Skip to content

Garmenting & stitching: each line’s hourly output and DHU, with sources

Ask how a line ran hour by hour, its DHU and the defects behind it, or where an order’s cut pieces are. The agent answers from your hourly sheets, end-line sheets and registers, shows the rows it used, and drafts any message for an admin to approve or decline.

Works from: Cut plan, Bundle card, Hourly production sheet, Daily manpower sheet

Where the cutting room and the lines lose minutes and pieces

Each problem below lives in one record: the hourly board, the end-line tally, the manpower sheet or the bundle register. Many are still paper or a supervisor’s own sheet, and the cost shows up as lost minutes, overtime and short shipments.

  • The hourly board is filled in late, and the gap shows at shift end

    The line supervisor writes target and actual on the board each hour, often late and sometimes several hours at once. By the time the floor in-charge sees the gap, the hour where the balance broke has passed. Its cause (an absent operator at the bottleneck, a machine down, bundles late from cutting) stays in someone’s memory. The playbook says to step in as soon as an hour’s output falls more than 15% below target.

    32%

    Share of Bangladeshi manufacturers, garments among them, that tracked no key performance indicator, in a 2019–2020 World Bank survey of about 900 formal firms with 8 or more workers, in five industries.

    Source: World Bank, Gearing Up for the Future of Manufacturing in Bangladesh (opens in a new tab), 2021.

  • DHU and rework reach the huddle after the bundles have moved on

    End-line checkers tally defects by hand, and the DHU report with its defect Pareto arrives after the shift. The playbook’s ladder needs the number while the line runs, so the huddle at DHU 3 or the line stop at 10 can happen the same shift. Until then, the operation making the open seams, skip stitches or puckering keeps making them, and every piece goes back for rework in operator minutes planned for new output.

    Around 8%

    Average share of garments needing rework, in daily line data from 33 large factories in and around Dhaka, 2012–2014 (median almost 6%). Rejects averaged only 0.6%.

    Source: International Growth Centre (IGC), The effect of political and labour unrest on productivity: Evidence from Bangladeshi garments (opens in a new tab), 2015.

  • Absent operators at key stations break the line balance

    Each morning the supervisor counts who is missing and moves operators across or within lines to cover the key operations. The bottleneck shifts, output drops and overtime follows. The playbook targets absenteeism of 6% or less and says anything over 10% breaks line balance. When key sewing positions stay empty for more than two days, the playbook’s escalation goes to the HR head and the floor in-charge.

    Over US$6 million a year

    Estimated yearly cost of unplanned absence alone to Myanmar’s garment sector, from 2020 ILO factory data scaled to the sector. The report calls it the very low end: it leaves out overtime and the time supervisors spend moving workers.

    Source: International Labour Organization (ILO), Skilled Workers Matter: The business case for addressing absenteeism and turnover in Myanmar’s garment sector (opens in a new tab), 2020.

  • Cut, bundled, sewn, packed: nobody can say where the pieces are

    The cut plan sets the order by size and colour. Cutting reports lays and bundles, the lines count input and output on their boards, and finishing and packing count again. A bundle waits for a re-cut panel or a missing trim, and nobody sees it until the order is short at packing. The cut-to-ship reconciliation is then done by hand from several registers, after shipment.

What the agent works out from your hourly sheets and registers

It works from hourly sheets, end-line records, cutting and bundle registers, tech packs and the Odoo production orders the department reads. Each answer names the rows or files it used, and any message it drafts waits for an admin to approve or decline.

  • Lines under target, and a note to their line chiefs

    Acts in your tools

    Connect the hourly production sheet your supervisors fill in Google Sheets. Ask which lines missed their hourly target since 8 am, and the agent reads each hour’s target, actual and remark from the sheet and lists the hours below target by line, citing the cells it read. It can draft a WhatsApp Business or Slack note to those line chiefs. A department admin approves or declines it, and the note goes only if approved.

    Ask“Which lines missed their hourly target since 8 am? Show target and actual for those hours, and what the supervisors logged.”

  • DHU by line and defect, counted from your inspection rows

    Answers with sources

    Upload the end-line sheet as Excel or CSV, with the tally on its first sheet. The agent counts defects from the rows for the line, week and defect type you name, as the sheet’s columns record them, sums the pieces checked and gives DHU per hundred. Counted type by type, it names the top three defects, the ones the playbook says get an operator-level fix (re-training or a mechanical fix), and cites the rows it used.

    Ask“What is line 4’s DHU this week from the end-line sheet, and which three defects make up most of it?”

