Four gaps between the SMV book and the sewing line
Your IE team sees these gaps in four records: the time-study sheet, the OB, the daily efficiency sheet and the capacity report. Each gap costs output, and each one starts as a figure nobody has checked.
Stopwatch studies on paper, and an SMV book nobody re-studies
Each SMV starts as a stopwatch study on a clipboard, typed into one IE’s Excel file, or never typed at all. Then a folder is added or an operation is split, and nobody re-studies it. When a buyer quotes a lower “industry SMV” to push the CM price down, the IE manager must find the study, rating and allowance behind each figure. Sometimes they sit on the laptop of an IE who has since left.
22%
In a 2016 study of Pakistani and Bangladeshi factories, one Bangladeshi t-shirt factory set its SMVs about 22% higher than another for the same operations. Factories may use different SMV standards, the authors note.
Source: International Growth Centre (IGC), Understanding productivity in the Pakistani garment sector (opens in a new tab), 2016.
Efficiency is on the board every day, but the month is rebuilt by hand
Each line’s output and efficiency go on the board and into a daily sheet, hour by hour. For the monthly meeting, someone copies a month of daily figures into one summary by hand. Even then, nobody can say why line 7 trails line 3: the style, the SMV or the headcount. The 2014 study behind the figure found the best of ten lines in a typical unit two-thirds more productive than the worst.
Around 45%
Average efficiency in a 2014 study of around 60 Bangladeshi garment factories, all suppliers to mid- and higher-end international buyers.
Source: DEGRP (ODI and ESRC-DFID), hosted by the Centre for Policy Dialogue, Enhancing productivity in Bangladesh’s garment sector: current policy and research debates (opens in a new tab), 2014.
The bottleneck moves every morning with the absentees
The OB and the layout assume a full line. Each morning some operators are absent, sometimes the one skilled operator on a long operation, and the supervisor reshuffles by feel. The IE re-balances on paper or in Excel after the line has already started, and the operators after the new bottleneck sit waiting for bundles, while work piles up in front of it.
2–3% to 6–8%
Daily absenteeism reported by managers of three Cambodian piece-rate garment factories in December 2019 interviews. Daily-wage factories reported 3–5%.
Source: Better Work (ILO and IFC), Discussion Paper 44, Piece Rate, Productivity, and Occupational Health in the Global Economy: Mixed Method Evidence from Cambodian Garment Factories (opens in a new tab), 2021.
A new style starts slow, but capacity was booked at full speed
Orders are smaller and styles change more often. Each changeover costs set-up time, then days of low output while operators get used to the new operations. If capacity was booked at the line’s usual efficiency, the gap shows in week one, the IE is asked why the SMV “is wrong”, and Planning and the merchandiser carry the risk to the ship date.
7–8 working days
Time an assembly line may take to reach peak output on a new style, in a NIFT worked example for 5-pocket jeans; a modular line, 3–4 days. The authors note there is no standard curve.
Source: Apparel Resources, IE in Apparel Manufacturing-9: Minimizing Start-up Losses (opens in a new tab), 2015.
How the IE agent works from your SMV book and line sheets
It reads your SMV sheets, OBs, capacity reports and incentive scheme, and calculates from the output sheets you upload, showing its files and rows. A message to the line or an email to Planning runs only if an admin approves it.
The SMV on record, operation by operation
Answers with sources
Upload your SMV sheets and OBs, or connect the Google Sheet your IE team keeps them in. Ask for a style and the agent lists each operation with its method, observed time, rating, allowance and SMV, and the study date your sheet records, citing the file and rows. Where an operation has no study on record, it says so and shows what the OB uses instead.
Ask“What SMV is on record for style 4471, operation by operation, and when was each one last studied?”
Line efficiency for the month, calculated from your rows
Answers with sources
Upload the daily line efficiency sheet, with each line’s efficiency for each day already on it, as a CSV file or the first sheet of an Excel file: one row per line per day. A workbook with a tab for each day needs combining first. Ask about one line and the agent averages its daily figures from the rows, gives its highest or lowest day, or sums its output for one style, citing the rows behind each figure.
Ask“From October’s efficiency sheet, what was line 7’s average efficiency for the month?”
A line-balancing draft from the OB on record
Creates files
The agent reads the OB on record, with each operation’s SMV and machine type, and drafts a balancing sheet: operators per operation, the station it reads as the bottleneck, and every assumption stated, such as whether helpers are counted. It is a draft for your IE to check against the line. Ask it to move an operator or merge two operations, and it revises only that part; earlier versions are kept.
Ask“Style 3308 moves to line 4 next week with 30 operators. Draft a line-balancing sheet from its OB and state your assumptions.”
