Courier Weight Disputes: Getting Back Money You Already Paid
Couriers bill on the higher of dead weight and volumetric weight, measured at their hub rather than your bench. The gap between what you declared and what they measured becomes a discrepancy charge with a short response window, and silence counts as agreement. This covers where the gap comes from, how to build evidence before you need it, and why COD parcels make every weight error twice as expensive.
Weight discrepancies are the least glamorous line item in a D2C operation and one of the few that leaks money quietly in both directions at once. You pay more freight than you budgeted, and because the charge arrives weeks later as an adjustment on an aggregator wallet rather than as an invoice you approve, it rarely gets attributed to the SKU that caused it. For a COD-heavy brand the leak is wider than it looks, for a reason that has nothing to do with weighing scales: a refused parcel is billed on both legs, so a packed weight that is wrong by 200 grams is wrong twice.
What you are actually billed for
Freight is not charged on the weight of your product. It is charged on billed weight, which is the higher of two numbers: the dead weight on the scale, and the volumetric weight derived from the parcel’s outer dimensions. The logic is straightforward — a vehicle runs out of space before it runs out of payload, so a large light parcel consumes capacity that a small heavy one does not.
The volumetric figure comes from a divisor applied to the parcel’s dimensions. Logistics aggregator Shiprocket states the formula as Volumetric Weight = (Length x Breadth x Height) / 5000, with dimensions in centimetres. Divisors vary by carrier and by service — air and surface services are often priced differently, and express services sometimes use a tighter divisor — so the number that matters is the one in your own rate card, not the industry-standard one. Check it, because a brand shipping bulky-but-light goods on a tighter divisor is paying a materially different price per parcel than the same brand on a looser one.
Two consequences follow immediately. First, the box matters as much as the product: shipping a 400-gram item in a box sized for comfort rather than fit can push the billed weight above the dead weight and keep it there for every unit you ever ship. Second, both of the inputs to billed weight are measured by the courier, at their facility, after the parcel has left you. Whatever you typed into the manifest is a declaration, not a measurement.
Where the gap comes from
Almost every recurring discrepancy traces back to one of a handful of causes, and they are worth separating because they have different fixes.
Catalogue weight instead of packed weight. This is the most common cause by a wide margin, and Shiprocket names it explicitly: discrepancies usually arise when sellers enter the product’s own weight rather than the volumetric weight of the packed parcel. The product weighs what the spec sheet says. The parcel weighs that plus the carton, the filler, the tape, the invoice sleeve and the branded tissue. If your catalogue was populated from supplier data, every single SKU is understated by the weight of your own packaging.
Dimensions that describe the product, not the box. The same problem in the other variable, and it is worse because dimensions are cubed in their effect. A box that is 2cm larger in each direction than it needs to be is not 2cm of waste; it can move the parcel across a pricing slab.
Multi-item orders. Single-SKU parcels are easy to model. The moment a customer buys three things, the packed dimensions depend on which three, which box the packer reached for, and how it was filled. If your system derives parcel weight by adding SKU weights, it will be wrong on every multi-item order, and it will be wrong in the expensive direction because it ignores the air inside the box.
Slab boundaries. Rate cards are banded — typically a base slab and then increments. A parcel at 490 grams and a parcel at 510 grams are not 4% apart in cost; they are a whole slab apart. SKUs that sit just under a boundary are the ones where packaging weight silently doubles your freight, and they are the first place to look when a discrepancy pattern appears.
Genuine measurement variance. Hub scales and dimensioning equipment are not laboratory instruments, and a parcel that has been handled, compressed or re-taped in transit does not present the same dimensions it did on your bench. Some share of discrepancies will be noise. The point of good records is to tell the noise apart from the pattern.
The dispute clock, and why it is the real constraint
The mechanics differ by provider, but the shape is consistent: the courier reports a measured weight, the difference becomes a proposed charge, and you get a defined window in which to accept it or contest it. If you do nothing, the charge goes through.
