In short

  • Margin leaks in two places: upstream at the factory, where the cost is baked in before goods exist, and downstream in the market, where you can only recover pennies on the dollar.
  • The biggest downstream drains are overstock, returns, and discounting, and most of them are caused by an upstream decision you made months earlier.
  • The biggest upstream drains are a wrong-size curve, sampling and decoration creep, hidden factory surcharges, and rework plus freight, which hits hardest on small runs.
  • Apparel returns commonly run 20-40%, and fit alone drives up to 70% of them. The refund is rarely the real cost of processing.
  • The Leak-Audit Checklist at the end is the tool: run it against your next order before you sign anything.

The model:

The Margin Drain:

A dollar that should have reached your bank account passes through four rooms on the way there:

  • Your order.
  • Your factory.
  • Your warehouse.
  • Your market.

Money leaks out of every one.

The trick to plugging leaks is knowing which room you’re standing in because the room changes everything about what a leak costs you.

How a dollar drains out

Upstream leaks cost you full margin the money is gone before you ever make a sale. Downstream leaks are recovery jobs you’re salvaging cents from a dollar you’ve already spent.

The Margin Drain diagram A horizontal pipe carrying a dollar from Order to Bank, with upstream leaks at the Factory stage and downstream leaks at the Warehouse and Market stages. $ ORDER FACTORY UPSTREAM WAREHOUSE MARKET DOWNSTREAM ✓ BANK size curve sampling creep surcharges rework · freight overstock returns discounting

Upstream: at the factory

The leak is sealed into the cost of the goods before a single unit sells. A wrong decision here costs you 100 cents on every dollar, and no amount of clever marketing gets it back.

Downstream: in the market

The goods already exist and the cash is already spent. Every fix here is damage control: a markdown recovers part of a sale, a return refund is money flowing backward out the door.

The Upstream Rule:

A leak you fix at the factory is worth more than a leak you fix in the market.

Downstream problems are loud: the unsold rack, the returns pile, and the discount you had to take. But they’re usually symptoms.

The cause sat upstream, in an order you placed before you had a single data point. Audit upstream first.

Downstream Leak 01:

Overstock & deadstock:

Symptom: units that won’t move 
Root cause: over-ordering on weak demand signal
 

Overstock is the most expensive kind of confidence.

You believed in the drop, ordered deep to win the unit price, and now a third of the run is sitting in a box.

The box isn’t free.

Carrying cost is the quiet part.

Stored inventory burns money every week it sits in storage: insurance and the depreciation of a style that ages out of relevance.

Worse is the opportunity cost: that’s cash you can’t spend on the next drop, frozen inside garments that are slowly becoming clearance.

~3.2%

The average net profit margin across the world’s 250 largest retailers.

There is almost no cushion in this business, which is exactly why a single overbuy can erase a drop’s entire profit.

Source: Deloitte, Global Powers of Retailing via ShipBob. ref

Aged stock is a cash problem before it’s a margin problem.

Even a well-run direct-to-consumer brand can carry roughly 147 days of inventory, nearly half a year of capital locked in product.

For a founder funding the next run out of cash flow, that’s the difference between a second drop and a stalled brand.

How to plug it:

Order to your demand confidence, not the factory’s minimum.

If you have no sell-through history, a tight first run protects cash even when it raises your unit cost, and a phased reorder lets you buy more after the market has voted.

This is the whole reason a small batch exists.

See small-batch vs. bulk production for how to size a first run to the cash you can afford to risk, not the unit price you wish you could hit.

Downstream Leak 02:

Returns:

Symptom: refunds and reverse shipping
Root cause: fit uncertainty & bracketing
 

Apparel is the most-returned category in retail, and it isn’t close.

Online clothing return rates commonly run 20-40%, meaning up to two in five units you ship may come back.

That’s not a policy problem you can fix with stricter rules.

It’s the structure of selling clothes people can’t try on.

~70%

The share of apparel returns driven by fit and sizing is the single largest cause, far ahead of everything else.

Get the size curve and the size chart right, and you’ve addressed most of your returns at the source.

Source: returns-industry benchmarks via Richpanel (2026), citing Coresight Research. ref

Then there’s bracketing: shoppers ordering two or three sizes intending to keep one and send the rest back.

It’s now a mainstream habit among online apparel buyers, up sharply from roughly 40% in 2018.

For every bracketed order book revenue you’ll partly refund, and for shipped units you’ll pay to recover.

The refund is not the cost:

Here’s the part founders miss.

The refund is just the sale reversing.

The loss is everything around it: reverse shipping, inspection, repackaging, restocking labor, and sometimes the unit comes back unsellable.

Processed end to end, a single return can eat 20-65% of the item’s original price.

Worked example: one returned hoodie

A $60 retail hoodie that costs you $22 to land.

