8–12% Daily Absenteeism: Rebalancing a Sewing Line and the ₹ Cost of Floaters

On a typical day, one in ten sewing operators in a large Indian garment factory does not turn up — and the line barely notices until absence crosses about 10%, after which efficiency starts to fall. The real decision is how you replace the missing hands: double-rate overtime, a paid pool of multi-skilled floaters, or both. On the illustrative numbers below, a floater pool of about 10% of the line is cheaper than chasing the shortfall with overtime.

By the Mama Editorial Team — we scope camera and floor-visibility projects for Indian factories. Last verified: 6 October 2026, against the research and government notifications linked inline. All ₹ figures are illustrative worked examples, not benchmarks; replace them with your own wage and output data.

Key takeaways

• 10–11% absent on a typical day is normal, not a crisis. That is what researchers measured across four Karnataka garment factories, with nearly all of it unannounced (Adhvaryu et al., HBS Working Paper 21-109).
• The bad days are what hurt. In the same data, any given line was at least 20% absent one day in ten, and efficiency fell from above 50% to below 45% between 10% and 20% absence.
• Plan capacity on the heads who turn up, not the muster roll. A 50-operator line averaging 11% absence is really a 44–45-operator line.
• Overtime is the most expensive cover. Under the OSH Code, in force since 21 November 2025, overtime is paid at twice the rate of wages.
• Floaters only pay if they can run your bottleneck operations.
• Cameras can count stations manned and show where bundles pile up. They can't tell you why someone is absent, and they shouldn't identify who.

What 8–12% absenteeism actually looks like

The best Indian data on this comes from a study of four ready-made garment factories run by a large exporter in Karnataka. Achyuta Adhvaryu and co-authors tracked six to seven months of daily attendance and line efficiency in factories with roughly 21–22 lines each. Their findings (HBS Working Paper 21-109; summary on Ideas for India) are a reference point for other clusters, not a measured norm for them:

• On a typical day, 10–11% of workers are absent. A typical line had 56 home-line operators, so 5 to 6 were missing every day.
• Nearly all of it is "unauthorised", meaning nobody tells the firm in advance. Common causes: health shocks in the family, festivals that need travel to the native village, and short-term work that pays better than a lost day's wage, such as harvesting coffee or areca nuts.
• The shocks are lumpy. The average is 11%, but any given line was at least 20% absent one day in every ten.
• Lines are seldom hit together. Absence on one line is only weakly correlated with absence on the next, which is why borrowing operators between lines works at all.

The study notes that most workers arrive just before the 9 a.m. start, and in those minutes the line manager must guess whether a missing operator is absent or just late, and whether to borrow someone.

The efficiency penalty kicks in after about 10%

The useful finding for an owner is the shape of the loss. In the Karnataka data, absence barely moved line efficiency up to about 9–10%, because supervisors absorbed it by reshuffling operations within the line. Past that point the line can't absorb it. The researchers report that each extra percentage point of absence above 10% cut efficiency by about 0.25 percentage points (Ideas for India). The paper's binned averages fall faster: from above 50% efficiency at under 10% absence to below 45% at 20% absence (HBS WP 21-109).

The mechanism is line balancing: the slowest operation sets the pace (see UPH, UPPH and line balancing). Lose one of three side-seam operators and the other two speed up. Lose the only person on collar attach and everyone downstream is starved.

Note the baseline too. The average line efficiency in that dataset was about 49%, which sits inside the 45–55% band commonly estimated for Indian lines, a directional industry figure rather than survey data (see what 1% of sewing efficiency is worth).

Worked example: one 50-operator line, one month, three choices

Here is the ₹ comparison for an illustrative line in the NCR (Gurugram–Manesar) belt, priced at Haryana's notified wage. Note the pairing: the absence profile below is borrowed from the Karnataka study, because no comparable dataset exists for NCR. Treat it as a stand-in and replace it with your own attendance register. Every state notifies its own wage, so plug in yours too.

Wage inputs (Haryana, from 1 April 2026). The Haryana Labour Department notification No. 2/25/26-2 Lab of 9 April 2026 (issued under the Code on Wages, 2019) fixes the basic minimum for a semi-skilled worker at ₹16,780.74/month or ₹645.41/day, with the day rate calculated on 26 days. It is CPI-linked, so check for a later revision. Assume your operators are graded semi-skilled; many units grade tailors as skilled, at ₹18,500.81/month (₹711.56/day). Assume an all-in cost of 1.25× the wage for the employer's EPF/ESI share, leave and bonus, giving ≈ ₹20,976 per operator per month. That multiplier is our assumption, not a statutory figure, so use your own payroll number.

Overtime rate. The four labour codes took effect on 21 November 2025. Under Section 27 of the OSH Code, overtime is paid at twice the rate of wages and needs the worker's consent. At ₹645.41 ÷ 8 hours ≈ ₹80.68/hour, overtime costs ≈ ₹161.35/hour. Total overtime hours are capped by the appropriate (usually state) government; option A below needs about 25 overtime hours per operator a month, so check it against your cap.

