Where Your Sewing Minutes Go: The Non-Productive-Time Map
You pay for every minute an operator sits at a machine, but on a typical Indian sewing line only about half of those minutes turn into garments. The rest — commonly estimated at 25–45% of attended time — leak into four big buckets: no feeding (the line starved of work), machine breakdown, style changeover, and waiting for a bundle. Each one is visible on the floor if you know what to look for, and each can be priced in ₹ once you know your labour cost per minute.
A low line-efficiency figure tells you half your minutes are gone; it doesn't tell you where. This is the map — the non-productive-time (NPT) categories, what each looks like when you walk the floor, how to measure it, and what it costs. It is the companion to what 1% of sewing efficiency is worth in ₹, which prices the recovery; here we find the leak.
By the Mama Editorial Team — we scope camera and floor-visibility projects for Indian factories. Last verified: 2 October 2026, against standard apparel industrial-engineering practice (lost-time / capture studies, hourly production boards, on- vs off-standard time). All ₹ figures are illustrative worked examples, not benchmarks.
Key takeaways
- Attended minutes split into on-standard (producing) and off-standard (everything else). The off-standard share is your NPT.
- Four buckets carry most of it: no feeding, machine breakdown, style changeover (plus the learning-curve ramp after it), and bundle-wait.
- Each bucket has a floor signature — you can name it from across the hall once you know the posture of each loss.
- Price it with one number: labour cost per attended minute. Multiply by the minutes lost in each bucket.
- This is about flow, not people. You are timing when the line stops, not judging who is slow.
On-standard vs off-standard: the only split that matters
Every paid minute on your floor is one of two things:
- On-standard — the operator is doing the sewing operation the garment's SAM was set for. These minutes earn standard time.
- Off-standard — the operator is present and paid, but not producing standard output: waiting, re-threading, in a changeover, reworking, idle because the line is starved.
Line efficiency is just on-standard ÷ attended. So "improving efficiency" is, concretely, shrinking the off-standard buckets. You can't shrink what you haven't named, so here is the taxonomy.
The non-productive-time map
| Bucket | What it is | How it looks on the floor | How to measure it |
|---|---|---|---|
| No feeding (starvation) | The line has no work to sew — cutting is behind, trims missing, or the previous process didn't hand off | Operators idle with hands still, looking up, tidying; whole sections quiet at once | Hourly board shows a dip across many stations together; lost-time log coded "no feed" |
| Machine breakdown | A sewing machine or special machine is down, waiting for the mechanic | One station dead while neighbours run; a mechanic crouched at a machine; WIP piling just upstream | Maintenance log (call time → fixed time); see why breakdowns take so long |
| Style / size changeover | Line is being re-laid for a new style — machines moved, folders/guides changed, first pieces slow | Line largely stopped; supervisor and mechanic walking the line; bundles of the old style clearing | Clock from last good piece of old style to first good piece of new; add the ramp (see below) |
| Bundle-wait (internal WIP) | Work is in the line but stuck between operations — unbalanced stations, big bundle sizes | One station swamped with bundles while the next sits empty; uneven piles along the line | Count WIP between stations; measure bundle dwell time; bundles = frozen cash |
| Rework / off-standard quality | Pieces come back to be re-sewn; operators doing work the SAM never counted | A repair table busy; operators re-opening seams; a check-point stacking rejects | Reject/alter log; rework minutes don't earn standard time, so they hit efficiency twice |
| Personal / allowed | Breaks, fatigue, bundle ticket handling, line-balancing slack | Normal, expected gaps | Allowance factor in the SAM; keep separate from the avoidable buckets above |
The first four rows are where the recoverable money sits. The last two matter but are partly designed-in (allowances) or quality-driven (rework, which you fix at the source, not by sewing faster).
The changeover trap: the ramp is part of the loss
Owners time a changeover as "line down for 40 minutes." The real loss is longer: after the line restarts, it runs below standard for a learning-curve ramp — operators re-finding rhythm on a new operation — before it reaches steady efficiency. A plant running many small export orders can lose more total minutes to frequent changeovers and their ramps than to any single breakdown. If your style mix is getting shorter-run, this bucket is quietly growing.