   Sale reversed (refund to customer): net zero
   Return shipping label: −$9
   Inspection + repackaging labour: −$6
   Restock/loss if unsellable: −$3 to −$22
   Cost of the return: −$18 to −$37
 

On a hoodie with ~$38 of gross margin, a single mishandled return can swallow half of it or all of it if the unit can’t go back on the shelf.

The above example is live results from one of our client.

How to plug it:

You can’t engineer returns to zero; the category sets the floor.

What you control is the fit accuracy that prevents the return and the cost of handling the ones you get.

Lock your grading and publish a size chart built from your actual garments, not a borrowed template.

Most fit returns are a measurement you never gave the customer.

Downstream Leak 03:

Discounting & markdowns:

This is the loudest leak and the most misunderstood.

A markdown feels like a marketing decision.

It’s actually a margin transfer straight out of your profit into the customer’s pocket, and the discount comes out of margin, never out of revenue.

The markdown math

Take a $100 unit at a 55% gross margin; it costs you $45.

Full price: $55 profit
30% markdown → $70, $25 profit
55% clearance → $45, $0 profit

A 30% discount doesn’t cost you 30%; it more than halves your profit.

A clearance price hands the whole drop’s margin away.

And the baseline has shifted under everyone.

Full-price sell-through has slid from a 70-75% norm to roughly 50% at many fashion retailers, so markdowns went from a clean-up tool to half the business.

Across the sector, markdowns are now estimated to consume 20-50% of net sales.

When a public retailer like American Eagle reported a 2.7-point gross-margin slip tied to markdowns and a write-down, its operating income fell by nearly half.

A small slide in how much you give back to discounting can swing the whole bottom line.

The contrast is instructive: by producing in smaller batches and restocking only what sells, Zara reportedly hits around 85% full-price sell-through against an industry average closer to 60-70% and spends far less on markdowns as a result.

Discipline upstream means fewer markdowns downstream.

How to plug it:

Stop treating markdown as a reflex and start treating it as a planned line.

Two rules carry most of the benefit: cap the opening cut shallow (10-25%) and trigger it on sell-through, not the calendar.

A small early markdown on a slow SKU beats a deep late fire sale.

And never blanket-discount a whole drop; a sitewide sale gives away margin on the units that would have sold at full price anyway.

The deepest fix, though, is upstream: the brands that discount least are the ones that didn’t over-buy. Broken-size runs and over-ordered styles are what force the markdown in the first place, which is the next two leaks.

Upstream Leak 04:

The wrong size curve:

Symptom: sold-out mediums, stranded XS/XXL
Root cause: ordering a flat curve

This is the leak almost no first-time founder sees coming, and it’s a direct cause of forced markdowns.

Here’s the trap: you order your run-flat in equal quantities of S, M, L, and XL because it feels fair and it’s easy to brief.

But demand isn’t flat. It’s a bell.

So your mediums and larges sell out in a week.

Those are lost sales customers who wanted the size you ran out of.

Meanwhile, your smallest and largest sizes sit because you made as many of them as your best sellers.

Those become clearance.

You lose at both ends of the same order, and the clearance you take on the tails is what drags your blended margin down.

Size Flat order (you) Real demand (typical) Outcome
S 25% ~15% Overstock → clearance
M 25% ~32% Sells out → lost sales
L 25% ~33% Sells out → lost sales
XL 25% ~20% Roughly balanced

Real curves vary by fit, audience, and gender mix; oversized streetwear skews larger than fitted.

How to plug it:

Briefly, a graded curve, not a flat one.

If you have past drops, use your own sell-through by size.

If you don’t, start from a realistic bell curve weighted to your fit, and keep the tails (XS, XXL) deliberately short on run one; you can always add them in a reorder once you’ve seen who’s actually buying.

A flat order is a markdown you’ve pre-ordered.

Sizing sits inside a bigger decision: what to make at all.

See how to define your product range for run one to keep the whole order tight enough that a bad curve can’t sink it.

Upstream Leak 05:

Sampling & decoration creep:

Symptom: the quote keeps climbing
Root cause: changes after the price was set
 

This leak is small per event and brutal in aggregate.

It’s the cost of changing your mind after the clock started, and it hides because each change feels reasonable in isolation.

Sampling creep:

Every sample round costs money and time.

The prototype is expected.

The problem is the fourth and fifth rounds are chasing a fit you could have nailed in two if your tech pack had been complete.

Each extra round adds a sample charge, courier fees both ways, and one to two weeks that push your drop closer to a season you’ll have to discount.

Decoration creep

Decoration is priced on inputs you keep adding.

Method What drives the cost up? Quiet add-on
Screen print Number of ink colors Each extra color = another screen + setup
Embroidery Stitch count A denser logo can double the run rate
Placement Number of print locations A front, a back, and a sleeve are three setups, not one.
Specialty Puff, foil, metallic, oversized Slower line speed, higher reject rate

None of these are scams.

