Knit tops are the category in casual womenswear most likely to look simple on paper and fail in the sample room. The same sketch, with the wrong stitch or a weight pushed too light, comes back thin, limp and short. Woven garments are shaped by cutting; knits are shaped by structure — loops stretch, shrink and change form in the wash. So the development order cannot be shuffled: the stitch sets the hand and where stretch comes from, the weight sets the season and the silhouette, the shrinkage decides whether the garment still fits after washing. This piece follows the order a factory actually works in. The broader question of knit versus woven sits in six ways to decide between knit and woven; here we stay inside knit tops.
1. Fix the stitch first: what actually separates jersey, rib and waffle
These are not the same cloth at three thicknesses; they are three different mechanical structures. The point buyers most often blur is where the stretch comes from — and that determines both how you write the spec and how you inspect it:
| Structure | Construction | Source of stretch | Typical yarn | Weight range | Typical styles | Main risk |
|---|---|---|---|---|---|---|
| Single jersey | Single-faced loops, smooth face, float arcs on the back | The loop structure itself gives widthways give; recovery needs elastane | Combed cotton, cotton/modal, cotton-poly | About 120–180 g/m² | Tees, light long sleeves, layering tops | Sheer when light, neckline distortion, lengthwise shrink |
| Rib, 1×1 / 2×2 | Alternating face and back wales form vertical columns | Structural elasticity plus elastane content, both matter | Cotton/elastane, viscose/elastane, poly/elastane | About 150–260 g/m² | Body-hugging layering tops, cropped tops, neckbands and cuffs | Over-elastane loses recovery; narrow width, high consumption |
| Waffle | Raised cellular grid, double-faced interlace with air pockets | Essentially none; shape comes from thickness and structure | Cotton, cotton-poly, low-elasticity polyester | About 200–300 g/m² | Long-sleeve tops, cardigan-style layers, sweatshirt alternatives | Collapses below weight; snagging; skipped stitches on thick seams |
The decision rule is blunt: close to the body and needs recovery — rib; flat, printable, economical — jersey; needs thickness and surface texture — waffle. The three can also be combined: rib for the neckband and cuffs, jersey for the body, waffle for a pocket or a panel. That "one garment, several structures" build is common on knit tops, but every additional structure adds a matching step and a seam, so the quote has to be built on panel count. For rib as a garment in its own right — weight, recovery and sagging — see rib knit tops: weight and elastic recovery.
2. Weight is not "heavier is better": a season-by-category map
Grams per square metre is the metric most easily abused in knit sourcing. The same 180 g/m² behaves completely differently as combed cotton jersey and as polyester waffle — hand, drape and opacity all diverge — which is why weight must always be written together with the structure. A common trade mapping (indicative ranges; confirm against structure and yarn count):
| Season band | Single jersey | Rib | Waffle | Corresponding styles |
|---|---|---|---|---|
| Spring–summer (above 20°C) | 120–150 g/m² | 150–180 g/m² (light stretch layering) | Rarely used | Short-sleeve tees, light long sleeves, camisole overlays |
| Spring–autumn (12–22°C) | 150–180 g/m² | 180–220 g/m² | 200–240 g/m² | Long-sleeve tees, layering tops, light waffle tops |
| Autumn–winter (below 12°C) | 180 g/m² and up (mostly as a layer) | 220–260 g/m² | 240–300 g/m² | Heavy waffle long sleeves, rib turtlenecks, layering pieces |
Two practical corrections. First, light colours go one band up, dark colours can go one band down: white jersey below about 140 g/m² is reliably see-through, while black can usually drop a step. Second, climate differences between Western and Gulf customers outweigh the season label: when a Gulf buyer asks for "heavier", it usually means opaque and fluid rather than warm, and a denser structure at the same weight achieves that better than adding grams. Longline modest tops carry an extra constraint — the longer the garment, the more its own weight pulls the shoulder line and hem, so weight has to be compensated by length. For how opacity is actually inspected, see opacity acceptance for modest fabrics. The wider weight-and-fit framework for the Western casual line is in casual tops: weights and silhouettes.
3. Shrinkage: the two numbers to measure during sampling
Among knit complaints, "it shrank a whole size in the wash" outranks "the colour is off" — and almost every case is measurable at sampling stage. Two numbers matter, not one:
- Lengthwise shrinkage: loops are stretched during knitting and contract in water, hitting body length and sleeve length first. Jersey usually shrinks more lengthwise than widthwise; waffle, being lofty, contracts even more visibly;
- Widthwise shrinkage: rib contracts the most across the width, especially high-elastane stretch rib, where chest and hem pull in by a full size.
The method need not be complicated: mark a 50 cm × 50 cm square on the sample fabric (or fixed points on a garment), wash once on the home-laundry programme of the target market, dry as labelled, then re-measure the same marks; the difference over the original length is the shrinkage. A commonly accepted range is within roughly 3–5% (confirm by category, structure and test standard); beyond that you either build compensation into the pattern or change the process. Three disciplines:
- Relax before cutting: knit rolls should be laid flat to relax, commonly 12–24 hours, before the cutting table — cutting straight off the roll bakes the contraction into the garment;
- Pre-setting beats correction: for structures prone to shrinking, pre-setting during dyeing and finishing is far more reliable than pulling a garment back to size after washing;
- Write shrinkage into the spec clause: state the wash programme, the drying method and the tolerance, or "it shrank" versus "it did not" becomes an argument. For clause wording see fabric spec clauses.
