Why I stopped guessing at size before producing more clothes

When I started a small shirt shop, I thought the hard part was the graphic. The shirt looked clean on the screen, the printer matched the color I wanted, and the fabric felt nice. My first repeat customer wrote back that the medium pulled tight across her shoulders and the sleeve cap sat too high. That note changed my habits. The issue was fit. After that, I stopped trusting a flat spec sheet from one supplier and began testing how a garment behaves on a body that moves.

Before I order more units, I ask what fits the customer I have, not the customer I imagine. The note I saved from roulens what fits helped me ask whether a garment matches the way a real buyer stands, reaches, and washes the item. I use it when a pattern feels almost right, because almost right costs money.

A sample order can lie

A single sample can hide a range of failure. I once received a shirt that looked perfect on a small sample. The second run used the same artwork, but the medium shoulder was half an inch short. The cause came from a small change in seam allowance and a stitch that pulled more than expected. I learned to ask for two samples in different sizes from the same production lot when a style matters to repeat buyers. If a vendor says the first sample proves the fit, I move on.

Body movement beats flat measurements

Flat measurements are easy to copy into an email. They do not tell me what happens when a person reaches up, lifts a child, or sits for six hours. For one relaxed henley, the chest sat fine while the model stood still. The moment he crossed his arms, the front pulled. That pattern needed more room around the underarm, not a bigger overall size. I began keeping a short list of motion tests for each style: reach, cross, bend, stretch. It sounds plain, but it saved several hundred units from dead stock.

Return rate is a fit signal

Small shops often treat returns as an administrative problem. I started reading them as product data. The words inside a return note are useful too. A customer might pick medium because the chest measurement looked close, then complain that the sleeve length felt short. That tells me my chart favors numbers over body shape. It also tells me the style may fit a different size group than my marketing photos suggest. Once I changed one product page image to show a side view with arms relaxed, the “size ran small” comments dropped in the next month.

A good vendor asks uncomfortable questions

The right supplier does not smile when you hand them a sketch and ask for a quick quote. They ask about target weight, stretch, wash behavior, hem depth, and where the body moves most. One factory owner asked why my graphic sat two inches above the waistband. I did not have a good answer, only taste. That question led to a test fit on three body types. I learned to pay for vendor curiosity. A vendor who cares about fit can save me from a pile of shirts that look correct and feel wrong.

Write the fit rule down before you order

My process is boring on purpose. Before a production run, I write one sentence that defines the fit promise. Then I attach a small photo or sketch that proves the customer can move in it. A rule like no pinching at the armpit during a normal reach beats a vague note like keep the shirt comfortable. If I cannot test it, I cannot ship it. For repeat buyers, that written check is part of the product, not extra paperwork.

• Measure the garment on a person who matches your size group, not a dress form only.
• Track the first ten returns by size, height, and comment word.
• Ask the factory to show how seam placement affects movement.
• Test the sample after one wash, not before it leaves the printer.

For my shop, clearer fit notes and partners who test movement are the difference between a small batch that sells and a stack of shirts that sits in the back office. I still care about design, but design has to pass the body. The work feels slower at the start. It pays off when a customer writes back because the shirt fit the way the page promised.