They’re real machine times.

But a design that drifts from a 2-color chest hit to a 5-color front, back, and sleeve package has quietly tripled its decoration cost, and if that happened after you set your retail price, the gap comes straight out of margin.

How to plug it:

Freeze the design before you price it.

Lock colors, placements, and method, then get the quote and treat any post-quote change as a budget decision, not a creative one.

And invest in a complete tech pack so sampling converges in two rounds, not five.

The cheapest sample round is the one you didn’t need.

Go deeper on rounds, costs, and sign-off in the streetwear sampling process, and compare print vs embroidery vs DTG economics in choosing a decoration method.

Upstream Leak 06:

Hidden factory surcharges:

Symptom: final invoice > quoted unit price
Root cause: per-unit price isn’t the full price.
 

The unit price on a quote is a headline, not a total.

Legitimate surcharges live underneath it, and a founder reading only the per-piece number budgets for a landed cost that doesn’t exist.

None of these are a factory cheating you.

They’re the cost of how your order is shaped.

Surcharge Triggered when… Plug
Low-MOQ uplift You order below the price-break quantity. Consolidate styles to share a minimum.
Dye lot / color Each colorway needs its own fabric dye lot Limit colorways on run one.
Fabric overage You’re billed the full knit/dye lot, not what the cut used. Design the lot, or bank the surplus for reorders.
Split-size/grading A wide size range adds pattern and marker work. Keep the range tight early.
Rush/priority You compressed the calendar. Book earlier and buffer the timeline.

The fabric overage one surprises people the most.

Custom knit and dye lots carry minimums often measured in hundreds of kilos per color.

If your run only consumes part of a lot, you’re frequently billed for the whole thing.

That surplus fabric isn’t lost if you reorder the same color, but on a one-off drop, it’s pure leak.

How to plug it:

Ask for the all-in landed cost per unit, in writing, before you commit, not the FOB unit price.

Make the factory list every surcharge that applies to your order as specified.

A good manufacturer will do this gladly; it’s how you both avoid a surprise on the final invoice.

Upstream Leak 07:

Rework & quality failures:

Symptom: defects, remakes, delays
Root cause: no agreed quality standard
 

Rework is the cost of doing it twice.

A bulk run comes back with crooked prints, loose seams, or off-shade panels, and now someone pays to fix or replace them in money, in time, or in units you quietly write off.

It’s a leak that compounds, because a quality failure caught late also delays your drop into the discount window.

The root cause is almost always the same: nobody defined what “acceptable” means before production started.

Without an agreed standard, every defect becomes an argument, and arguments cost time you don’t have.

AQL

The acceptable quality limit is the industry’s shared language for “how many defects are too many.”

Agreeing on an AQL standard up front and the inspection that enforces it turns a vague dispute into a clear pass/fail you can both act on.

How to plug it:

Two moves.

First, seal a golden sample, the approved reference every bulk unit is judged against.

Second, agree on an AQL and a pre-shipment inspection in your terms so quality is a checkpoint, not a hope.

Rushing sampling to save a week is what causes most bulk-quality failures; the saved week comes back as rework, late.

In-transit Leak 08:

Freight & landed cost:

Symptom: shipping eats the unit margin.
Root cause: small runs + the wrong mode.
 

Freight is the leak that punishes small orders hardest.

A sea container spreads its cost across thousands of units, so per-unit freight is trivial.

A small first run can’t fill one, so you’re choosing between a part-container rate, consolidation, or air.

And air freight on a panicked timeline can cost several times the sea rate per unit, quietly turning a healthy margin thin.

Then there’s the gap founders forget entirely: the difference between the factory price and the landed price.

Your true cost per unit isn’t what leaves the factory; it’s that plus freight, insurance, duties, customs clearance, and any broker fees, divided across units that actually arrive sellable.

Check our full guide on freight.

Cost layer Often forgotten because…
Sea vs. air freight The quote was in factory (FOB) terms only.
Import duty/tariff The rate depends on the garment’s classification code.
Customs clearance + broker A flat fee that stings hardest on small runs
Last-mile to your 3PL Treated as separate from “shipping”

Duty is the one most likely to blindside a first-time importer, because the rate hangs on how the garment is classified and where it’s coming from, and those rates move.

Tariff, duty, and customs figures change and depend on your destination country, the garment’s classification, and trade agreements in force. Confirm your specific duties with a licensed customs broker before you budget.

How to plug it:

Budget in landed cost per unit from day one, and get a freight quote before you finalize run size.

Sometimes ordering slightly more to hit a better freight bracket is cheaper per unit than the smaller order plus air.

Book early enough that sea is an option; air freight is usually a tax you pay for a calendar you didn’t protect yourself from.

Freight folds into the same landed-cost picture as your factory surcharges plan them together, never separately.