  • Absentees by line and by supervisor, from your manpower sheet

    Answers with sources

    Upload the daily manpower sheet your line supervisors keep. The agent counts absences from the rows for a line, a supervisor or a month the sheet records, so the playbook’s monthly Pareto by supervisor is one count per supervisor. From the skill matrix you upload, it reads out operators trained on an operation that is short today. The supervisor decides every move. HR’s attendance and payroll files stay in HR.

    Ask“How many absences did the manpower sheet record on supervisor Karim’s lines in October?”

  • The cut-to-ship reconciliation, built as a sheet

    Creates files

    Upload the cutting register, the bundle register and the packing list as Excel or CSV files. The agent sums cut and packed quantities for the order from the rows, reads the bundle register and lists the bundles it finds with no later stage recorded, by last recorded date, oldest first, then builds the reconciliation as a spreadsheet showing which rows it used. It knows only what the registers you uploaded record.

    Ask“Reconcile cut against packed for order 5520 from the registers, and show where the missing pieces sit.”

  • The tolerance the line needs, quoted from the tech pack

    Reads your documents

    Upload the tech pack, PP-meeting notes and trim card. Asked the chest-width tolerance or the side-seam stitch density, the agent quotes the file and the version it states, so keep replaced tech packs out or marked as replaced. A scanned tech pack, or a measurement chart pasted in as a picture, has no text to read: ask the buyer for the spec sheet in Excel. It never judges a garment against the spec.

    Ask“What does the current tech pack give for chest width and its tolerance on style 2231? Name the version.”

From pieces short at packing to a DHU spike: a week on the sewing floor

  • An order’s cut quantity does not match what reached packing

    Cutting in-charge

    Tuesday afternoon: packing counts order 4417 short in sizes M and L, and the merchandiser asks whether a re-cut is needed.

    1. Asks

      “For order 4417, compare cut and packed quantities by size, and list the M and L bundles with no stage recorded after cutting.”

    2. The agent

      Sums cut quantity by size from the cutting register and packed quantity by size from the packing list, both uploaded, from the rows. Reads the bundle register for the M and L bundles and lists those it finds with no later stage recorded, with its line and last recorded date, oldest first. It cites the rows behind each figure. A bundle moved without a register entry does not show.

    3. Result

      The cutting in-charge goes to the floor with a list of bundles to find, by line and date, before anyone asks for a re-cut.

  • The day’s production summary, before the evening meeting

    Floor in-charge

    Shift end, with the day’s hourly sheets filled in and the end-line tally uploaded.

    1. Asks

      “Prepare today’s production summary from the hourly sheets, with DHU and the top three defects for lines 2, 4 and 5 from the end-line sheet, and send it to the production manager on WhatsApp.”

    2. The agent

      Reads each line’s hourly target, actual and cumulative output from the Google Sheet. For lines 2, 4 and 5, it works out DHU and the top three defects from the uploaded end-line rows. Builds a spreadsheet: output against target and efficiency as recorded for each line, DHU and top defects for those three, and the remarks logged for the hours that missed. Then drafts a WhatsApp Business message to the production manager with the main points.

    3. Approval

      The WhatsApp draft sits with the department admin, who approves or declines it. Nothing reaches the production manager unless it is approved, and the decision is recorded with who made it.

    4. Result

      A line-by-line summary for the evening meeting, each figure tied to a row, and a WhatsApp message that is sent only if approved.

  • DHU passes 5 on line 4 before lunch

    Sewing QC in-charge

    Mid-morning, the end-line checkers report more defects than usual on line 4, and their tally sheet is uploaded.

    1. Asks

      “From this morning’s end-line sheet, what is line 4’s DHU, which three defects drive it, and what did the last CAPA for this style say?”

    2. The agent

      Counts defects by type, and by operation where the sheet records it, from the uploaded rows. Divides by the pieces checked for DHU and lists the top three defects with their counts and the rows used. Reads the uploaded CAPA log for earlier actions on those defects on this style. Quotes the playbook’s ladder: above 5, the floor in-charge and IE take targeted operator action the same shift; above 10, stop the line.

    3. Result

      The QC in-charge takes the DHU, the three defects and the last CAPA on record to the floor in-charge while most of the shift is still ahead.

Where a sewing floor’s margin goes

These are published figures on garment production, cited at the source; two of the three come from Bangladesh. None is a Mavick result. Before a pilot, record each line’s DHU, efficiency and cut-to-ship gap, then compare them as the pilot runs.

  • The hour a line fell behind, found the same day

    About 51%

    Average sewing-line efficiency in daily line data from 28 large woven and light-knit factories in and around Dhaka, 2012–2014. One study’s sample, not a national benchmark.

