Knitted Bag Risk Register: Eighteen Risks in Five Families, Built on the First Day

The Programme That Built a Register on Day One

The programme was a spring launch: 12,000 pcs across four colourways, one body shape, two handle lengths, a lead time that put despatch in the second week of March, and a buyer who had run the same bag in a woven version the year before.

The woven history was the problem. Everyone involved assumed the knit version would behave like the woven one, and almost every assumption was wrong: the shade would not match across four lots, the handle length could not be graded the way a strap is graded, the packing count would not be the same, and the compliance file for a knitted body is not the file for a woven one. A register built on day one does not prevent those discoveries; it assigns them. By the end of week one, each of them had a name, a score and an owner, which is what made the difference later.

Six risks were caught early and handled inside the normal calendar. Four were accepted knowingly, which is a legitimate outcome. One was caught late and cost eleven days: a lining material that had been carried over from the woven version and had never been tested against the compliance target for the destination market. It was on the register; it had been scored as low likelihood and given to an owner who did not act on it. That is the failure mode this article is most interested in, because a register that lists everything and changes nothing is worse than no register at all.

What a Risk Register Is Actually For

A risk register on a knit programme exists to turn a list of worries into a list of owned decisions, and its output is not the document but the four actions it produces: mitigate, monitor, accept or escalate.

Most risk documents fail because they stop at identification. A list of eighteen things that might go wrong is not useful in itself; every experienced buyer could write it. What makes it useful is that each line ends in a decision. A risk with a mitigation plan, a trigger date and a named owner is managed. A risk without those three is a worry that has been written down, which is marginally better than a worry that has not. The register should therefore be short on description and strict on the last three columns, and a line without an owner should be treated as unfinished.

There is also a commercial reason to keep one. A programme that can show which risks were known, when they were identified and what was done about them resolves disputes quickly. When a shade problem appears at inspection, the difference between a programme that booked a lot reservation and one that did not is the difference between a shared problem and a claim.

The Eighteen Risks, in Five Families

Eighteen risks cover a knitted bag programme in five families — specification, material, capacity, compliance and release — and the distribution is uneven: material and compliance together account for nine of the eighteen, which is where knit programmes differ most from cut-and-sewn ones.

Family One: Specification Risk, Risks 1 to 4

The four specification risks are a GSM stated as a description, a handle drop that was never dimensioned, an artwork that cannot survive the knitted size, and a tolerance set that was never agreed — and all four are eliminable at kickoff for the price of an hour.

They sit at the top of the register because they are inexpensive to fix and expensive to discover late. A GSM written as 'medium weight' produces different bodies from different programmers and consumes a correction round. A handle drop described rather than measured produces a bag that fits the drawing and not the shoulder, and on a seamless knit body the handle is part of the loop structure, so correcting it re-opens programming rather than a strap pattern. An artwork drawn for print and knitted at 20 mm has about nine needle columns at E12 to exist in. A tolerance set left unagreed means every subsequent measurement is arguable: plus or minus 3 mm on dimensions, plus or minus 2 mm on handle drop and plus or minus 5% on GSM are the values worth writing down.

Family Two: Material and Shade Risk, Risks 5 to 9

Five material risks dominate a knit programme: lot-to-lot shade drift, a custom shade booked late, a substitute yarn in sampling, setting parameters derived from the wrong yarn, and recycled-content documentation that does not travel with the lot.

Shade drift is the highest-scored of the eighteen, and it is structural rather than careless. Yarn is a dyed natural-and-synthetic product with lot-to-lot variation, and the variation is largest on dope-dyed and yarn-dyed shades where colour is fixed before knitting. The mitigation is a triangle: an archived reference swatch from the approved lot, an agreed Delta E tolerance of 1.0 measured under D65, and a pre-production shade check on each new lot before the bed is threaded. The second half of the mitigation is a yarn reservation, commonly 15% over-booking, so that a repeat can draw from the same lot family rather than starting a fresh dye lot.

• Risk 5: lot-to-lot shade drift beyond Delta E 1.0 — mitigate with an archived swatch and a pre-production check.
• Risk 6: custom or dope-dyed shade booked late — mitigate by sampling in the nearest stock shade first.
• Risk 7: substitute yarn used in sampling — mitigate by sampling in the production yarn wherever the lot allows.
• Risk 8: setting parameters derived from the wrong yarn — mitigate by re-running trials when the yarn changes.
• Risk 9: recycled-content documentation not travelling with the lot — mitigate by requesting it at booking.
• Shared mitigation across all five: reserve yarn at 15% over-booking so repeats stay inside one lot family.

