How to Start a Clothing Brand in 2026:
A Step-by-Step Guide

What the first year of a clothing brand actually costs you

The honest answer: it costs time before it costs money, and then it costs money in a lump. Nine to twelve months from first sketch to first shipment is a normal, well-run timeline — not a slow one.

Three costs surprise first-time founders every single time. The first is sampling (the paid rounds of prototypes before bulk production), which almost always takes more rounds than planned. The second is landed cost (the true per-unit cost once freight, duty, and fees are added), which is not the number on the factory quote.

The third is returns. Retailers forecast a 15.8% overall return rate and 19.3% on online sales (NRF & Happy Returns), and apparel sits at the high end of that range because fit is the hardest thing to judge from a product page. Every point of return rate comes straight out of the margin you costed in Step 6.

“A first collection is not a creative project with a budget attached. It's a costing exercise with a creative brief attached.”

None of that is a reason not to start. It is a reason to start in order — and the failure patterns are consistent enough that we mapped them separately in the four operational mistakes behind most US clothing-brand failures.

Step 1 — Pick a niche you can say in one sentence

A workable niche names a specific customer, a specific problem, and a specific product category. “Elevated basics for everyone” is not a niche. “Heavyweight cotton tees cut for men over 6'2" who can't find length” is.

The test is repeatability. If a customer can describe your brand accurately to a friend in one sentence, your positioning works and your marketing gets cheaper. If they can't, you will pay for every single customer, forever.

Narrow also fixes a production problem. A tight niche means fewer styles, fewer fabrics, and fewer size curves — which means lower minimums, faster sampling, and a cost sheet you can actually understand.

“A niche you can say in one sentence is a niche a customer can repeat to a friend. Everything else is a mood board.”

Step 2 — Build a range plan, not a mood board

A range plan (the list of styles, colourways, and sizes you will actually produce) is the first document with money attached. It converts a Pinterest board into a number of units.

Keep the first drop to three to five styles. Every extra style multiplies through fabric minimums, sampling rounds, grading, photography, and inventory risk — and it is the fastest way to turn a launch budget into a warehouse.

Then decide the manufacturing route, because it changes everything downstream. Private label versus custom manufacturing is the real fork, and beginners are rarely shown the trade-off side by side:

Step 3 — Register the entity and clear the name before you print it

Form the business entity, get an EIN, open a business bank account, and register for sales tax where you have nexus. That part is administrative. The part that ends brands is the name.

Search the USPTO register before you buy the domain, and file early. Since the January 2025 fee restructure, a US trademark application costs a $350 base fee per class, with surcharges that stack — $100 per class for insufficient information and $200 per class for writing your own free-form description instead of using the ID Manual (Federal Register; USPTO fee schedule).

Apparel is Class 25. Filing before your first labels are woven costs a few hundred dollars; discovering a conflict after 500 garments carry the name costs the run.

Step 4 — The tech pack is the contract

A tech pack (the technical specification a factory manufactures from) carries the flat sketch, the graded measurement chart, the fabric and trim specification, stitch types, seam allowances, label placement, and packing instructions. It is the difference between ordering a garment and describing one.

Send a sketch and a reference photo, and the factory will fill every gap with the cheapest reasonable option — different thread, different interlining, a hem that reads wrong. That is not sabotage. It is what an unspecified field means.

A tech pack is also the only way to compare quotes honestly. Two factories quoting from the same spec are quoting the same garment; two factories quoting from a photo are quoting two different products. We broke the document down field by field in what a tech pack is and why you can't manufacture without one.

“Everything you don't write down, someone else decides for you — usually the cheapest way.”

Step 5 — Choose fabric by spec, then sample until the fit is boring

Fabric is chosen on four specifications, not on how a swatch feels in a showroom: composition, GSM (grams per square metre — the weight of the cloth), construction, and finish. Those four decide drape, durability, shrinkage, and roughly half your cost sheet.

Judge cloth the way a buyer does — off the roll, against the light, at the weight you'll actually produce in. A square of fabric on a table tells you almost nothing about drape.

Ask for a wash test before you commit. A jersey that loses 6% in length after three washes will generate returns that no photography fixes — and the return, not the garment, is what your customer remembers. Our fabric sourcing guide walks the selection criteria in detail.

Then sample. Expect two to three rounds: a proto for the pattern, a fit sample on a real body in your base size, and a pre-production sample that is signed off and becomes the quality benchmark for the whole run. Approve nothing on a photo.

Step 6 — Cost the landed price, not the factory quote

This is the step that separates brands that survive from brands that are quietly unprofitable at full price. The factory quote is an input. Landed cost is the number you price from.

