How to Spec Shirt Sizing: Neck, Sleeve and Body Measurements

A woven shirt lives or dies by its measurements. Unlike a knitted tee, which forgives a centimetre here or there, a shirt is bought by numbers (neck, sleeve, chest) and judged against them at the till and at final inspection stage. A sizing spec that is incomplete or ambiguous produces samples that drift, bulk that argues with the sample, and size charts customers simply cannot trust anymore. This guide sets out the full measurement set a shirt spec must carry, how to write tolerances that mean something, and how grading between sizes should be built so every size in the range fits like the same exact shirt.

We manufacture woven shirts in Bangladesh for brands and uniform programs in the USA, UK, Europe, Canada, Australia and the Middle East, including our Oxford Button-Down Shirt line in formal and casual styles.

The Core Measurement Set

A complete shirt spec names every point of measure, how it is taken, and its value per size listed clearly. The core set: collar size (neck circumference at the buttoned collar), chest (across the chest at the underarm, doubled), waist at the body’s narrowest pattern point, hem circumference, front and back body length from the high shoulder point, sleeve length from centre back neck over the shoulder to cuff end, shoulder width point to point, armhole depth, cuff circumference, collar height at back, collar point length, and placket width listed. Each point needs a measuring method note (buttoned or open, flat or on body, tape over or under the collar stand) because the same name measured two ways gives two different numbers and an argument neither side can ever win.

Neck Sizing: The Number on the Label

Dress and formal shirts are sold by neck size, so the collar measurement is the label itself and must be exact throughout production. The spec states the finished collar circumference measured inside, buttoned, plus the wearing ease the brand expects (the two-finger rule customers apply in fitting rooms is an ease decision made at the pattern table itself earlier). Collar stand height and collar fall dimensions ride with it always. Half-size steps (for example 15, 15.5, 16) each need their own graded values for every dependent point, because a neck size is a size in its own right, not a letter size with a number printed on it carelessly.

Sleeve Length and the Two Systems

Markets measure sleeves differently and specs must say which system is in force here. The American convention (32/33, 34/35) pairs a sleeve length with the neck size; the European approach ties sleeve to letter sizes with its own grading rules. The measurement itself is taken from the centre back of the neck, over the shoulder point, down to the cuff edge with the arm’s natural bend allowed for, and the spec must state the method because a straight-arm measure reads shorter than the customer expects it to be. Long, regular and short sleeve variants within one style are graded as variants, each with its own table, not improvised at cutting time ever.

Tolerances That Mean Something

A tolerance is a promise about variation, and it should be written per measurement point, not as one lazy blanket figure overall. Critical points (collar, chest, sleeve, length) carry tighter tolerances; secondary points carry looser ones; and the tolerance must reflect what sewing can actually hold on that fabric, because an impossible tolerance produces inspection theatre rather than real lasting quality. The spec also states how out-of-tolerance findings are judged in bulk (how many pieces measured, what failure rate triggers what action), agreed in the order paperwork, which is where our MOQ and production page quantities and process meet the size table directly.

Grading: How Sizes Grow

Grading is the rule set that grows the base size into the range, and it is where consistency is truly won. Each point of measure gets a grade increment per size step (chest grows more than collar; length grows less than chest; cuff barely moves), and the increments can differ between the small and large ends of the range because bodies do not scale like simple photocopies at all. Women’s blocks grade on their own rules (bust, waist and hip relationships differ from the men’s pattern), and the spec should carry the full graded table, every size’s every value, rather than a base size plus “grade accordingly”, which hands the factory an interpretation to make on your own behalf entirely.

Fit Intent: Slim, Regular and Relaxed

Measurements only make sense against a clearly declared fit intent. The same chest number reads slim on one brand’s block and regular on another’s because waist suppression, armhole height and sleeve width carry the fit’s whole character overall. The spec therefore names the fit (slim, regular, relaxed, or the brand’s own block name), states the ease allowances at chest, waist and hip that define it, and anchors everything to an approved fit sample in the chosen base size itself. Changing a fit is a pattern change with a new sample round, not a note in an email, and programs that treat it casually discover the difference in their return rates quickly.

