Knitted Bag Sampling Steps: Three Routes Through the 7-Day Window, Compared Stage by Stage

What Counts as a Sampling Step, and What Does Not

A sampling step is any stage that changes the physical piece; a sampling milestone is the point at which a decision is taken — and conflating the two is why buyers and production teams disagree about how long sampling takes.

On a knitted bag there are six milestones and roughly eleven operations behind them. Programming and file preparation is an operation, not a milestone, because nothing physical changes when a file is written. The first knit is the operation that opens the physical sequence. Heat setting follows, because a knitted body has to be stabilised before any fit judgement means anything. The fit assessment is the first true milestone, since it is where the buyer says yes or no. Structural correction is the second. Trim and lining is the third, measurement and packing the last. Buyers who count operations see eleven steps and expect eleven days; buyers who count milestones see six and expect seven.

The distinction matters commercially because only milestones need the buyer in the room. Eleven operations can run inside 7 working days if the buyer answers at three of them. The same eleven operations can take four weeks if the buyer answers at nine, which is the most common reason a sampling window overruns and the least often acknowledged.

Route A: Digital-First, Built From a Tech Pack

Digital-first runs programming before any yarn is committed: the body is written as a knitting file from dimensions, gauge and stitch structure, then knitted, set and fitted — and it is the correct route when the buyer holds a tech pack with a stated GSM and tolerance.

Route A puts the risk at the front. Programming a whole-garment body from a drawing is the stage where a silhouette either works or does not, and it costs little to iterate because no yarn has been bought and no bed has been booked. Once the first knit passes the fit assessment, the remaining milestones are mechanical: correction, trim, measurement. The route assumes the buyer can supply numbers rather than only images, specifically gauge, yarn count, finished GSM, body dimensions and handle drop. A tech pack with those five figures plus the trim list is enough to programme against; a mood board is not, and a programme that starts from a mood board is on Route B whether or not anyone says so.

• Best for: a buyer with a tech pack, a defined GSM and a toleranced dimension set.
• Front-loads: programming, which is the least expensive stage at which to discover that a silhouette does not work.
• Needs from the buyer: gauge, yarn count, target GSM, dimensions, handle drop and trim list.
• Produces: three sample pieces, one of which is kept as the approved reference.
• Runs to: 7 working days, with two correction rounds inside that window.

Route B: Physical-First, Built From a Reference Sample

Physical-first starts from an existing bag: the reference is measured, taken apart where necessary, and digitised into a knitting programme — and it costs one working day more than Route A because digitising a physical object is a separate operation from writing a file.

The extra day is measurement and interpretation. A reference bag has to be measured in the relaxed, conditioned state, which means conditioning it for 24 hours, then recording at least twelve points, then deciding which of those points were intended by the original designer and which are artefacts of how that particular bag was finished. The last judgement is the difficult one. A reference bag that has been washed will have shrunk, and copying its measurements copies the shrinkage. A reference bag that has been stuffed for photography will read wider than it was made. Route B produces four pieces rather than three, because one is knitted to match the reference rather than the brief, so the buyer can compare like for like.

Route B is the correct choice when the brief is genuinely visual — a buyer who has a competitor's bag and wants that silhouette, with no documentation behind it. It is the wrong choice when a tech pack exists but has not been written down, because in that case the same day buys a permanent specification rather than a one-off reading.

Route C: Concurrent, With Yarn and Trim Booked in Parallel

Concurrent sampling takes the six milestones and runs yarn booking and trim sourcing in parallel with programming, which removes roughly two working days from the critical path — but only if the yarn count and shade are fixed before day one, and only if the trim set is drawn from stock.

The saving is real and the condition is absolute. Yarn in a specified count and shade is the longest-lead item in any knit programme, and on a custom or dope-dyed shade it can run to ten days or more before it is on the cone. Booking it during programming rather than after the first fit is the single largest compress available, and it is available to Routes A and C alike. The reason it is not universal is commitment: booking yarn before the first fit means committing to a count and a shade before the body has been seen. A buyer who is confident about the yarn and unsure about the shape should run Route C. A buyer who is unsure about both should run Route A and accept the longer calendar.

The Comparison: Six Milestones Across Three Routes

All three routes run the same six milestones; what differs is the order, the number of pieces produced and whether yarn is committed before the first fit — which is why Route C is fastest on paper and slowest when a shade change is requested at milestone three.

