The line runs seven days. The control plan assumes five.

Here is an uncomfortable observation. The most consequential quality directive of this quarter did not come out of a core-tools manual, a customer-specific requirement or an IATF clause. It came out of a scheduling decision. Maruti Suzuki has instructed its Indian suppliers to cap the working week at six days, and read properly, that instruction is one of the most honest quality documents an OEM has issued in years. This is not a labour story with quality side-effects. It is an admission, from the largest carmaker in the country, that the seventh day is where process discipline goes to die.

Anyone who has run a plant knows the seventh-day shift before the data confirms it. The machines sound the same. The andon works. The cycle times hold. And nothing on the dashboard admits that Monday morning's scrap is already being made.

What the seventh day quietly spends

I built the QA/QC department at SNOP, a 900-plus-employee greenfield plant, and I designed the weekend coverage myself: layered supervision, a rotating quality technician, maintenance on call. The escape data redrew that design within two quarters. The clusters sat on Sunday nights and Monday mornings, on the lines that had run longest. Never Tuesday afternoons. Not once.

A seven-day week does not merely add hours. It cancels, quietly, the hours that keep the other hours honest:

• Maintenance windows. Planned maintenance is written for a five-day week. Run the line Saturday and Sunday too, and the PM task does not move to another slot – it evaporates. Nobody signs the cancellation.
• Layered process audits. The Sunday auditor is the shift leader already covering two lines. An audit by someone with no time to audit is a signature, not evidence.
• PFMEA reviews and QRQC follow-ups. The quality engineer who owns the containment is on the one day off they were promised, so verification slides from Monday to Wednesday and the escape buys itself another fortnight.
• Supervision. The seventh day runs on the second team, which is arithmetic, not insult. The deepest process knowledge is at home resting – as designed – exactly when the process runs longest unsupervised.

Fatigue is a process variable

Treat fatigue the way you treat torque or melt temperature. It drifts. It has a distribution, across the shift and across the week. It interacts with other inputs – changeover pressure, heat, manning gaps – and it appears on no control plan I have ever signed. We write reaction plans for tool wear measured in microns, then ignore the input holding the gauge. The torque wrench gets recalibrated on schedule; the hand that positions it gets a vending-machine coffee and hour eleven of day seven.

Monday-morning QRQC sessions taught me this the expensive way. Same press, same coil of steel, fixture verified against the same master – and a defect signature that tracked, week after week, to parts made on the seventh consecutive working day. Once we started logging consecutive days worked beside the defect data, the correlation was embarrassing in its clarity. Not enough for a courtroom, no. Enough for a plant manager who had just signed a €60,000 sorting invoice.

The OEM owns the calendar it creates

Suppliers do not run seven days by preference. They run seven because OEMs bank weekends, pull forward end dates and schedule volumes as though the supplier's calendar were elastic. Across the years I led quality initiatives over a 2,000-plus workforce on multiple sites, I watched customers treat our weekends as free inventory. The cost always arrived later, wearing a different label – a debit note, an 8D, a sort at the customer's dock – and the name on the cover page was always ours, never the scheduler's.

That is why the better auditors ask for working-time records. In the VDA 6.3 audits I have led, working-time and maintenance logs predicted my findings more reliably than any control-plan binder. A plant running twenty-seven days a month against a maintenance plan written for twenty-two has told you everything you need to know before you cross the gate. IATF 16949 requires controlled production scheduling for a reason: a schedule starved of maintenance and supervision is an uncontrolled variable with a purchase order attached.

The instruction lands in a quarter already crowded with supply-chain quality headlines – an airframer re-inspecting hundreds of undelivered narrow-bodies over production defects, consumer-electronics launches trimmed over yield problems. What distinguishes Maruti's move is the direction of the admission: an OEM pricing its own contribution to escape rates instead of invoicing suppliers for theirs. That is rare enough to deserve its own headline.

The seventh day does not create capacity – it converts maintenance windows into escapes and books the difference as output.

A capacity problem in a scheduling costume

If you genuinely need the seventh day, you do not have a scheduling problem. You have a capacity problem in a scheduling costume, and there are exactly two honest responses. Buy the capacity – a shift, a line, a changeover attacked with the seriousness you give a customer audit. Or price the escapes into the contract and let the customer choose with open eyes. What you cannot do is run the loan off the books and call it output.

Maruti did not publish a labour policy. It published a failure-mode analysis of its own supply base and signed it. The suppliers who read the instruction as a quality document – cap the week, fix the capacity, price whatever remains – are the ones still shipping a decade from now without an apology attached. The rest pay the interest instead: in debits, sorting, quietly transferred volumes. Capacity bought with the calendar is borrowed against escapes. The plants that price that loan honestly keep their customers.