Line assumptions (all illustrative): 50 home-line operators, an 8-hour shift, a 12-SAM garment, 50% efficiency at full strength, so 1,000 pieces/day and 26,000 pieces/month planned. Over a 26-day month we assume this attendance pattern, which averages about 11% absent and has roughly one day in ten above 20%, close to the study's profile:

We apply the conservative 0.25-point penalty per point of absence above 10%. We also assume absent operators are unpaid, which the study reports as the main consequence of unauthorised leave, and that a floater placed on the line performs at the line's average efficiency. That last one is optimistic.

Three things stand out:

1. Most of the shortfall is missing hands, not lost efficiency. In option A, only about 240 of the 3,201 missing pieces come from the efficiency penalty. The rest is simply 11% fewer operators. If your planning department loads lines on the full muster, every order starts behind.
2. Floaters beat double-rate overtime at these numbers, because a floater costs about 1.25× the normal hourly rate and overtime costs 2×. On this example, the 10% pool saves roughly ₹61,000 a month against option A for one line. Across a 10-line floor that is about ₹6.1 lakh a month, still illustrative.
3. Diminishing returns are real. Moving from 3 to 5 floaters saves only ₹13,000 a month, because on the 6 low-absence days the extra floaters have no gap to fill. A sixth or seventh floater would cost more than it saves.

The result flips if your floaters can't run the bottleneck operations: a slow floater on collar attach still leaves the line starved, and you pay twice. It also flips if your absence is rare and spiky. A line that is fully staffed most weeks is better served by occasional overtime than by a standing pool.

Floaters vs multi-skilling: what the money actually buys

A floater is extra headcount; a multi-skilled operator is flexibility inside the headcount you have. You usually need both.

• Floaters pay where absence is chronic and predictable, such as festival weeks or harvest months, if your own attendance data shows them. Keep the pool factory-level, not per line, because absence is weakly correlated across lines.
• Multi-skilling lets the supervisor rebalance at 9:05 instead of borrowing. Build a skill matrix of operators against operations (graded trainee / can do / can do at target) focused on the bottleneck operations of current styles. A line where three people can run collar attach absorbs a lost collar operator; a line where only one can is one sick day from stopping. (On EMS lines, where some stations need a certificate, see the multi-skilling matrix.)
• Borrowing between lines is the free option most factories already use informally. The study found it narrow: managers kept at most about five trading partners, forgoing 15–17 others on the floor, and 72% of exchanged workers moved between lines no more than 20 feet apart (Ideas for India). A floor-level morning huddle that matches surplus to shortage across all lines, not just neighbours, costs nothing.

The 9 a.m. rebalancing routine

A practical sequence for each line, before the first bundle moves:

1. Count stations manned by section, not heads at the gate. The line runs on who is at a machine.
2. Check the bottleneck operations first. If every critical operation is covered, a small gap elsewhere can be absorbed by neighbours, as the data shows up to about 10%.
3. Fill critical gaps from the floater pool or the skill matrix, not whichever machine happens to be empty.
4. Re-target the line honestly. At 84% strength, the hourly target board should say so.
5. Watch the bundles, not the operators. WIP piling up in front of one operation by 10:30 tells you the rebalance failed there. See where your sewing minutes go for the waiting loss it creates.

What cameras can and cannot tell you here

A camera over the line can show: - how many stations are manned in each section at 9:15, 11:00 and 14:00, as an aggregate count with no names - where bundles pile up, meaning which operation became the bottleneck after the morning rebalance - how long the line ran short-handed, and whether the gap was covered or just absorbed

A camera cannot show: - why someone is absent, or whether they're late or not coming - an operator's skill grade (output data and the supervisor tell you that) - next month's festival or harvest spike (your attendance register is the better forecaster) - anything that justifies tracking or ranking individual workers. Identifying people isn't needed to balance a line, and under India's DPDP Act any CCTV over workers needs a clear notice of purpose. See the worker CCTV notice your factory needs.

Where Mama fits

Mama reads the cameras already on your floor, or ones we place. Every morning it sends the owner a short WhatsApp note on where yesterday's hours leaked. For absence, that means which lines ran short-handed and for how long, and where work piled up behind an uncovered operation. It reports on stations and flow, counting people only in aggregate. Send a short phone video of your floor and we'll come back with where the hours are going and a camera plan to watch those spots.

FAQ

What is a normal absenteeism rate for an Indian sewing line? In a large published dataset from four Indian garment factories in Karnataka, 10–11% of workers were absent on a typical day and any line hit 20% or more about one day in ten. Yours will vary with season, location and wage, so measure it by line and day of week.

How many floaters should a garment line keep? There's no universal ratio. On our illustrative 50-operator line, a pool of around 10% beat overtime; more added cost because extra floaters sat idle on low-absence days. Size the pool to your own absence pattern, share it across lines, and only count floaters who can run your bottleneck operations.

Is overtime cheaper than hiring floaters? Rarely. Overtime is paid at twice the rate of wages under the OSH Code; a floater costs roughly your normal all-in rate. Overtime suits a rare spike, within the legal cap.

Does absenteeism always reduce line efficiency? Not at first. In the Karnataka data, supervisors absorbed absence up to about 9–10% by reshuffling operations. Beyond that, efficiency fell by roughly a quarter of a point for every extra point of absence, and faster on the raw averages.

Can a camera system track which operators are absent? It shouldn't, and it doesn't need to. Balancing a line needs only how many stations are manned and where work is piling up. Who is absent is a matter for your attendance register and HR, handled under a proper DPDP worker notice.