2026 made this unavoidable. CBP indefinitely suspended the $800 de-minimis exemption on 24 June 2026, so low-value shipments now require formal or informal entry and carry duty, taxes, and fees (Federal Register). The “ship it direct and skip the paperwork” model that a lot of launch advice still assumes is gone.

Here is a 500-unit run of cotton tees, classified under HTS 6109.10.00, which carries a 16.5% general duty rate (USITC Harmonized Tariff Schedule):

The factory quote is 77% of the real cost. A founder who prices at 4× the quote thinks they have built a $48 tee at a healthy margin; they have actually built a $48 tee on a $15.58 cost, before returns, shipping, payment fees, and acquisition. Price off the wrong number and the gap never shows up until the third month of settlements.

Treat those figures as a worked structure, not a quote. Duty depends on the exact classification, the country of origin, and whatever trade actions are live when your goods clear — get the HTS code confirmed for your specific garment before you commit to a retail price. Lead times move too, which is why the design-to-doorstep timeline belongs in the same spreadsheet.

“The factory quote is about three-quarters of what the garment actually costs you. Price off the quote and you've already given the margin away.”

Step 7 — Labels are federal law, not decoration

Every garment you sell in the US needs three things on a permanent label: fibre content, country of origin, and the identity of the manufacturer or the business responsible — either a company name or a registered identification number (FTC, Threading Your Way Through the Labeling Requirements).

A fourth requirement comes from a separate rule. The FTC's Care Labeling Rule requires permanent care instructions on textile wearing apparel, and the brand must have a reasonable basis for the instruction it prints (16 CFR Part 423).

Put all four in the tech pack in Step 4. Adding them after the run means relabelling every unit by hand, and that cost lands the week you were meant to launch.

Step 8 — Launch small, sell through, then scale

Order the smallest run your economics tolerate. Unsold inventory is the most expensive mistake in apparel because it consumes the cash you need for the reorder of the thing that did sell.

Then measure two numbers before designing anything new: sell-through rate (the percentage of a run sold in a set window) and return rate by size. Sell-through tells you whether the product works; returns by size tell you whether the grading does.

The counterexample — when to break this sequence. If you already have an audience, run it backwards: take pre-orders or run a small made-to-order drop from the sampling stage, and let real demand size the first production run. Founders with distribution should validate before they invest in tooling; founders without distribution should build the product first. Choose based on which one you actually have, not on which is more comfortable.

“Sell through the first run before you design the second. That one rule prevents most inventory graveyards.”

The 12-month build order, in one view

Bands overlap in practice — sampling runs while the trademark sits in examination, and photography is booked while goods are on the water. What does not overlap is the dependency chain: no fabric decision without a tech pack, no cost sheet without a fabric decision, no retail price without a cost sheet.

The terms you'll hear, in plain English

Most of the confusion in a first manufacturing conversation is vocabulary, not concepts. These are the words that will come at you, and what each one actually means.

Tech pack

The technical specification a factory manufactures from — sketches, measurements, fabric, trims, stitches, labels, packing. Your contract with the maker.

GSM

Grams per square metre: how heavy the cloth is. A 180 GSM tee is light and summery; 240 GSM reads premium and structured.

MOQ

Minimum order quantity — the smallest run a factory or mill will accept, usually set per style, per colour, and per fabric.

FOB

Free on board: the price of the goods loaded at the origin port. It excludes freight, duty and every fee after that point.

Landed cost

What a unit truly costs once freight, duty, customs fees, brokerage and inbound logistics are added. The number you price from.

HTS code

The Harmonized Tariff Schedule classification that sets your US duty rate. Cotton knit tees sit at 6109.10.00.

Grading

Scaling an approved base-size pattern up and down into a full size run while holding the fit relationships.

Proto / fit / PP sample

The three sampling rounds: proto proves the pattern, fit sample is worn on a real body, pre-production sample is signed off as the quality benchmark.

Colourway

One colour version of a style. Each colourway is a separate production line item and its own inventory risk.

Sell-through rate

The percentage of a production run sold within a set window. The number that tells you whether to reorder or stop.

Lead time

Elapsed time from placing the bulk order to goods arriving. Fabric availability, not sewing, is usually the long pole.

3PL

Third-party logistics: the warehouse that stores your inventory and ships orders to customers on your behalf.

The bottom line

Starting a clothing brand in 2026 is not a creativity problem. It is a sequencing problem: the founders who get to a second run are the ones who refused to move to the next step until the current one produced a document they could point at.

If we were in your shoes, we'd spend this week on Steps 1 and 6 — the sentence and the cost sheet — because those two decide whether everything in between is worth building. So: can you say your niche in one sentence, and do you know your landed cost per unit? If either answer is no, that is where the next month goes.

FAQs