Shiprocket’s published process is a useful concrete example, and worth reading as a template rather than a universal rule. The company states that a seller has 7 working days to raise a dispute once an incorrect charge has been identified — its own FAQ on the same page says 7 days rather than 7 working days, so treat the shorter reading as the safe one. It states that disputed amounts are put on hold and cannot be used by either of the parties until the dispute is resolved, that disputes are resolved within 5-6 working days, and — the line that costs brands the most money — that in case the sellers do not respond to the dispute within 7 working days, Shiprocket auto accepts, and the amount is debited from the seller’s wallet. It also notes that weight disputes are independent of tracking and do not affect the shipment itself, which is worth knowing: contesting a charge does not put the parcel at risk.
Your own provider’s window and process may differ, and the terms change, so verify yours rather than inheriting these numbers. But the structural lesson holds everywhere: this is a process with a default outcome, and the default favours the party that already has the money. A discrepancy queue that nobody checks is not a backlog — it is a series of decisions being made in your absence.
That makes ownership the first fix, before evidence, before packaging, before anything. Someone has to own the reconciliation queue, with a named day of the week, because the window is short enough that a fortnightly review guarantees you lose the ones raised at the wrong moment. If your team is small, the day matters more than the person.
Evidence that wins, and when it has to exist
The critical insight about weight disputes is one of sequencing. By the time a discrepancy is raised, the parcel is gone. You cannot go and re-measure it. Whatever you are going to argue with has to have been captured at pack time, days or weeks earlier, by someone who had no idea this particular parcel would be questioned.
So the evidence process is not a dispute process. It is a packing process.
What providers ask for is relatively modest. Shiprocket’s support documentation describes selecting the order, adding a product category, remarks and an optional product URL, and then uploading images of your package as a piece of evidence; its customer-facing page adds that the seller shares an image of the product along with a description and category. Package images are the substance of it.
That modesty is an opportunity rather than a relief. A reviewer comparing a claimed weight against a hub measurement is looking for something that makes your number credible. Images that do that have a few properties in common:
The AWB or order number is legible in the same frame as the parcel. An unlabelled box on a scale proves that some box weighed something. The label is what ties the image to the shipment under dispute, and without it the image is decoration.
The reading is readable. A scale display at an angle, in warehouse lighting, with the operator’s hand in shot, is not usable. One clean overhead shot where the digits can be read without zooming is worth more than five hurried ones.
Dimensions are shown, not asserted. If the dispute is volumetric rather than dead weight — and for light bulky goods it usually is — a photograph of the sealed parcel against a tape measure on each axis is the only thing that addresses the actual question.
It is systematic rather than selective. Capturing images only for parcels you expect to be queried produces exactly the records you cannot use, because you cannot predict which parcel gets flagged. The cheap version of systematic is a fixed station: scale, tape, phone on a stand, one photograph per parcel at the moment the label goes on, filed by AWB. The marginal cost is seconds per parcel and the marginal benefit is the entire dispute capability.
If capturing every parcel is genuinely out of reach, capture by exception in a way that still works: every parcel above a value threshold, every multi-item order, and a rotating sample of each high-volume SKU. The sample is what lets you argue about a SKU pattern even when you cannot argue about an individual parcel.
Prevention is worth more than recovery
A won dispute returns the overcharge. It does not return the labour spent filing it, and it does not stop the same charge arriving next week on the same SKU. The durable work is upstream.
Weigh and measure every SKU as packed, once. Not the product — the finished parcel, in the box it actually ships in, with the filler it actually gets. Record both the dead weight and the dimensions, and write those into your shipping system as the defaults. For a catalogue of a few hundred SKUs this is a day of work and it eliminates the single largest cause of discrepancies permanently.
Re-measure when packaging changes. A new box supplier, a switch from bubble wrap to paper, a seasonal mailer, a change in how a garment is folded — each one invalidates the numbers you recorded. Packaging changes are made by people who have no reason to think about freight slabs, which is exactly why this needs to be a checklist item rather than an instinct.
Rationalise box sizes against your slabs. Work out, for your actual rate card, where the slab boundaries fall in both dead and volumetric terms, then check which of your boxes push common orders over one. Dropping a box size, or reducing filler on a specific SKU, is often worth more per month than any dispute you will ever win.
