Restarting After Diwali: Getting Lines Back to Full Output in Week One
The first week after Diwali is not lost because people are missing. It is lost because the plant opens every line with whoever turned up. Run fewer lines at full strength, start with the orders that carry a ship date, and have the machines warmed and checked before the operators arrive. In our worked example that choice alone recovers about 4,200 units, or ₹1.3 lakh of contribution, in six working days.
By the Mama Editorial Team. We scope camera and floor-visibility projects for Indian factories. Last verified: 13 October 2026, against the Government of India 2026 holiday list, Haryana's April 2026 minimum-wage notification, the Labour Codes and the cluster reporting cited inline. All ₹ figures are illustrative.
Key takeaways
- Diwali 2026 is on Sunday, 8 November. That is the date in the Department of Personnel and Training's 2026 holiday order, as reproduced by a central government office. Chhath falls in the week after it, and Guru Nanak's Birthday is on 24 November. Workers come back in waves.
- Reported return rates swing widely. In mid-November 2025, Surat processors put their labour shortfall at more than 30%, with only about 60% of dyeing and printing mills back up. In 2024, Ludhiana production dipped to 50–60% over the festival weeks.
- Consolidate, don't spread. A line with holes on its bottleneck stations loses more output than its share of missing people. Four full lines usually beat six thin ones.
- Sequence lines by order and skill, not by which supervisor came back first. Start with shipments that carry a date and lines whose bottleneck skills have returned.
- Machines need a Day 0. Compressors, boilers, hydraulics, dies and first-piece approvals all take time after a week of standing still. Do that work before the full shift walks in.
The 2026 calendar is the plan
Put three dates on the wall before you announce the break:
| Date (2026) | What it means for the floor |
|---|---|
| Sun 8 Nov: Diwali (Deepavali) | Most units close for a few days around it. Bonus has been paid and trains are full. |
| The week after Diwali: Chhath | Workers from Bihar and eastern UP often stay home until it is over. In 2024 a Ludhiana woollen-industry leader said factories return to normal only after Chhath, once the workforce comes back. |
| Tue 24 Nov: Guru Nanak's Birthday | A gazetted holiday in the same 2026 list. It matters in Punjab and wherever your workforce marks it. |
The DoPT order (O.M. No. 12/2/2023-JCA, 3 July 2025) is for central government offices, so it fixes the festival dates rather than your plant's holidays. Your state's festival-holiday rules and your plant's notified holiday list decide which days are paid holidays, and some regions celebrate on a different day, so check yours.
With north-Indian migrant crews, expect a partial first week and a second step up after Chhath. Plan for two restarts, not one.
What the clusters actually report
There is no national return-rate statistic. What exists is cluster reporting, and it varies by year and by order book:
- Surat, November 2025. Mill owners and contractors put the shortfall at nearly 30%; the Southern Gujarat Textile Processors Association said more than 30%. A printing machine that normally needs eight to ten workers ran with four to six, owners paid contractors to bring workers back, and the association hoped for full capacity only after 25 November. That year the Bihar elections were stacked on top of Chhath.
- Ludhiana, October 2025. After Chhath, the World MSME Forum president put the share who had gone home at 35–40%, "slightly higher than usual"; the woollen manufacturers' chairman said 50–60%. The Tribune reported that incentives, bonuses and in-factory celebrations kept many lines largely intact.
- Ludhiana, November 2024. Production dipped to 50–60%, and garment output to 40–50%. "This happens every year so we are used to it," the woollen manufacturers' association president said.
- Tiruppur and Coimbatore, October 2022. One Tiruppur owner told his workers to return after ten days and told clients deadlines would move by ten days.
Retention effort and order visibility change the numbers. The cost of people who never come back is a separate problem, priced in our piece on what non-returning operators cost a sewing line.
Before you close: the last working week
Most of a good restart is done before the break.
Build a return roster by line and by skill. Don't just record a headcount. For each line, write down who holds the bottleneck skills: the welding fixture, the press tool setter, the test station. Ask each person their expected return date and record it. If two of your three setters are from the same district in Bihar, you already know which line opens late.
Pre-stage the first three days. Cut panels, kitted parts, packing material and released drawings for the orders you'll run first should be at the line before you lock up.
Freeze new styles and new jobs. Don't schedule a new style, a new die or a new customer's first article in the week after the break. Our garment style-changeover piece shows how much a first hour costs on a normal day.
Use the shutdown for maintenance, but finish it two days before the floor opens.
Pay any return incentive on top of the statutory bonus, never out of it. The festival-churn piece covers the bonus and journey-allowance rules. Get your labour adviser to look at how a "back by date X" payment is worded.
Day 0: the machines come back first
Bring maintenance and a skeleton crew in the day before the operators:
- Air and power: compressors started and drained, leaks fixed, DG set and UPS batteries checked.
- Heat: boilers, ovens, furnaces and mould heaters brought up to temperature. These take hours, not minutes.
- Hydraulics and lubrication: presses and moulding machines cycled empty, sewing machines oiled and run, CNC axes homed and warmed up.
- Tools and dies: first dies mounted and first-piece samples approved the day before, so inspectors aren't the queue at 9:15 a.m. See waiting for first-piece approval for what that queue costs on a normal day.
- Material: store open, first loads issued, forklift batteries charged.
Days 1 to 6: which lines run first
On Day 1 every supervisor will want his line open. Resist that. Fill lines in this order:
- Orders with a ship date in the next two weeks, especially export orders or ones carrying a penalty clause.
- Lines whose bottleneck skills are back. A line with 36 people but no one for its bottleneck operation is a line making WIP, not product.