First-week targets for a new style, set to the learning curve
Creates files
Ask before a style loads, and the agent lists what is still open on the pre-style checklist you uploaded: OB, SMV per operation, layout, pilot target. It drafts first-week targets from its OB at the playbook’s ramp, about 60% of standard output in week one and 85% in week two, with each assumption stated. An email of the targets to Planning waits for an admin to approve or decline it; it is sent only if approved.
Ask“Style 6012 loads on line 5 on Monday. What is still open on its pre-style checklist? Draft first-week targets from its OB and email them to Planning.”
The incentive scheme, read slab by slab
Reads your documents
Upload the incentive scheme with its slabs, thresholds and DHU conditions. Ask where a line stands and the agent quotes the slab and clause that apply, beside the line’s output and QC-passed output summed from the sheets your team uploads. Wage structures and payroll stay with Human Resources; the IE agent cannot read HR’s files.
Ask“Line 3 averaged 68% in October. Which slab of the incentive scheme applies, and what does the scheme say about DHU?”
From the morning headcount to the kaizen review
A buyer pushes back on your SMV
IE manager
The buyer’s costing team says the “industry SMV” for a basic polo is lower than yours and asks for a lower CM.
Asks
“The buyer says our SMV for style 5120 is too high. Show each operation’s SMV, the rating and allowance used, and when it was studied.”
The agent
Reads the SMV sheet and the OB for style 5120 and lists each operation with its method, observed time, rating, PFD allowance, SMV and study date, citing the rows. Where a value is not recorded, it says so. It marks the operations whose study date it reads as more than six months old, for the IE manager to confirm, and quotes the playbook: industry SMV is style-specific, so answer with your own SMV book. It builds the list as a document.
Result
The IE manager answers with the study behind each SMV, and a short list of operations to re-study before the next costing.
Four operators short on line 6
IE officer
Line 6 starts the shift with 34 of its 38 operators, and one of the absentees runs the collar operation.
Asks
“Line 6 has 34 operators today on style 2210. Draft a balance from the OB and send the allocation to the line supervisor on WhatsApp.”
The agent
Reads the OB for style 2210, the line layout and the manpower sheet uploaded this morning, and reads from its rows who is present on each operation. It drafts the balance for 34 as a spreadsheet, names the station it reads as the bottleneck from the SMVs on record, and states each assumption. It then drafts a WhatsApp Business message to the line supervisor with the allocation.
Approval
The message waits for the IE manager, as department admin, who approves or declines it. It is sent to the supervisor only if approved, and the audit trail keeps the IE manager’s decision.
Result
A balance for 34 operators that the IE checks before anyone moves, with the bottleneck station named, instead of a reshuffle done by feel after the first hour.
Did the folder kaizen on line 2 hold?
IE manager
The weekly kaizen review, three weeks after a folder was fitted on line 2’s hemming operation.
Asks
“What was the line 2 hemming kaizen expected to deliver? And what was average hourly output at hemming in weeks 40 and 44, on shift A and on shift B?”
The agent
Reads the kaizen log entry and the method sheet: issue, root cause, intervention and the output it was expected to deliver. From the station-wise hourly sheet uploaded for line 2, it works out four averages of hourly output, each over the rows where station is hemming and the week and shift match: weeks 40 and 44, shifts A and B. It lists the four figures with their rows and builds the close-out note as a document.
Result
A close-out note with week 40 beside week 44 for each shift, all from your rows. If one shift’s output did not move, you see it before the kaizen is marked closed.
What a point of efficiency can be worth, and what to baseline
These are published figures from ILO, Better Work and IGC studies of garment factories, each cited at the source. None is a Mavick result. Before a pilot, take your own baseline for the IE KPIs below; the pilot is read against it, not against these studies.
Each line’s month, worked out from its daily rows
US$48,288 a year
What each 1% of efficiency was worth to one Thai sportswear factory with 382 sewing operators, starting from 46%: a saving potential modelled in a 2017 ILO paper.
How the agent helps
The agent does not count pieces on the line. Ask about a line and it averages the daily efficiency figures your uploaded sheet records for it, or sums its output for the month, and sets the style and the SMV on record beside it. Line by line, the IE starts from the lines and styles that fall short, with the rows behind each figure.
Track: Operator efficiency %, by line and month
Source: International Labour Organization (ILO), Redistributing value added towards labour in apparel supply chains: Tackling low wages through purchasing practices (opens in a new tab), 2017.
SMVs that hold up when the price is argued
Up to 1.33×
How far apart two Bangladeshi factories set SMVs for the same three styles from one European buyer, in a 2016 IGC working paper (1.08× to 1.33×). The authors say helper ratios explain part of the gap.
