Style Changeover on a Sewing Line: The First Hour, the First Three Days and the ₹ Cost

A style changeover on a garment sewing line costs far more than the hour or two the line stands still. The bigger bill comes afterwards, in the two or three days the line takes to climb back to its normal efficiency. In our worked example for a 40-operator line, one changeover consumes about 1.25 line-days of paid labour, roughly ₹40,000. Better preparation and feeding can bring that down to about ₹19,000.

By the Mama Editorial Team. Mama scopes camera and floor-visibility projects for Indian factories. For owners of Indian garment units with 200 to 1,000 operators. All ₹ figures are worked examples on stated assumptions, not benchmarks. Replace them with your own wage, SAM and output data.

Key takeaways

• A changeover has three costs, not one: the dead gap between the old style's last piece and the new style's first, the slow first hour while the line fills, and a learning ramp lasting several days.
• The ramp is usually the larger cost. One widely used trade ramp (30%, 40%, 50%, then 60%) loses a full line-day of output over three days. That is more than a two-hour gap.
• Most of the gap is decided before the line stops. A pre-production meeting that ends with real decisions, cut work at the line ahead of time, and special machines set up offline all shorten it.
• Catching up with overtime more than doubles the bill. Overtime needs the worker's consent and is paid at twice the ordinary rate of wages, on top of the wages already paid for the lost time.
• Cameras can time the gap and the restart. They can't explain the ramp. Video shows when the line stopped, when bundles arrived and where work piled up. Operation-level learning still comes from your IE data.

Why this matters more every season

The shorter your orders, the more often the same line changes style. Online Clothing Study, a long-running Indian apparel IE site, describes short-run orders as those that run on a line for only 3–5 days. It notes that a line given 4–5 such orders in a month loses productive time to set-up and style changeover each time. In the site's order-quantity example, a 4-day order averages 45% line efficiency and a 10-day order 54%, on the same ramp. The author's conclusion: "when a product run becomes longer, the average line efficiency increases."

Our map of where your sewing minutes go names changeover as one bucket of lost time, and the SAM capacity check shows the start-up loss eating quoted capacity. Here we take one changeover apart.

The anatomy of one changeover

Part 1: the dead gap

The standard definition, given in this Online Clothing Study guide, is the time between the last piece of the previous style leaving the last machine and the first piece of the new style passing output. Its worked example has Style A's last piece out at 12:30 and Style B's first piece through at 14:30: two hours with nothing coming off the line.

During the gap the line is in three states:

• The tail of the old style. The end of the line finishes the last bundles; the front is already out of work.
• The re-lay. Mechanics move machines, change folders and attachments, reset stitch and thread; the IE reassigns operations.
• The first bundles of the new style. They arrive only if cutting has delivered them. The same guide lists "cut part feeding not available" among the causes of long changeovers.

Part 2: the first hour (line fill)

After restart, end-of-line output stays near zero until the first bundles pass every operation. Meanwhile, work builds up in front of the slowest new operation: the bottleneck the IE predicted, or one nobody expected. That pile is the first sign of a wrong balance, and cash sitting still (see the cash your WIP holds).

Part 3: the learning ramp

Even with every machine set correctly, operators sew a new operation slowly at first. Prasanta Sarkar of Online Clothing Study puts this down to a "new operation (job), a different machine with different settings and guide" and "a different kind material need to handle." His conservative ramp runs at 30% line efficiency on day 1, 40% on day 2, 50% on day 3 and 60% from day 4. He also notes that the build-up normally happens over the first 1–3 days of loading.

Treat these as illustrative, not a benchmark. The shape matters: the ramp lasts days and hits every operator at once.

What the research says about cutting it

The best published evidence comes from garment factories in developing countries, much of it from Bangladesh:

• Across three action-research cases, Mohammad, Hamja and Hasle (International Journal of Lean Six Sigma, 2024) report that SMED combined with a RACI responsibility matrix cut changeover time by 50% to 64%. They measured it with two indicators: throughput time and time to reach peak production. The second one counts the ramp, not just the gap.
• In two Bangladeshi apparel case organisations, Ali and co-authors (Benchmarking: An International Journal, published online 2024) cut changeover time by 70.76%, from 434.56 to 127.08 minutes, and by 42.12%, from 2,664 to 1,542 minutes.

The method is ordinary SMED: list every changeover task, move as many as possible from "internal" (line stopped) to "external" (done beforehand), and give each a named owner.

Before the line stops: PP meeting and feeding

Make the PP meeting about the changeover

Online Clothing Study's description lists the factory manager, production manager, QA, R&D, IE, merchandiser, pattern master and machine mechanic in-charge as attendees. It is held after the PP sample is approved and materials are sourced, to review construction, flag problems, confirm fabric and trims are in-house and assign responsibilities. Add four changeover decisions to it:

1. Machine list and layout for the new style, with gaps against the current layout. Which special machines, folders and attachments are needed, and which will be set up and test-sewn offline on spare machines beforehand. The changeover guide above recommends allotting spare machines for exactly this.
2. The operation bulletin and the expected bottleneck. The IE names the operation most likely to choke the line, and who will run it.
3. Cut-work readiness date. Online Clothing Study suggests having at least two days' cutting WIP ready before loading. Agree the date cutting will deliver it, and to which line.
4. Day-wise targets for the ramp. Write down the expected output for day 1, 2 and 3 so that a slow day 2 is spotted on day 2, not at shipment.