- The highest contribution per operator-hour, if you still have people left over.
Fill each chosen line to full strength before you open the next one. People left over become floaters who cover breaks and early absences, or do kitting and cleaning for the lines that open on Day 3 or Day 4. A skill matrix tells you who can move where. See the multi-skilling matrix guide to build one.
Two more rules for the week. Count heads at the line, not at the gate, because contractors sometimes send substitutes under the same headcount (see contract labour headcount). And open a closed line only when it can be staffed to full strength for two days running, not on a good morning.
Worked ₹ example: six thin lines or four full ones
Assumptions (illustrative; swap in your own):
- 6 lines × 40 operators = 240 operators. A full line makes 1,000 units a day in an 8-hour shift, so 6,000 units a day plant-wide.
- Contribution (selling price minus material) of ₹30 a unit.
- Wages from Haryana's notification No. 2/25/26-2 Lab of 9 April 2026 under the Code on Wages: ₹645.41 a day (₹16,780.74 a month) for semi-skilled workers from 1 April 2026. We assume operators are graded semi-skilled.
- Attendance rises from 60% on Day 1 to 88% on Day 6.
- A thinly staffed line loses 1.67 times its missing share (15% missing costs 25% of output), because holes fall on bottlenecks. That is an assumption, not a measurement.
- Restart ramp on both plans: 80% of normal pace on Day 1, 90% on Day 2, full pace after that.
Plan A spreads everyone across all six lines. Plan B fills whole lines and turns the remainder into floaters or a partial line.
| Day | Attendance | People present | Plan A output | Plan B output | B − A |
|---|---|---|---|---|---|
| 1 | 60% | 144 | 1,594 | 2,666 | +1,072 |
| 2 | 68% | 163 | 2,507 | 3,600 | +1,093 |
| 3 | 75% | 180 | 3,495 | 4,165 | +670 |
| 4 | 80% | 192 | 3,996 | 4,666 | +670 |
| 5 | 85% | 204 | 4,497 | 5,000 | +503 |
| 6 | 88% | 211 | 4,789 | 5,000 | +211 |
| Week | 20,878 | 25,097 | +4,219 |
A normal week is 36,000 units. Plan A ends the week about 15,100 units short and Plan B about 10,900 short. The wage bill is identical in both plans because the same people are paid. The difference, about 4,200 units, is roughly ₹1.27 lakh of contribution in one week.
Now suppose you have to make the shortfall up. Under the OSH Code and Code on Wages, overtime is paid at twice the ordinary rate of wages. ₹645.41 for an 8-hour day is about ₹81 an hour, so an overtime hour costs about ₹161. Assuming overtime runs at the normal 3.125 units per operator-hour, that is about ₹52 of overtime wages per unit recovered. Catching up Plan A's shortfall costs about ₹7.8 lakh in overtime wages. Plan B's costs about ₹5.6 lakh. Overtime also needs each worker's consent, and your state sets the quarterly ceiling, so it is not an unlimited tap.
Plan B's edge shrinks as attendance climbs: most of the loss sits in the first three days, where sequencing pays.
What cameras can and cannot tell you here
Cameras on the floor can help with the line-level half of the restart:
- Which machines actually ran on Day 1, and from what time. A press that was "ready" but didn't cycle until 11 a.m. shows up as a late start, not as a vague bad morning.
- Occupied stations per line, hour by hour, as an aggregate count, against the plan.
- Where WIP piled up, which tells you which open line is missing its bottleneck skill.
- Your own recovery curve. How many days each line took to reach its pre-Diwali running pattern, so next year's plan uses your plant's numbers rather than a newspaper's.
Cameras cannot tell you who is coming back, why someone stayed home, what a contractor promised, whether a returning operator is rusty or whether the first pieces are within tolerance. They also can't see a die that is mounted but unapproved. Floor counts are aggregates; they should never be used to identify or rank individual workers. If cameras watch your floor, workers must be told what is recorded and why. Our DPDP worker CCTV notice template covers that.
Where Mama fits
The hard part of a restart week is seeing, next morning, which lines and machines actually ran. Mama reads your existing floor cameras, or cameras added where you have none, and every morning sends the owner a short plain-language note: which lines started late, which machines stood idle, where work piled up, and roughly what that was worth in ₹. It watches machines and flow, and it counts people only in aggregate, never by name. Mama can draw a camera plan for the lines you restart first. Leave an email or phone in the form below.
FAQ
When is Diwali in 2026? Sunday, 8 November 2026, per the Government of India's 2026 holiday order for central offices. Chhath follows the next week. Your state's rules and your plant's notified holiday list decide your paid holidays.
What share of workers come back in the first week after Diwali? There's no national figure. In 2025 Ludhiana bodies put departures at 35–60%, though incentives kept many lines running, and Surat's processing mills reported a shortfall of more than 30% in mid-November. Record your own return roster by line for two seasons and plan on that.
Should I open every line on the first day back? Usually not. A line missing people on its bottleneck stations produces far less than its headcount suggests. Fill the lines carrying urgent orders to full strength first, use the remainder as floaters, and open the rest as attendance allows.
Can I run Sundays to catch up after Diwali? Only within the law. The OSH Code limits a worker to six days a week and requires a compensatory holiday if a weekly holiday is lost. Overtime needs the worker's consent, is paid at twice the ordinary wage rate and is capped by your state's rules. Check the current position with your labour adviser.
How do I know which line recovered slowest? Compare each line's output for the six working days after the break with its average for the six weeks before, from the shift register or, more reliably, from what each machine actually did.