Family Three: Capacity and Calendar Risk, Risks 10 to 12

Three calendar risks matter: a peak-season slot not reserved 60 days ahead, a missed vessel turning a three-day slip into a full shipping cycle, and a fragmented run split across two slots to fit it in.

Capacity behind a programme is 7 production lines at about 200,000 pcs per month, and it is booked rather than held open. A standard 30-day bulk run wants its bed slot reserved about 45 days ahead and a Q4 despatch about 60. The vessel risk is the one that surprises buyers: a run finishing three days after the intended sailing does not arrive three days late, it catches the next sailing, and the retail date moves by the cycle rather than by the delay. The fragmentation risk is self-inflicted and common — splitting a run across two slots to fit it in adds a second set-up and a second inspection, and both are avoidable with a flexibility band offered at booking.

Family Four: Compliance Risk, Risks 13 to 15

Three compliance risks carry the longest tails: a target market named after yarn is booked at day 10, a carried-over material never tested against the destination standard, and documentation requested after the run ships rather than at booking.

The second is the one that cost the programme in the opening case eleven days. A lining carried over from a woven version of the same bag had been tested against a previous market and not against the current one; it was on the register, scored as low likelihood, and the owner did not act. The mitigation that would have worked is unglamorous: list every material in the bag — body, lining, trim, hardware, print — and put a test line beside each against the named standard. That is a table with five rows, and it takes twenty minutes to build at kickoff. Testing costs are charged per standard per material, so the same table is also the one that keeps the development invoice predictable.

Family Five: Quality and Release Risk, Risks 16 to 18

The three release risks are inspection against a superseded revision, measurement taken before the body has conditioned, and a packing count set by what fits rather than by weight and recovery.

Inspection against the wrong reference is the quietest of the three and the most damaging. A programme revision changes, the physical reference on file does not, and bulk is then measured against a piece that no longer describes the bag: everything fails, or everything passes for the wrong reason. The second is conditioning: a knitted body keeps relaxing after it leaves the bed, so a piece measured on the knitting day is measured before it has settled. Conditioning for 24 hours at 20 °C and 65% relative humidity is what makes a measurement mean anything. The third is packing: a carton filled to what physically fits compresses the pieces harder than one filled to the 12 kg limit, and compression held across a long voyage sets a crease.

What a Register Does to a Commercial Conversation

A register changes disputes into administration: when something goes wrong, the question becomes what the register said and whether the mitigation was done, rather than who is to blame.

That is worth more than it sounds. A shade claim at inspection is either a shared problem with a documented reservation and tolerance, or an argument about whose expectation was reasonable. The first resolves in a day and usually ends with a partial re-run at an agreed cost split; the second can consume a relationship. The same is true of a missed date: a programme that reserved its slot 45 days ahead and offered a flexibility band has a very different conversation from one that placed an order and hoped. None of this requires a register to be adversarial; it requires it to be written down before anyone needs it.

Scoring: Likelihood, Impact and the Owner Column

Score each risk from 1 to 5 on likelihood and 1 to 5 on impact, multiply for a priority between 1 and 25, and treat anything at 12 or above as requiring a mitigation plan rather than a note.

Likelihood scale

1 rare to 5 near-certain, judged against this programme rather than in general

Impact scale

1 negligible to 5 launch-threatening, judged in days and money

Priority

Likelihood multiplied by impact, 1 to 25

Mitigation threshold

12 or above requires a plan with a trigger date

Review rhythm

Fortnightly while sampling, weekly during bulk

Owner column

One named person per risk; an unnamed owner means an unfinished line

Closure

Confirmed at the pre-shipment inspection under AQL 2.5, ISO 2859-1

The scoring is less important than the discipline around it, and one rule does most of the work: no risk may be scored without an owner. The lining risk on the opening programme was scored at 4, which was defensible on the day, and it had an owner who did not act on it. Had the owner column been enforced with a review date, the score would not have mattered. In practice, a fortnightly review during sampling and a weekly review during bulk is enough to keep a register alive, and each review should ask only three questions: has anything changed, has any mitigation been done, and does any score need to move.

A register with those four columns survives contact with a real programme. Anything less is a list. For the handover that carries the register from development into bulk, see knitted bag programme handover.

Frequently Asked Questions

People Also Ask

The Owner Column Is the Whole Document

Identification is easy and is not where registers fail. Anyone with experience can list eighteen risks in an afternoon. What fails is the last column: a named person, a mitigation and a review date. The lining risk that cost eleven days was correctly listed and correctly scored, and it still happened.

Write the four columns, review fortnightly, and the register does what it is supposed to do. For the design-ownership questions a register often surfaces late, see knitted bag design rights and ownership. For the programme framework, see custom knit bags.