How the agent helps
Ask for a style and the agent lists each operation’s SMV, rating, allowance and study date from your SMV sheet, and marks the operations whose study date it reads as more than six months old. It builds the re-study list as a spreadsheet, so your IE can see which SMVs behind a costing are due for a re-study.
Track: SMV currency: share of operations re-studied within six months
Source: International Growth Centre (IGC), Understanding productivity in the Pakistani garment sector (opens in a new tab), 2016.
Incentives that pay on QC-passed output
Nearly 50%
Share of Cambodian garment workers paid piece rate, in Better Work worker surveys from 2015–2018 (paper published 2021). Managers interviewed for it said quality has a greater potential to decline when pay rewards speed.
How the agent helps
The playbook ties bonus to QC-passed output only, with DHU above a threshold reducing it. The agent quotes your scheme’s slabs and conditions, and sums a line’s output, or its QC-passed output, for the month from the sheets you upload. The IE then sees whether the scheme pays for pieces QC did not pass.
Track: Incentive payout % against design
Source: Better Work (ILO and IFC), Discussion Paper 44, Piece Rate, Productivity, and Occupational Health in the Global Economy: Mixed Method Evidence from Cambodian Garment Factories (opens in a new tab), 2021.
What to baseline before a pilot
Measure each one before the pilot starts, then again at its end.
| KPI to baseline | Why it matters |
|---|---|
| Operator efficiency %, by line | The playbook’s target is 75% or more, measured as output × SMV over available minutes. It is only as honest as the SMV behind it. |
| SMV currency: operations re-studied within six months | The playbook sets 80% or more, and a quarterly re-study of the top three bottleneck operations on each line. An SMV that has never been re-studied, it says, is fiction. |
| Kaizen pace: closed kaizens per line per month | The playbook’s target is at least one a month on every line, each documented with the method and the output before and after. |
| Incentive payout % against design | Read monthly. Above 120% of design, the playbook calls the scheme too generous; below 80%, too tight or a skill mismatch. |
| Output against target on confirmed-PO styles | When a line misses target by more than 15% for over two days, the playbook’s rule is to take it to IE, PPC and the plant head, to re-balance or re-plan. |
| Line balance ratio, by line and month | The playbook reports it per line each month and aims for 85% or more. It warns that a ratio above 90% usually means slack in the bottleneck. |
What the agent works from
Records it reads
- SMV sheet
- Operation breakdown (OB)
- Time-study sheet
- Capacity report
- Line layout
- Daily line efficiency sheet
- Kaizen log
- Incentive scheme
In your ERP
From Odoo, the IE agent reads production orders by order, product, quantity and status, and stock transfers by partner, status, scheduled date and source document. It can draft a new planned start, or a new assignee, for a style’s Odoo production order, or a note on a transfer; an admin approves or declines it. SMVs, OBs and manpower come from your sheets.
One ERP connection serves the whole factory: SAP Business One, SAP S/4HANA, Odoo, Oracle NetSuite, Microsoft Dynamics 365, WFX Cloud ERP or Datatex NOW. This agent reads only the records named above; we check yours on the call.
In your tools
Reads the IE team’s Google Sheets when you ask: the SMV library, the hourly output sheet, the manpower sheet. Prepares a row for the kaizen log, a WhatsApp Business or Slack message with a line’s allocation, or a Gmail or Outlook email listing the operations to re-study. Nothing is sent or written unless an admin approves it; a declined draft never runs.
Tools named today: Gmail, Outlook, Google Sheets, Slack and WhatsApp Business.
Never made up
- SMV values, time-study readings or allowance percentages
- Capacity figures, machine counts or manpower numbers
- Efficiency percentages or output rates
- Incentive rates or payout calculations
If it is not in your connected data, the agent says so.
Belongs to another department
- Running production on the line, and line supervisionGarmenting & Stitching
- Order sequencing and ship datesPlanning & PPC
- Wage structures and payrollHuman Resources
- Machine repair and availabilityMaintenance & Engineering
What the Industrial Engineering playbook covers
SOP: setting an SMV from a time study
At least 10 cycles per operator, and at least three operators per operation. Rate every cycle against 100 as standard pace: an unrated study only records the speed of whoever was timed. Basic time is observed time × rating. SMV is basic time × (1 + allowances), with PFD allowances typically 15–20%.
Decision rule: a line stuck at 80% of plan
The playbook gives two causes: the SMV is wrong, or the operators are below skill. Test the SMV first, by re-studying the bottleneck station with your best operator. If the SMV holds, re-train or re-allocate the operators.
