is set by route tightness and pay design, both of which are planner decisions. Treating door accuracy as a training issue addresses the least significant variable.

What should a driver capture at a collection?

Structured reason codes covering address findability, customer availability, item condition and attempt number, because those predict future collection success better than anything in the plan. The important part is budgeting the seconds to enter them, since codes entered under time pressure look like data without being it.