Attendance Is Full but Output Is 15% Short: Where Are Your Sewing Operators?

Your attendance register says everyone came, yet the line made 15% fewer pieces than planned. As of 17 Oct 2026, the usual answer is that the operators are present but not sewing: moved off their machines, waiting for bundles, or pulled away to fix alterations. On an assumed 500-operator Noida floor, at an assumed ₹580 a day, that gap costs about ₹14 lakh a month in wages alone.

By the Mama Editorial Team. Last checked: 17 Oct 2026. Every ₹ figure is a labelled assumption for you to replace with your own payroll and production data.

Key takeaways

• Count three numbers, not one. On the roll, at a machine, and actually sewing. Output follows the third number; the attendance register only tells you the first.
• The gap hides in three places. Operators and helpers moved to other work, operators seated but waiting for bundles, and operators pulled off-line to re-sew alterations.
• A full muster does not mean a full line. Borrowed operators still show present on their home line's register.
• Price the gap with one figure. Labour cost per attended minute times the minutes that were present but not sewing.
• Count stations and flow, never individuals. Aggregate counts by section answer the question; naming people does not.

Why a full register can sit next to a short line

Most production plans are loaded on the muster: so many operators on the roll, so many minutes, a target at the planned efficiency. If the register shows full attendance and output still falls 15% short, the plan's minutes went somewhere other than sewing.

Full attendance is already better than normal. In a published study of four Shahi Exports garment factories in Karnataka, 10–11% of workers were absent on a typical day (Adhvaryu et al., summarised on Ideas for India, 8 Sep 2021). If your register beats that, absence is not your problem this week, and the absenteeism and floater maths will not explain the shortfall.

The shortfall comes from the space between "present" and "producing". That space has three layers.

Three counts: on the roll, at the machine, actually sewing

The gap between count 1 and count 2 is a deployment problem. The gap between count 2 and count 3 is a flow problem. Owners usually see only count 1 and the final output, so the whole difference gets blamed on "efficiency".

For the full list of what happens to minutes at the machine, see where your sewing minutes go. This page is about the reconciliation: how many of the people you paid for were sewing, and at what times.

Gap 1: present, but not at their machine

Operators moved to other work. When a shipment is due, supervisors pull operators to finishing, thread trimming, checking or packing. They stay present on the register and stay on the sewing section's cost, but their machines stand empty.

Helpers moved, so operators do the helpers' work. A sewing line runs on helpers who feed bundles, cut threads and move work along. If helpers go to finishing, the operators fetch their own bundles and trim their own threads. Every operator is still seated, but each one sews fewer minutes per hour.

Operators borrowed between lines. Borrowing is normal and often useful; the same Karnataka study found line managers trading workers informally, mostly with neighbouring lines (Ideas for India, 8 Sep 2021). The problem here is the bookkeeping, not the move. A borrowed operator is counted on the home line's register and sews on another line's target. Unless someone records the move, one line looks over-staffed on paper and short on the floor.

Other pulls. Sampling, training, a medical visit, a gate pass. Each is small and legitimate; together they add up over a shift.

Gap 2: at the machine, waiting for work

The operator is seated, the machine works, and there is nothing to sew. Cutting is late, a trim has not arrived, or the operation upstream is the bottleneck and bundles are piling there instead of here.

This gap is the hardest to see from the office because the floor looks full. Every seat is taken. The tell is uneven piles: one station buried in bundles, the next with an empty bin.

Gap 3: off-line for alterations

When the end-line check rejects pieces, someone re-sews them. Often that is an operator taken off the line to a repair table, or an operator re-opening seams at their own machine. Either way, those minutes produce no new pieces against today's target.

Alteration minutes hurt twice: the line loses output, and the re-sewn pieces never earned standard minutes in the first place. The ₹ side of the alteration loop is worked through in end-line DHU and the alteration loop.

Worked example: 500 operators on the roll, all present

Wage input. Assume a Noida semi-skilled operator earns ₹580 a day in basic plus DA (₹15,080 a month over 26 days; illustrative), and a skilled tailor ₹650 a day. Take the current notification for your trade and zone from your state labour department (Uttar Pradesh Labour Department); each state notifies its own rates and most revise the DA part twice a year.

If you grade your tailors as skilled, the ₹650 rate raises every ₹ figure below by about 12%. If you run lines in Tirupur or Bengaluru, swap in your own state's current rate for your employment.

Cost per minute. ₹580 a day ÷ 480 minutes = ₹1.21 per minute in wages. Assume an all-in cost of 1.25× the wage for PF, ESI, bonus and leave: ₹1.510 per attended minute.