    How the agent helps

    When the hourly board is a Google Sheet, the floor in-charge can ask at any hour which lines are under target and what was logged, and get the rows behind the answer. The agent drafts the note to the line chiefs, sent only if an admin approves. It does not balance lines; your supervisors and IE do.

    Track: Line efficiency % by line, against the playbook’s target for the garment type

    Source: International Growth Centre (IGC), The effect of political and labour unrest on productivity: Evidence from Bangladeshi garments (opens in a new tab), 2015.

  • Fewer panels cut twice: each cut piece traced to the carton

    10%

    Share of the material going into clothing worldwide that is lost during garment production, for example as offcuts, in a 2017 global estimate of clothing material flows (2015 data).

    How the agent helps

    A missing panel is often cut again, from fresh fabric. For one order, the agent sums cut quantity from the cutting register and packed quantity from the packing list you upload, and lists the bundles the bundle register shows with no later stage, to look for before a re-cut. Re-cuts and rejects show only where recorded. It counts pieces, not metres, and makes no markers.

    Track: Cut-to-ship ratio by order

    Source: Ellen MacArthur Foundation, A New Textiles Economy: Redesigning fashion’s future (opens in a new tab), 2017.

  • Last season’s puckering fix, found by a new supervisor

    15%

    Share of line supervisors who had received any formal supervisory training, in a Bangladesh survey from about 2014 cited in a 2019 ILO working paper. Another survey it cites puts the share at 20%.

    How the agent helps

    Supervisors pick up line setting, the DHU ladder and last season’s fixes on the job. Upload your SOPs, checklists and CAPA notes, and a new supervisor can ask what fixed puckering on a style last season and get the CAPA note, or a clear answer that no such note is on record. The agent trains no one; it keeps the record within reach.

    Track: Days for a new style to reach line balance (the playbook expects day 3)

    Source: International Labour Organization (ILO), Securing the competitiveness of Asia’s garment sector: A framework for enhancing factory-level productivity (opens in a new tab), 2019.

What to baseline before a pilot

Measure each one before the pilot starts, then again at its end.

Garmenting & Stitching: the KPIs to baseline before a pilot, and why each matters
KPI to baselineWhy it matters
Line efficiency % by line and garment typeThe playbook holds a denim line and a T-shirt line to different targets, so compare each line with the target for the garment it is sewing, not with the floor’s average.
DHU by line and shiftThe playbook’s target is 3 or less, read per line and per shift with its defect Pareto, so one line’s spike does not hide inside the floor’s average.
Operator absenteeism % by line and by supervisorThe playbook targets 6% or less and says over 10% breaks line balance. It asks for the absenteeism Pareto by supervisor, not by line, every month.
First-pass FRI pass rateThe playbook targets 98% or more, and says the end-line audit catches more than the FRI, which is all buyers see. The FRI verdict itself is Quality’s to report.
Throughput per line per day, against the SMV-based targetThe playbook’s line target is available minutes divided by the bottleneck SMV, times line efficiency. Reading output against it gives each line’s daily gap in pieces.

What the agent works from

Records it reads

  • Cut plan
  • Bundle card
  • Hourly production sheet
  • Daily manpower sheet
  • DHU report
  • End-line audit report
  • Tech pack
  • CAPA log

In your ERP

Reads Odoo production orders (order, product, quantity, status) and stock transfers (partner, status, scheduled date, source document). Can draft, in Odoo, a production order’s planned start or assignee, or a transfer’s note or source document, for an admin to approve or decline. Output, DHU and bundles come from your sheets. No other ERP gives it records.

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 hourly or manpower sheet in Google Sheets when someone asks. Prepares the daily summary or a huddle note as a WhatsApp Business or Slack message to line chiefs, a row of the day’s totals for your production sheet, or a CAPA email to send from Gmail or Outlook. Each waits in Mavick for an admin to approve or decline, and goes only if approved.

Tools named today: Gmail, Outlook, Google Sheets, Slack and WhatsApp Business.

Never made up

  • Hourly or daily output figures and line efficiency
  • Cut quantities, bundle numbers or WIP positions
  • DHU figures and defect counts
  • Audit results, measurement readings or buyer approvals

If it is not in your connected data, the agent says so.

Belongs to another department

What the Garmenting & Stitching playbook covers

  • SOP: DHU control, one ladder on every line

    DHU counts defects per hundred units, so it can pass 100, and it is measured at end-line and at final. Above 3, hold a quality huddle and act on the top three defects. Above 5, the floor in-charge and IE act on targeted operators the same shift. Above 10, stop the line, find the root cause and restart only after the fix.

