How to Build a Size Chart for Your Clothing Brand

A size chart looks like the least interesting document in a range and it is quietly one of the most expensive to get wrong. Returns attributed to quality are frequently sizing failures, and sizing failures are almost always decisions nobody made deliberately.

Here is how to build one that works, in the order the decisions have to happen.

Body measurements and garment measurements are different documents

This is the distinction that causes the most confusion, and getting it clear removes half the problem.

A body measurement chart says what size a person is: chest 102 cm, waist 88 cm. It is what your customer uses to pick a size, and it belongs on your website.

A garment specification says what the finished garment measures, flat, at each point: chest 56 cm across, body length 74 cm. It is what the factory builds to and what inspection checks against, and it belongs in your tech pack.

The difference between the two is ease — the room designed into the garment. A base layer might have negative ease; a shell worn over a mid-layer might have 15 cm or more. Confusing the two is how a jacket meant to fit over a fleece ends up fitting a bare chest.

Decide the market before the numbers

US, EU and UK sizing are not the same and are not interchangeable. A garment graded for one market and sold into another produces returns that look like a quality problem and are a specification problem.

Two things worth deciding up front:

• Which market is your primary one? Grade for that, then produce a conversion table for the others rather than trying to average across them.
• Are you selling into a region with its own convention? Gulf markets, for example, commonly label in EU sizes while also recognising UK and US, and buyers there frequently specify longer body length and sleeve than a standard Western block. That is a tech-pack conversation, not a labelling one.

Pick the points of measure, and show them

Name every measurement and show it on a diagram. "Sleeve length" is measured from at least three different origins depending on who you ask — centre back, shoulder seam, or armhole — and a chart without a diagram is an invitation to build the wrong garment.

For an upper-body garment the working minimum is:

• Chest, measured flat one inch below the armhole
• Body length from high point shoulder
• Across shoulder
• Sleeve length, with the origin stated
• Cuff opening
• Hem or sweep
• Neck opening

For trousers: waist relaxed and extended if there is elastic, hip, thigh, knee, leg opening, inseam, and front and back rise stated separately. Rise is the measurement customers cannot name and return garments over — a short back rise pulls the waistband down every time the wearer bends, and the complaint arrives as "the waist is too big".

Grade properly — it is not a percentage

Grading is the step most often done badly by people scaling a base size up and down uniformly. Different measurements grow at different rates: chest grows faster than neck, and body length grows far more slowly than either. A range graded by multiplying everything by the same factor produces XXL garments that are absurdly long and XS garments that are cramped in the body.

State the grade rule — the increment per size for each point of measure — in the tech pack rather than only listing the sizes. It makes the intent explicit and it lets a factory extend the range later without guessing.

Set tolerances, or you cannot hold anyone to anything

No garment is made to the millimetre. Fabric moves, cutting has variance, stitching has variance. What matters is agreeing how much variance is acceptable at each point.

Tolerances are normally tighter on the measurements customers notice — chest and body length — and looser on hems and cuffs. Put them in the tech pack next to each measurement. A tolerance that is written down is a tolerance you can enforce; without one, a dispute is two opinions.

Test it on real bodies before you commit

A chart validated only on a table will pass and still fail in the market. Get the sample onto people of the right size and have them do what the customer will do — raise both arms, layer it, sit down, wear a pack, put the hood up over a cap.

Then look at where it binds. A garment that measures correctly and restricts the shoulder has a pattern problem rather than a sizing problem, and the two are fixed differently.

Publish it, and publish it usefully

Once it works, the chart is a sales asset rather than an internal document.

• Show how to measure, with a diagram, not just the numbers. Most returns come from customers measuring differently than you intended.
• Give conversions for the markets you sell into.
• Give both body measurements and finished-garment measurements. Experienced buyers want the second, because it lets them compare against a garment they already own and like.
• Version it. Put a reference and a revision letter on the document, so when it changes there is no ambiguity about which one an order was built to.

Our own published charts show the format we work in — points of measure lettered against a diagram, conversion columns, and both body and finished-garment tables: UAE and Gulf sizing, cargo trousers and safari shirts. They are a reasonable starting point to mark up rather than starting from a blank page.

The shortcut, if you have one garment that fits

If you already own a garment your customers love, the fastest route to a good chart is to measure it flat, point by point, and send those numbers with the garment itself. That single step removes more ambiguity than a page of description, and it is how most of our first samples land close on the first round.

Our standard run is S–3XL and grading shifts to match the market you name. If you do not have a chart yet, we will provide our block as a starting point — but a chart built from a garment your customers already like will beat anyone's standard block.