KPI: line balance ratio, aimed at 85%
The sum of station SMVs, divided by the number of stations times the bottleneck station’s SMV. The target is 85% or more. Above 90% usually means slack in the bottleneck, the playbook warns. Below 70% for a sustained period goes to the IE manager and the production head, for a redesign within a week.
Edge case: the first weeks of a new style
Expect output at about 60% of the book-SMV rate in week one, 85% by week two and 90–100% by weeks three to four, depending on style complexity. First-day output is not the SMV’s fault, so the playbook says not to re-study before week three.
Written as general textile practice and reviewed by four independent AI models. Not yet signed off by a human subject-matter expert. Upload your own SOPs and standards, and every answer shows which source it used.
Ten questions it ships with
The template’s starter questions. Your team can edit them, or ask in its own words.
SMV on record for a style
What is the SMV on record for this style, how was it broken down by operation, and how does actual time compare?
Bottleneck operation on a line
Which operation is the bottleneck on this line, what does the capacity study show, and how much output is it costing?
Capacity available for an order
What capacity is available for this order at our current efficiency, and how many days would it take at the planned manpower?
Kaizen result against expectation
What was this kaizen or layout change expected to deliver, and what did the production records show afterwards?
Incentive payout at current output
At current output levels, what does the incentive scheme pay out for this line? Show the calculation from the scheme on file.
Line-balancing proposal
Draft a line-balancing proposal for this style from the operation breakdown on record, and state the assumptions you used.
Efficiency trend by line
Show the efficiency trend by line over the last month and tell me which lines improved and which slipped.
Manpower against standard
Compare the manpower deployed on this line with the standard manning for the style, operation by operation.
Operations without a time study
Which operations on this style have no time study on record, and what is being used for them instead?
Machine utilisation by type
Show machine utilisation by machine type for this month, and identify which types are underused.
Industrial Engineering: questions before a demo
Our time studies are on paper clipboards. Can the agent still use them?
Only once they are typed in. A photo of a study sheet, or a PDF made from one, has no text for the agent to read. Key the cycle times into Excel or a Google Sheet, one row per cycle with the operation and the operator, then upload or connect it. From an uploaded sheet, the agent calculates the average, lowest or highest observed time for an operation from its rows. It does not run the study or rate the operator: your IE still does that.
Can it set the SMV for a new style from the tech pack?
No. It does not work out an SMV for a style nobody has studied, and it has no GSD or other predetermined time codes of its own. It reads the SMVs your IE team set. For a new style, it can list the operations in your SMV library with the same name, with their SMVs and the date each was studied, for your IE to judge. The playbook’s pre-style checklist still applies: OB drafted from the tech pack, SMV set per operation, layout reviewed.
Can it time an operation from a video of the operator?
No. It cannot open a video file, and it does not time cycles, rate an operator’s pace or count elements. Your IE still films or times the operation. Once the readings are typed into Excel or a Google Sheet, or exported from your time-study software as an Excel or CSV file, the agent reads them like any other time-study sheet, citing the rows. A table holds up to 5,000 rows, so upload one style or one month of studies at a time if your file is larger.
Planning asks if line 4 can take 12,000 pieces by the 20th. Can the agent answer?
It reads the capacity report on file for that line and garment family, and the SMV on record for the style, and shows both with their sources. It can draft a sheet for your IE to check, setting out each input with its source (available minutes, the efficiency assumed, the SMV) and its own working of the days needed: arithmetic the agent writes into the draft, not a figure calculated from your rows. If your capacity report does not cover the style’s garment family, it says so instead of guessing. It does not forecast or schedule: order sequencing and the ship date stay with Production Planning.
Will the IE agent see payroll, or what each operator is paid?
Not from HR. Wage structures and payroll belong to Human Resources, and the IE agent cannot read HR’s files; asked about pay, it says HR owns it. Salary and bank-account fields are removed from ERP records before the agent reads them. It sees only what is connected to Industrial Engineering, such as your incentive scheme and the output and QC-passed output sheets your team uploads. If you upload a payout sheet to IE, it can read that sheet. It never pays anyone or changes a wage.
Can it re-balance a line by itself when operators are absent?
No. It drafts a balance when someone asks, from the OB on record and that day’s manpower sheet, and states every assumption it used, such as which operators it treated as multi-skilled. Your IE checks it and decides. It does not run an optimisation or move anyone. If you ask it to send the allocation to the line supervisor on WhatsApp Business or Slack, the message waits for an admin, who approves or declines it, and it is sent only if approved.
Bring Industrial Engineering. We’ll build it on the call.
Book a 30-minute demo (opens in a new tab)- support@mavick.si
- Demo calendar
- calendly.com/safock/30min (opens in a new tab)