A shirt with a 40-inch chest, 26.5-inch back length, and 17-inch shoulder width gives flat pattern measurements, while the shopper experiences fabric around a three-dimensional body. A pattern cutter records these figures at the factory from cut fabric laid flat. They accurately describe the pattern, but not the wearing experience on a real body.
Flat specs and lived fit are different, which makes sizing difficult. Treating a garment measurement as a body measurement leads fit systems toward poor recommendations. To read specs usefully, you have to follow how a flat pattern becomes three-dimensional fit on a body.
Ease: the designed space between body and garment
Ease is the difference between a body measurement and the garment measurement at a given point. A structured blazer with a 40-inch chest may be designed for a 37-inch chest, leaving 3 inches for movement and layers. A fitted t-shirt measuring 38 inches at the chest may also be designed for a 37-inch chest, leaving 1 inch for a slim silhouette.
Ease reflects design intent, not measurement error. A 37-inch chest can fit both garments correctly, yet the blazer feels relaxed while the t-shirt feels fitted. Neither fit is incorrect. Comparing body and garment chest measurements alone would label both garments the same size for this shopper, a technically accurate but practically misleading result.
Direct garment-to-body comparisons have a major limitation. Fit recommendations must account for each garment's intended ease, or interpret its measurements in relation to the shopper's particular body geometry and wearing experience.
How fabric stretch changes the measurement equation
Woven fabrics have no stretch. A woven shirt's chest spec therefore predicts the largest chest circumference that can pass through without distortion. If body measurement plus required ease is greater, the shirt will fit too tightly.
Knit fabrics stretch, so their chest spec is taken without fabric tension and shows the unstretched dimension. The garment can fit a larger body measurement until its stretch limit. A knit top listed at 36 inches flat may suit a 40-inch chest when the fabric stretches 15% before distortion. The same 36-inch spec in lower-stretch fabric may suit only a 38-inch chest.
For stretch garments, prediction depends on recoverable fabric stretch and its relationship to body measurement, not the flat spec alone. A flat-measurement-only system will over-recommend size-ups in stretch categories and understate the fabric's accommodation.
That is why our fit spec records fabric composition and stretch data with garment measurements. Flat measurements work reasonably well for woven garments. In stretch categories, stretch range changes the prediction.
Proportional fit points: where flat specs become complex
Some fit points translate well from flat measurements. Chest circumference at the widest point is one example. Others do not, because the shopper's experience depends on three-dimensional geometry rather than one flat dimension.
Shoulder fit shows the issue clearly. A shoulder spec is the flat, point-to-point distance across the back shoulder seam. It describes the garment, not where that seam will land on the shopper. Placement depends on garment shoulder width, body shoulder width and slope, and armhole position. Two shoppers with the same shoulder width can have different seam placement when their shoulder slopes differ.
This is where a three-dimensional body model adds detail that flat specs miss. From two photos, we capture shoulder width, slope, and its spatial relation to the armhole. Comparing that geometry with the garment's shoulder construction gives a more useful prediction than comparing two flat numbers.
What manufacturers measure, and what they leave out
A technical spec sheet is built for production control, not primarily for fit prediction. Manufacturers measure points that support pattern consistency and quality checks, including seam allowances, hem depths, collar stand heights, and cuff circumferences. These details say little about whether a garment fits a particular body.
Fit prediction needs body-facing dimensions at key points. These may appear in a spec sheet within construction measurements that need interpretation. For example, "waist at seam" describes construction. To use it for body fit, you must know where that seam sits on the body.
With new pilot partners, we work with brand technical design teams to separate body-relevant specs from construction specs during data setup. The sheet already exists. We add the translation from its measurements to fit-relevant measurements.
Why garment categories need different measurement sets
Fit-predictive measurements depend on garment category. Fitted trousers need waist, hip, inseam, and thigh circumference, since the thigh can be tight even when waist and hip fit. A blazer needs chest, shoulder, back length, and sleeve length. A woven shirt needs chest, shoulder, neck, and sleeve length. A knit dress needs bust, waist, hip, and fabric stretch range.
Using one measurement set across categories improves recommendations where that set happens to fit and weakens them where other inputs matter. For brands with several categories, uneven results can look like a system issue when the underlying problem is incomplete category data.
We define a schema for each garment category and use it when importing a brand catalog. A fitted woven shirt and a stretch knit top follow different data paths because their fit variables differ. Catalog setup is more involved, but mixed assortments receive better predictions than they would from one shared schema.
The limits of measurement-based fit prediction
Even carefully collected garment specs describe the garment at rest. Fit changes with movement. A trouser can sit correctly at the waist and hip while standing yet bind at the seat when sitting if its curve misses the shopper's body geometry. A woven shirt can have enough chest room yet pull when raised shoulders expose inadequate armhole depth.
Our recommendations do not predict dynamic comfort. They estimate static fit at rest and account for stretch in stretch fabrics. Predicting comfort through movement needs more data than two photos and a spec sheet. The result tells shoppers whether a garment falls within the range of their body measurement and fit preference. It does not replace trying it on, but it reduces wrong guesses that lead to returns.