Floor assumptions (all illustrative): 500 operators on the roll, all present, one 8-hour shift, so 2,40,000 attended minutes a day. The plan was loaded on all 500 at the line's normal efficiency, so every minute lost from sewing comes straight off output.

Three things follow from these numbers.

1. The 15% shortfall needs nobody to be absent. Thirty people moved and forty minutes of waiting each explain most of it.
2. The biggest bucket is the one the floor hides best. On these assumptions, waiting at a seated machine costs more than moving people off the line, and nobody walking past would call those operators idle.
3. Wages are the floor of the loss, not the whole loss. The ₹14.1 lakh is pay for minutes that made nothing. A missed ship date or a lost reorder costs more, and that part depends on your orders.

The overtime trap. Owners often close a 15% gap with overtime. Overtime needs the worker's consent and is paid at twice the ordinary rate of wages (basic plus DA) under OSH Code, 2020, s.27; your state's rules set the overtime ceiling. At the assumed ₹580 a day, ₹580 ÷ 8 hours = ₹72.50 an hour normal, so ₹145 an hour on overtime. Paying double to recover minutes you already paid for once is the most expensive fix available. See also whether rising overtime is work or habit.

How to run the three counts for a week

You can do this without software: one supervisor or IE with a clipboard is enough. Allow roughly 15 minutes per count on a 500-operator floor (an estimate; it depends on your layout).

1. Fix three times. For example 10:00, 12:00 and 15:30. Avoid the first and last half hour of the shift, which have their own losses.
2. Count by section, not by person. For each section, write three numbers: on the roll, at a machine, sewing. No names.
3. Write the reason beside every gap. "6 moved to finishing", "cutting late, 14 waiting", "4 at repair table". The reason is what you act on.
4. Do it for five working days. One day is noise. A week shows whether the gap is a habit or a one-off.
5. Add up the minutes and price them. Use the cost per minute above, or your own from payroll. The absenteeism cost calculator handles the count-1 side if you also have absence.

What you usually learn: gap 1 is a supervisor habit (pulling operators for finishing every afternoon), gap 2 is a cutting or line-balancing problem, and gap 3 is a quality problem upstream. Each has a different owner in your management team.

What cameras can and cannot tell you

A camera over the sewing floor can show:

• how many machines in each section have an operator seated at set times, as an aggregate count
• how many seated stations show sewing activity, and for how long stations sat waiting
• where bundles pile up, which tells you where the line is starved or blocked
• how the three counts moved hour by hour across a week

A camera cannot show:

• why an operator was moved, or who authorised it; your supervisor's note does that
• whether a borrowed operator was booked to the right line in the register
• the skill grade of the person at a machine, or whether the right person is on the bottleneck operation
• anything that justifies tracking or ranking named workers. Balancing a line needs counts, not identities. Footage of recognisable workers is personal data under s.2(t) of the DPDP Act, 2023; tell workers in writing what is recorded and why (see the worker CCTV notice under DPDP).

A camera also cannot fix any of the three gaps. It shortens the time between "output was short" and "here is where the minutes went", which is usually days if you rely on a weekly IE study.

Where Mama fits

Mama reads the cameras already on your floor, or cameras placed for the purpose. Each morning it sends the owner a short plain-language note on where yesterday's hours leaked: sections that ran with empty machines, stations that sat waiting for work, and roughly what that was worth in ₹ at your own cost per minute. It counts stations and flow, never named individuals. To see where the minutes go on your own floor, leave an email or phone in the form below and Mama will get back to you with a camera plan for those spots.

FAQ

Why is output low when attendance is full? Because the register counts people who punched in, not people sewing. Operators moved to other jobs, operators waiting for bundles and operators re-sewing alterations are all present and all paid. On the assumed 500-operator example above, those three gaps add up to 15% of attended minutes.

How do I find out where my present operators are? Count three numbers by section at three fixed times a day for a week: on the roll, seated at a machine, and actually sewing. Write the reason beside every gap. The pattern across five days tells you whether it is deployment, flow or quality.

Is 15% present-but-not-sewing time normal? The 15% on this page is an assumption chosen to match the shortfall in the question, not a benchmark. Measure your own baseline first; the useful number is the trend and which of the three gaps dominates.

Should I recover the shortfall with overtime? Overtime needs the worker's consent and is paid at twice the ordinary rate of wages under OSH Code s.27, so it is the most expensive way to recover minutes you already paid for. Fix the biggest gap first and keep overtime for genuine peaks.

Does counting operators mean monitoring individual workers? No. The three counts are totals by section. They show where the line stalled, which is a management problem, and they need no names. If cameras are involved, tell workers in writing what is recorded and why.