The table understates one thing: Route C's five days assume nothing changes. If the fit assessment at milestone three leads to a different yarn count, the yarn already booked is wrong, and Route C becomes slower than Route A because the booking has to be redone. That is the entire trade. Route C buys two days by giving up the option to change yarn without rebooking, and the option is worth more than two days on any programme where the hand and drape are still being decided.

Where the 7 Working Days Actually Go

Inside the 7-working-day window the time is unevenly distributed: programming and the first knit take a day and a half, heat setting half a day, the fit assessment one day, structural correction one day, trim and lining two days, and measurement, photography and packing one day.

Trim and lining being the largest block surprises buyers who expect knitting to dominate, and it is the block most worth attacking. Two days of a seven-day window are spent placing a zip, a lining, hardware and a label, all of which are discrete operations requiring a jig and a check. A programme that samples with no trim at all can compress to five days, and many buyers do exactly that on the first round: approve the body first, then approve a trimmed sample second. The cost is a second sampling cycle; the benefit is that the structural decision is made on a body rather than on a body carrying hardware that distracts from it.

Programming and file preparation

1.5 days, including the first-piece check

First knit

0.5 day, three pieces at a 9-minute cycle

Heat setting and conditioning

0.5 day, plus the overnight hold on piece one

Fit assessment

1 day, buyer decision point

Structural correction

1 day, round one of two included

Trim and lining

2 days, the largest single block

Measurement, photography and packing

1 day, twelve points plus GSM

Correction Rounds: How Many, and Who Pays for Each

Two correction rounds sit inside the 7-working-day window, and the distinction that matters is not the count but the type: a structural correction changes the knitting programme, while a specification correction changes what the buyer asked for, and only the first is inside the quoted sampling scope.

A structural correction is one where the programme did not deliver the brief — the body came out 6 mm narrow, the handle drop is wrong, the base transition creases, the reinforcement band sits too low. Those are programme errors and they are corrected at no change to the sampling calendar as long as they are identified at milestone three. A specification correction is one where the brief changes after the piece has been seen: a different GSM, a different handle drop, a different yarn. Both are legitimate, but the second consumes a correction round and can consume a yarn booking. Naming which of the two a request is, at the moment it is made, is the single most useful habit a buying team can adopt in sampling.

Sampling Fabric Versus Production Fabric

A sample knitted on the production bed in the production yarn is predictable; a sample knitted on a different gauge, in a substitute yarn or without the production setting cycle is not, and that difference is the source of most disputes at bulk approval.

The trap is a sample that looks right for the wrong reasons. A body knitted at E14 when production will run at E12 will look finer and drape differently; a body knitted in a substitute yarn because the specified lot has not arrived will differ in hand and in GSM; a body that skipped the setting cycle will measure correctly on the day and move overnight. All three produce an approval that does not survive contact with the bulk run. The defence is a sample record that states gauge, yarn lot, GSM, setting parameters and measurement conditions alongside the approval, so that bulk is compared against a documented piece rather than a photograph.

From Approved Sample to Bulk: The Handover Checklist

An approved sample becomes a production specification only when five things are written down — gauge, yarn lot reference, finished GSM, setting parameters and the signed-off trim list — because a bulk run that matches an approved sample visually but not numerically is a bulk run waiting to be rejected.

Bulk then runs to 30 days from that handover: yarn booking across the first 10 days, knitting across days 8 to 20, finishing and trim across days 18 to 26, and inspection, packing and the pre-shipment release across days 24 to 30. MOQ is 100 pcs per style and per colourway, capacity behind it is 7 production lines at about 200,000 pcs per month, and the release is against AQL 2.5 under an ISO 2859-1 sampling plan. FOB Xiamen is the standard term with CIF and DDP available, and the reference unit cost for a standard E12 tote at specification is USD 4.80. For the file preparation that feeds milestone one, see digital knitting file preparation.

Frequently Asked Questions

People Also Ask

Choosing a Route Is a Commercial Decision, Not a Technical One

The three routes differ in what they commit and when, not in the quality of the sample they produce. Digital-first keeps the yarn option open until the shape is approved. Physical-first buys a silhouette from an object. Concurrent buys two days by committing to yarn and trim up front. A buyer should choose on the basis of which uncertainty is larger: the shape, or the material.

For the reinforcement decisions that most often drive a correction round, see reinforcement zones in a seamless knit body. For the programme windows around sampling and bulk, see custom knit bags.