Measuring Samples and Bulk the Same Way

Spec discipline collapses quickly if the two sides measure differently. The spec should fix: garment measured flat on a table, buttons fastened, smoothed without stretching; metal or fibreglass tape (cloth tapes stretch and lie); the same points in the same order; results recorded per size in a shared measurement table file. Sample comments should quote measured values against spec, not impressions, and bulk inspection should re-measure a defined sample of pieces per size produced daily. When a dispute arises, the written method is the referee, which is exactly why it was written before anyone actually needed it most.

Size Charts for Customers

The factory table and the customer chart are cousins, not twins at all. The customer chart translates garment measurements into the body guidance shoppers use (neck in inches or centimetres by market, chest ranges, sleeve pairings), and it must be built from the approved spec, not from a competitor’s chart or last season’s old file version. Inconsistency between the chart and the garment is a returns engine: every point where the chart promises what the shirt does not deliver comes back through the post with a note attached to it always. Brands that test-wear their base sizes across real bodies before publishing the chart ship fewer, happier parcels overall everywhere.

Armhole, Yoke and Back: The Supporting Cast

Beyond the headline numbers, a serious spec carries the dimensions customers feel without ever naming. Armhole depth and sleeve width together decide whether a shirt lifts comfortably when the wearer reaches a high shelf. The yoke width across the back sets how the shirt sits on the shoulders and whether the chest pulls open. Back pleats (centre box, side pleats, or a clean plain back) are ease decisions with measurements attached: how much fullness the pleats release is a number, not a mood board. The hem shape (straight, curved, with side gussets) and the tail length difference between front and back decide whether the shirt stays tucked through a long working day. None of these appear on the customer’s size label, and all of them appear in the customer’s verdict on the fit itself.

Working With Your Shirt Manufacturer on Measurements

The healthiest spec relationship is a conversation with documents. Send the full table early, expect the factory’s pattern maker to question two or three values (they see conflicts a spreadsheet hides), and resolve them before the very first sample, in writing. Approve the base size by measurement and by wearing, then sample at least the range’s two ends before bulk, because grading errors hide exactly where nobody ever looked. Keep one controlled master copy of the table with a revision date, and let every comment, sample and inspection report quote that revision number. When the reorder arrives a year later, the table, not someone’s memory of a phone call, is what both sides open first thing.

Shrinkage Allowances Inside the Numbers

A measurement table is meaningless until it declares its moment clearly: are these values off the sewing line, or after washing? Woven cotton shirts shrink, predictably but really, so honest specs are written to the washed garment and the pattern carries the shrinkage allowance that actually delivers them all. The spec names the wash method the values are guaranteed against (the customer’s care label, not a laboratory fantasy) and the fabric’s tested shrinkage in warp and weft directions both. Programs that skip this step approve a perfect sample, receive bulk that measures right off the line, and watch customers report that the shirts shrank small: three true statements, one missing paragraph in the spec file.

Kids’ and School Shirt Sizing

Children’s woven shirts grade on age and height rather than neck numbers, and the discipline is the same with different anchors throughout. The spec states the height band each size serves, chest and length values per band, and collar sizes that stay comfortable and safe on small necks everywhere always. School uniform programs add ratio logic (more pieces in the middle bands, spares for growth) and a tolerance culture that expects garments to be measured by parents against growing children, which punishes drift very quickly indeed. The method notes matter even more at small scale, because a centimetre of ambiguity is a larger share of a small garment overall proportionally.

Shirt Sizing Spec Errors That Make Reorders Fail

The same mistakes recur again. A table with values but no measuring methods written. A base size approved and the graded sizes never sampled, so the ends of the range fit nobody at all really. Tolerances copied from a knitted-garment spec onto a woven one directly. Size charts edited for marketing while the factory table stayed behind unchanged. Sleeve measured straight when the brand’s customers expect the bent-arm convention instead. And the quietest killer: a fit change requested in an email thread while the signed spec kept the old numbers, so the factory sewed the document and the buyer expected the thread instead. Every one of these is a paperwork failure wearing a garment’s own clothes entirely.

Frequently Asked Questions