4. Necklines: three finishing methods and how well they hold shape
The neckline is almost always the first thing to fail on a knit top, and whether it stays put is half method, half trim — the match between trim and body structure matters more than the method itself:
| Method | How it is made | Best for | Advantage | Risk |
|---|---|---|---|---|
| Self-fabric band | Strip cut from the body cloth, on the bias or straight, overlocked and twin-needle stitched | Jersey tees, light long sleeves | Zero colour mismatch, low cost | If the cloth lacks stretch it will not return; wavy edge after washing |
| Rib neckband | Separate rib strip, usually with elastane, attached at a defined width and stretch ratio | Crew-neck tees, sweatshirt-style tops, heavier weights | Good recovery, crisp shape | Colour mismatch against the body; excessive stretch ratio causes puckering |
| Bound / taped | Self or contrast binding wraps the edge, secured with topstitching | Waffle, relaxed silhouettes, V-necks | Structurally stable, allows contrast design | Skipped stitches and bulk ridging where thick structures are bound |
There is one hard acceptance test: stretch the neckline to its stated ratio and see whether it returns to the original measurement. In practice, fit the neckline over a form in the correct size, leave it, measure recovery, then wash once and repeat. Elastane content is the decisive variable — below about 5% a rib will not return after repeated stretching, above about 8% it tends to pucker and the seam ripples after washing, so 5–8% adjusted per style is the usual working range (confirm by structure and testing).
5. Machines and seams: three stitch types, three jobs
Eighty percent of seam quality on knit tops comes down to whether the factory owns the right machines, not how fast the operators are. Three questions worth asking on a factory visit:
- Overlock: for side seams and armholes under load; four-thread versus five-thread differs by whether a safety stitch is included. Put knits on a plain lockstitch and the seam simply breaks when stretched;
- Coverstitch: for hems and cuffs; twin- or triple-needle coverstitch produces the elastic stitch that knit hems require, while a flatbed hem will crack;
- Four-needle six-thread: joins and decorative stitching in one pass, common on sporty fitted styles and seam-free panel work; the equipment is expensive, and it shows up in the quote.
For waffle, ask one more question: are there thick-material needle plates and a walking foot? Sew thick cellular structures on an ordinary flatbed and the layers feed at different rates, producing mismatched lengths — and unpicking costs far more than the equipment saved. Fleece hoodies are a different seam and ribbing problem; see hoodie manufacturing: fleece and ribbing.
6. Pre-order checklist
- Write the structure, then the weight, then the neckline method — fill the spec sheet in that order, and never send weight alone;
- If a style uses more than one structure, confirm the quote and the matching requirements by panel count;
- Light colours one band heavier, dark colours one band lighter; compensate modest longline styles for length;
- Request pre-wash and post-wash measurement charts plus a 50 cm square shrinkage test, and put the tolerance in the spec;
- Confirm relax time before cutting, and whether pre-setting is done;
- Test the neckline on a form by size and re-test recovery after one wash;
- Confirm overlock and coverstitch are in-house; for waffle, ask about thick-material plates and a walking foot.
Knit top development comes down to translating "hand feel" into three inspectable parameters: the structure fixes where stretch comes from, the weight fixes the season and the silhouette, the shrinkage decides whether the garment survives its first wash. Get those three right and the sample passes first time; leave them vague and every later round of re-sampling is tuition paid on that first fuzzy spec sheet.
Frequently asked
- What should I check when sourcing knit tops — can you give me a priority order?
- Structure, then weight, then shrinkage, then neckline. The structure sets where stretch comes from and what the hand feels like; get it wrong and everything downstream is wrong. Weight must be written together with the structure — weight alone means nothing. Shrinkage decides whether the garment survives washing, with roughly 3–5% as the commonly accepted range. The neckline is the first part to fail, so test whether it returns after stretching. The first three are all measurable at sampling stage, before bulk.
- Same sketch, same factory brief — why did my knit top come back a size shorter after one wash?
- Almost always unmeasured lengthwise shrinkage with no pattern compensation. Knit loops are stretched during knitting and contract in water, hitting body and sleeve length first; on top of that, rolls cut without relaxing (commonly 12–24 hours flat) bake the contraction into the garment. The fix: a 50 cm square wash test at sampling, shrinkage written into the spec tolerance, relax before cutting, and pre-setting during finishing for structures prone to shrinking.
- Which season is a waffle top for, and what weight should I specify?
- Waffle holds shape through thickness, commonly about 200–300 g/m²: 200–240 g/m² for lighter spring-autumn long sleeves, 240–300 g/m² for autumn-winter layering. It is rarely a spring-summer cloth; if you need one, take the tightest structure at around 200 g/m² and stabilise the neckline with contrast binding. Below the range the grid collapses after two washes, which is the most common reason waffle tops come back.
- My necklines go slack after two wears — can a factory actually guarantee they will not?
- To this extent: stretched to the stated ratio it returns, and after one wash it is still within tolerance. The decisive variables are elastane content and how the trim matches the body — rib neckbands commonly run 5–8% elastane adjusted per style, and below 5% they stop returning under repeated stretch; self-fabric bands need cloth with real stretch or they wave after washing. Acceptance: fit the neckline over a form in the correct size, measure recovery, wash, measure again.
- For small knit orders, is the minimum per style or per colour?
- Per style — about 50 pieces per style — with colour as an allocation inside the style. The real constraint is knitting and dyeing minimums: one colour means one dye lot, so more colours means thinner volume per colour and small-lot dyeing cost spread across the unit price. Keep the first order to two colours per style and hold the third and fourth for the re-order, where whichever colour drew the most questions goes up first.