01 — The brief
An education provider tenders for a vocational training contract every three years, and the tender requires each bidder to submit a physical evidence pack: eleven sections, each with its own printed tab, in a branded ring binder. 2,400 binders.
The date was fixed by the tender authority and could not move. The provider won the contract with four weeks’ notice — which turned a comfortable job into a six-week job compressed into four, with the binders needed before the submission deadline, not after.
02 — The hard part
Eleven tab positions, and they must be indexed correctly on every single binder. This is the failure that looks impossible and happens constantly: a tab set collated with section 7’s divider behind section 6. In a bid submission, a misfiled tab is not a cosmetic defect. The evaluator opens the pack to find the wrong section and the bid looks careless.
At 2,400 binders × 11 tabs, that is 26,400 hand-collation decisions. Human collation at that volume has a real error rate. We did not want to accept a sampling plan for an error that cannot be tolerated at all.
Second problem: the deadline was already gone. Four weeks for 2,400 welded vinyl binders with printed tab dividers normally means straight to express everything, and express on the divider print run was quoted at a premium that ate the margin.
Third: the tab stock tore. Standard tab dividers on 250 gsm stock tear at the tab shoulder when a heavy section is pulled. In a 300-page evidence pack, every section is heavy.
03 — What we built
Kill the collation error entirely. Rather than collating by hand and inspecting, we printed the tabs as a single imposed sheet per binder set, in order, so that the dividers come off the guillotine already sequenced. The tab number is printed on the back of the preceding divider as well, so a misfile becomes visible before the binder is closed. Collation stopped being a manual step and became a cutting constraint.
Absorb the deadline. We split the job: the vinyl binder covers are our own high-frequency welding line, which we ran immediately, in parallel with the divider print. Cover welding and divider printing are independent, so the two ran on two tracks instead of queueing. That recovered roughly eight days.
Tab strength. Dividers moved to 300 gsm with a reinforced tab shoulder and a drilled-and-reinforced ring hole. The shoulder is where the tear starts; reinforcing the fold costs very little per unit and is invisible when done well.
04 — Production
2,400 binders, two parallel tracks, one final assembly and packing stage.
QC: every binder set checked for tab sequence against a printed reference strip at packing (a 15-second check per set, which at this volume is affordable); ring-pull cycle test on a sample per run; 100% visual on the welded cover seam.
Packed 40 sets per carton, palletised and shipped DAP to the submission hub rather than the client’s office — the packs had to be delivered where they would be handed over, and a split shipment across two addresses is how these jobs go wrong.
05 — Outcome
- All 2,400 sets delivered with eight days to spare before the submission deadline.
- Zero tab-sequence defects. Not “within an acceptable AQL” — zero, because the process removed the possibility rather than inspecting for it.
- The provider won the contract.
- They have now used the same sequenced-tab method on two subsequent tenders.
The transferable idea: when a defect is intolerable rather than merely undesirable, do not add inspection — change the process so the defect cannot be produced. Hand-collating 11 tabs and then sampling for errors is a plan for shipping some wrong binders and hoping nobody opens them.