  • Decision rule: a line below target by mid-day

    Find the bottleneck first: the station with the longest WIP queue. Then choose a path. Re-balance by moving operators between stations, which can be done the same day. Or re-engineer the operation breakdown and SMV with IE, which takes until the next day.

  • KPI: line efficiency, set by the garment on the line

    The playbook’s efficiency targets differ by garment: 55% or more for woven shirts, 65% or more for T-shirts and 50% or more for denim bottoms. The playbook asks the floor to report efficiency, DHU and the top three defects for each line every shift, and to review each line’s efficiency week by week.

  • Heuristic: when not to judge a line yet

    Line balance on a new style comes on the third day, so plan for a learning curve. The playbook also reads a Monday-morning DHU as half absenteeism, not skill, and says to wait until Tuesday afternoon before judging it.

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.

  • Hourly output against target

    Show hourly output against target for this line yesterday, and tell me which hour the balance broke and what was logged.

  • Cut plan progress for an order

    For this order, how much is cut, bundled, sewn, and packed against the cut plan? Show the quantities at each stage.

  • DHU by defect for a line

    What is the DHU for this line this week, and which defect types account for most of it?

  • End-of-line audit findings

    What did the end-of-line audit find for this style, and has the corrective action been closed?

  • Bundles stuck at a stage

    Which bundles are stuck at a stage, at which stage, and for how long? Show the oldest first.

  • Measurement spec from the tech pack

    What does the current tech pack specify for this style’s critical measurements and tolerances? Cite the version.

  • Line efficiency across the floor

    Rank the sewing lines by efficiency for the last week and show the output and manpower recorded for each.

  • Rework quantity and reasons

    How many pieces went to rework this week, what reasons were recorded, and which lines do they trace to?

  • Daily production summary

    Prepare a daily production summary from the hourly boards: output by line, efficiency, DHU, and anything that needs attention tomorrow.

  • Cut-to-ship reconciliation

    Reconcile cut quantity against packed quantity for this order and account for the difference from the records.

Garmenting & Stitching: questions before a demo

Our hourly boards are on paper. Can the agent still help?

Only once the numbers are typed in somewhere it can read. It cannot read a whiteboard, and a photo of the board has no text to read. If supervisors fill a Google Sheet each hour, connect it, and the agent reads it whenever someone asks. A daily Excel or CSV export works too: from an uploaded file (for Excel, the first sheet) it sums output and counts defects from the rows. Meanwhile, upload the SOPs, tech packs and audit reports you already keep as files and start there.

Can it count pieces from our sewing machines or read bundle barcodes?

No. Mavick does not connect to sewing machines, piece counters, barcode or RFID systems, and it does not inspect garments with cameras. It reads what your factory records: the hourly sheet, the bundle register and the end-line sheet you connect or upload, and the Odoo production orders and stock transfers the department reads. If another system exports its piece counts to an Excel or CSV file, upload it and the agent reads that file.

Will it balance our lines or tell supervisors where to move operators?

No. It does not compute a line balance or assign operators. Ask, and it lists what is on record: each hour’s output against target, and the operators your skill matrix shows as trained on an operation. It also quotes the playbook’s two paths for a line behind by mid-day: re-balance the same day, or re-engineer the OB with IE overnight. The supervisor and the floor in-charge decide the moves. SMV, line layout and capacity studies belong to Industrial Engineering.

Can supervisors ask the agent from the floor on WhatsApp?

No. Members ask in the Mavick workspace, and the agent does not answer messages sent to your WhatsApp Business number. When a member asks, it can read the messages there, such as a line chief’s note about a machine down. It can also draft a WhatsApp Business message to a line chief, which is sent only if an admin approves it.

Can the agent change our production orders in the ERP?

Only a few fields, and only in Odoo. It can draft a production order’s new planned start or the person it is assigned to, or a note or source document on a stock transfer. Each draft waits for an admin to approve or decline it, and it runs only if approved. It never writes output, DHU, bundle counts or stock quantities to any ERP, and a buyer’s delivery date is merchandising’s to agree with the buyer. Other ERPs take no writes from garmenting today.

Merchandising keeps the tech packs. Will the garmenting agent see them?

Only the copies it is given. Files uploaded to merchandising stay in merchandising. Upload the current tech pack, PP-meeting notes and trim card to garmenting, or share them across the factory, and the agent answers from that copy and names the version it states. Which version is current stays merchandising’s decision. A scanned tech pack, or a measurement chart pasted in as a picture, has no text to read.

Bring Garmenting & Stitching. We’ll build it on the call.

Book a 30-minute demo (opens in a new tab)
Email
support@mavick.si