Pricing the map in ₹
You need one number: labour cost per attended minute, straight from payroll (the derivation is in what 1% of efficiency is worth).
Illustrative, one 30-operator line:
- 30 operators × 480 minutes = 14,400 attended minutes/shift.
- Illustrative all-in labour ₹15,000/operator/month ÷ 26 days ≈ ₹577/day → ÷ 480 ≈ ₹1.20 per attended minute (illustrative; use your own).
- Say the line runs at 50% efficiency: 7,200 off-standard minutes a shift, costing
7,200 × ₹1.20 =₹8,640/shift in paid-for, non-producing time.
Now split that by bucket from a lost-time study — an illustrative shift might look like:
| Bucket | Minutes lost (illustrative) | Cost @ ₹1.20/min |
|---|---|---|
| No feeding | 2,500 | ₹3,000 |
| Bundle-wait | 1,900 | ₹2,280 |
| Style changeover + ramp | 1,500 | ₹1,800 |
| Machine breakdown | 800 | ₹960 |
| Rework | 500 | ₹600 |
| Total off-standard | 7,200 | ₹8,640/shift |
Over 26 days that single line's leak is on the order of ₹2.2 lakh a month — and the ranking tells you to fix feeding and bundle flow first, not to chase the mechanic. The ranking is the point; the exact split is yours to measure.
How to measure it without stopping the line
Three methods, cheapest first:
- Hourly production board. Already on most lines. A dip across many stations at once = a feeding or changeover loss; a single dead station = a breakdown. Crude, but it localises the bucket.
- Lost-time capture study. An IE or supervisor walks the line at set intervals and codes each idle station against the buckets above. A day or two of this reveals the ranking. Labour-intensive and sampled, so it misses the short, frequent gaps.
- Continuous observation. The short feeding gaps and micro-stops — the ones that dominate the "no feeding" and "bundle-wait" buckets — are exactly what sampling misses, because they're brief and constant. Continuous capture of running-vs-stopped, by station and by hour, is the only way to count them honestly. (This is the same blind spot as machine performance losses in OEE: the gaps nobody logs are the ones that sink the number.)
Flow, not people
A warning that decides whether this is useful or toxic: you are measuring the line, not the operator. "Station 7 was starved for 90 minutes" is a flow problem — cutting was late, or the line is unbalanced. It is not a verdict on the person at station 7, who was sitting idle because you didn't feed them. Keep the language and the metrics on the stream of work: where it stalled, how long, and why. The moment a non-productive-time map becomes a stick to beat individuals, operators learn to look busy, and your data dies.
A precise map of your floor
Walk any sewing hall and you can see these buckets — the quiet starved section, the mechanic crouched at a dead machine, the bundles mounded at one station. What you can't do by eye is add them up across every line and every hour of every shift, which is what you'd need to know where to spend first.
That is what Mama does. It watches the video from your floor cameras and each morning messages the owner, in plain language on WhatsApp, where the lines lost running hours yesterday — the starvation, the stopped line, the changeover that ran long — and roughly what it was worth in ₹ at your own labour cost per minute. It reports flow, never people. It runs on your existing cameras, or we place our own.
For your actual map rather than an illustrative one: send a short walk-through video of your shop floor and we'll return where the minutes are leaking, a camera plan to watch those spots, and a proposal. Then read the cash your line's WIP holds for the bundle-wait bucket's second, hidden cost — and production line monitoring with cameras for how continuous output visibility works.
FAQ
How much of a sewing line's time is non-productive? Commonly estimated at 25–45% of attended time on typical Indian lines (directional industry estimate, not survey data). The only reliable figure is your own: attended minutes minus earned standard minutes.
Which bucket is usually biggest? On many floors it's feeding and bundle-wait — the line starved or work stuck between unbalanced stations — rather than breakdowns. But rank your own; the whole value is in knowing which to fix first.
Do I need an MES to measure this? No. Start with the hourly board and a short lost-time study. The limit is that sampling misses short, frequent gaps; continuous capture closes that, which is where camera-based observation helps.
Isn't this just surveillance of workers? No, if it's done right. The measurement is of the line and the flow of work — when and where the stream stalled — not a ranking of individuals. Starvation and breakdowns are management problems, not operator ones.