Model multi-item parcels properly. Rather than summing SKU weights, define the handful of box sizes you genuinely use and have the system pick a box and apply its recorded packed profile. Approximate but honest beats precise but wrong.
Treat the discrepancy report as a data source. The pattern is more valuable than the money. Sort flagged shipments by SKU and by courier: a SKU that appears repeatedly has a catalogue error, a courier that flags across your whole catalogue is either measuring differently or on a different divisor, and a sudden cluster usually means a packaging change nobody told you about.
Why COD makes all of this twice as expensive
Everything above applies to any brand shipping parcels. Two things make it sharper if a large share of your volume is cash on delivery.
The first is double billing on failure. A refused parcel is carried out and carried back, and both legs are priced off the same billed weight. If a SKU is understated by 150 grams and that pushes it a slab higher, every RTO on that SKU pays the error twice. The expected cost of a weight error is therefore a function of your RTO rate, which means weight accuracy quietly matters most in exactly the categories where refusals are most common.
The second is margin headroom. COD parcels already absorb a handling charge, a longer cash cycle and a refusal rate that prepaid orders do not have. On a discount-led basket, there is very little left for an avoidable freight slab. The arithmetic that decides whether a COD order is worth accepting at all depends on a per-parcel freight figure — and if that figure is the one from your rate card rather than the one you are actually being billed, every downstream decision built on it is slightly wrong in the same direction.
This is also the argument for reconciling freight at the order level rather than the monthly-total level. A blended cost per parcel hides the SKUs and lanes where you are losing money, and those are precisely the ones where a COD or prepaid-only decision would pay for itself.
When not to dispute
Not every discrepancy is worth contesting, and a team that disputes everything burns credibility along with hours.
Work out roughly what a filed dispute costs you in labour, and treat that as a floor — below it, accept and move on, but still record the SKU so it feeds the pattern. Where the courier’s measurement is probably right because your own data is thin, accept it and fix the catalogue entry; disputing a charge you cannot evidence is a slower way of accepting it. Where the gap is small and sits within normal measurement variance, let it go. And where the same SKU is flagged every week, stop disputing parcel by parcel and fix the SKU — the recurring charge is information, not an injustice.
What is always worth contesting: large single gaps, anything where the measured weight is implausible for the goods, and any case where you hold a clean, labelled, dated image of the parcel on a scale. Those are the ones where the evidence does the work.
What to track
Four numbers are enough to run this well. The share of shipments flagged for a discrepancy, which tells you whether this is a process problem or background noise. The average gap between declared and billed weight, which tells you whether your catalogue is systematically understated rather than occasionally wrong. Your dispute win rate, which is mostly a measure of your evidence quality. And actual freight cost per parcel by SKU, which is the number everything else is in service of.
If only one of those gets a dashboard, make it the last one. A freight cost per parcel that reflects what you were really billed, rather than what the rate card said, is the input to your COD thresholds, your free-shipping floor, your packaging decisions and your category-level margin. Most brands do not have it, and the ones that do tend to find that a handful of SKUs are responsible for most of the gap.
The short version
You are billed on the higher of dead and volumetric weight, measured at the courier’s hub, not on your bench. Most discrepancies come from catalogue weights that describe the product rather than the packed parcel. The dispute process has a short window and a default outcome that favours the courier, so ownership and a weekly rhythm matter more than argument quality. Evidence has to be captured at pack time, with the AWB legible in the frame, because it cannot be created later. Prevention — weighing every SKU as packed, re-measuring when packaging changes, and sizing boxes against your slab boundaries — returns more than recovery ever will. And if you ship COD, every weight error is priced twice on each refused parcel, which is why weight accuracy belongs in the same conversation as RTO.
COD King does not sit in this part of the stack: freight reconciliation is between you, your aggregator and your couriers. The connection is indirect but real — a refused parcel is what makes a weight error expensive, so verification, partial advances and COD rules by pincode and order value reduce the number of parcels that ever get billed twice. The weighing scale, the box sizes and the discrepancy queue stay yours.
Figures and process details attributed to Shiprocket are from its weight dispute management page and its weight discrepancy support article, checked 18 November 2026. Terms change; verify against your own provider.