The tool:

The Leak-Audit Checklist:

Run this against your next order before you sign anything.

It’s organized by room upstream first, because that’s where the Upstream Rule says the cheapest fixes live.

Tick each box only when you have a real answer, not an assumption.

Upstream before production:

The size curve is graded, not flat:

Built from past sell-through or a realistic bell weighted to your fit, tails kept short on run one.

Design is frozen before pricing:

Colors, print placements, and decoration methods were all locked before you accepted the quote.

The tech pack is complete:

Detailed enough to converge sampling in ~2 rounds, not 5.

The all-in landed cost is quoted in writing:

Not the FOB unit price, but the per-unit cost with every applicable surcharge listed.

Surcharges are itemized for your order:

Low-MOQ uplift, dye-lot minimums, fabric overage, and rush fees are all named.

Colorways are deliberate:

Each one carries its own dye-lot cost; you’ve limited them on the first run.

Run size matches demand confidence:

Sized to the cash you can risk, with a reorder path, not the factory minimum.

Quality & transit:

The golden sample is sealed:

One approved reference every bulk unit is judged against.

AQL standard + inspection agreed:

Written into terms, with a pre-shipment check that enforces it.

Freight quote secured before final run size:

Mode chosen (sea vs. air) and per-unit freight known.

Duty confirmed with a broker:

Classification and rate checked for your destination, not guessed.

Downstream after it lands:

Size chart built from your real garments:

Published measurements, not a borrowed template, are your biggest lever on fit returns.

Reverse-logistics cost is known per return:

You can state what one return actually costs you, end to end.

A markdown plan exists before launch:

Shallow opening cut (10–25%), sell-through trigger, style-level not sitewide.

Carrying cost is on the books:

You’re tracking what aged stock costs in storage and tied-up cash.

If you tick only three

Make them the graded size curve, the all-in landed quote, and the markdown plan.

Those three sit at the top of the three biggest drains, and the first two are upstream, where a fix is worth the most.

From the manual

This is one chapter of The Complete Streetwear Manufacturing Guide

These chapters cover taking a streetwear brand from idea to shipped units, costing, sourcing, tech packs, QC, freight, and launch. All written from the Sialkot factory floor.

Open The Full Guide

Questions founders ask

What's the single biggest place clothing brands lose money?

For most small brands, it’s the discounting they were forced into, but that’s a symptom. The real culprit usually sits upstream: over-ordering and a flat size curve create the unsold stock that requires the markdown.

Fix the order, and the discount mostly disappears.

That’s the upstream rule in one line.

Because the refund is only the sale reversing, you’re back to where you started on that unit.

The loss is everything else: reverse shipping, inspection, repackaging, restocking labor, and the chance the unit comes back unsellable.

Processed fully, a single return can eat a large share of the item’s original price, which on tight apparel margins can wipe out the profit on the sale entirely.

Start from a realistic bell, weighted to your silhouette; boxy streetwear skews larger than fitted cuts, and keep the extreme sizes (XS, XXL) deliberately short on the first run.

Then let the first drop tell you the truth and correct the curve in your reorder.

A flat curve feels fair, but it’s effectively a markdown you’ve pre-ordered.

Ask for the all-in landed cost per unit in writing, and have them name every surcharge that applies to your order as specified: low-MOQ uplift, dye-lot minimums per color, fabric-lot overage, size-grading, and any rush fee.

A good factory itemizes these without being chased; it protects both sides from a surprise on the final invoice.

It’s proportionally largest exactly when your run is smallest, because freight and clearance costs don’t scale down neatly.

A small order can’t fill a sea container, and air freight on a rushed timeline can cost several times more per unit.

Sometimes ordering slightly more to hit a better freight bracket lowers your per-unit cost.

Always budget in landed cost, not factory price.

Industry analysis suggests the difference between disciplined and reflexive discounting is worth several hundred basis points of gross margin purely from changing how you discount, not whether you do.

Shallow-and-early beats deep-and-late, and style-level cuts beat sitewide sales that give away margin on units that would have sold at full price.

Rarely. Most of these leaks are recovered by ordering better, not pricing higher; a tighter run; a graded curve; a frozen design; a landed-cost budget; and a markdown plan.

You’re not asking the customer for more; you’re keeping more of what they already pay you.

On the floor · Sialkot

Written by

Faizan Ahmad

Chief Apparel Technologist & Head of Manufacturing, Gibben Clothing · Sialkot, Pakistan

Faizan leads production at Gibben Clothing, a cut-and-sew streetwear manufacturer in Sialkot, with 8+ years turning raw yarn into retail-ready hoodies, tees, bottoms, jackets, tracksuits, and headwear. He doesn’t just write about clothing; he works the floor, so every guide here is grounded in real fabric behavior, QC standards, and production data from live runs.