Feed the new style while the old one tails off

The cheapest gap reduction is back-to-back loading: each operator who finishes the old style starts the new one instead of waiting for the line to clear. It works only if new-style bundles already sit at the head of the line.

The ₹ cost of one changeover: a worked example

Line and wage assumptions. A 40-operator line, SAM 12 minutes, an 8-hour (480-minute) shift and steady efficiency of 60%. That gives 40 × 480 × 0.60 ÷ 12 = 960 pieces a day.

For wages we use the Haryana Labour Department notification No. 2/25/26-2 Lab of 9 April 2026, issued under the Code on Wages. It fixes the basic minimum from 1 April 2026 at ₹16,780.74 a month for a semi-skilled worker. It is CPI-linked, so check for a later revision, and use your own state's order and grading if you are outside Haryana. Assume an all-in cost of 1.25× that wage for the employer's EPF/ESI share, leave and bonus. That 1.25× is our assumption, not a statutory figure, and comes to about ₹21,000 per operator a month. At 26 days that is ₹32,308 of labour per line-day, or about ₹1.68 per operator-minute.

Two scenarios. A is a typical unprepared changeover: a two-hour gap and the 30/40/50/60 ramp. B is a prepared one: a 45-minute gap thanks to offline set-up and cut work waiting at the line, and an assumed ramp starting at 40% because operations went to operators who had sewn them before.

The difference is about ₹21,200 per changeover. A line taking four short-run orders a month changes over four times. Across a 10-line floor that is 40 changeovers and roughly ₹8.5 lakh a month of paid labour, on these assumptions. Run your own figures in the changeover time cost calculator, adding the ramp to the changeover minutes.

What the table leaves out. It counts labour only, not lost contribution, first-piece rejects or delivery penalties. And many units recover the loss with overtime. Under the OSH Code, in force since 21 November 2025, overtime needs the worker's consent and is paid at twice the ordinary rate of wages, within the hours your state allows. "Wages" here means basic plus DA, so at the Haryana basic minimum (₹645.41 a day) recovering scenario A's 1.25 line-days costs about ₹64,500 in overtime wages (1.25 × 40 × ₹645.41 × 2), on top of the ₹40,400 already paid. The changeover is cheaper to fix than to pay for twice.

What cameras show, and what only IE data can tell you

Cameras on the end-line checkpoint and the head of the line can time most of the anatomy above:

• The real gap. The time the last old-style piece crossed the checkpoint and the time the first new-style piece did, from footage, not memory.
• Feeding. Whether cut bundles of the new style were waiting at the head of the line when the old style ran out, or arrived 40 minutes later on a trolley.
• The re-lay. How long machines sat unused while mechanics reset them.
• Restart and fill. When stations restarted, and where bundles piled up in the first hour.
• Return to rhythm. Whether hourly output at the checkpoint got back to normal on day 2 or day 4, line by line and changeover by changeover.

What video cannot tell you:

• SAM and the operation bulletin. Whether the new style's standard minutes are right, or how they split across operations, is IE work.
• Why an operation is slow. A camera sees the pile, not whether the cause is a folder setting, a hard fabric or the method.
• Quality of the first pieces. First-day defects come from your inline and end-line inspection data (see end-line DHU and the alteration loop).
• Upstream reasons. Late approvals or trims stuck in store happen off the floor.
• Individual learning. Video should look at the line, not rank people. Track operator skill in your skill matrix, not on camera.

If cameras cover work areas, workers must be told what is recorded and why. Our worker CCTV notice under DPDP has a template.

Where Mama fits

Mama reads the video from the cameras you already have, or a few added ones, and every morning sends the owner a short plain-language note on where the sewing floor lost running hours the day before. That includes changeovers: how long each line was dead between styles, whether new-style bundles were waiting, and how many days the line took to return to rhythm. Mama watches lines, machines and the flow of work, never individual operators, and doesn't rank people. Mama can draw a camera plan for the sewing lines and bottleneck operations where changeover hours pile up. Leave an email or phone in the form below.

FAQ

How long should a style changeover take on a sewing line? The Online Clothing Study guide sets single-digit minutes as the SMED goal, but a whole-line style change with a new layout usually takes longer. Published garment-factory case studies report cuts of 42–71% from each factory's own baseline, so measure yours first. The gap is the time from the last old-style piece to the first new-style piece at output.

Is the gap or the learning ramp the bigger loss? Usually the ramp. On a 30/40/50/60% ramp, the first three days lose a full line-day against steady output, four times a two-hour gap.

What should a pre-production meeting decide about the changeover? The machine list and layout gaps (and which machines get set up offline), the expected bottleneck operation, the date cut work reaches the line, and day-wise output targets for the first three days.

Does back-to-back loading really help? Yes, if new-style bundles are already at the head of the line, so the gap overlaps the old order's tail.

Can cameras measure the learning curve? They can show when line output at the checkpoint returns to its normal hourly rhythm, changeover by changeover. They can't explain operation-level learning or SAM accuracy. For that you need your IE's time studies and the operation bulletin.