Skip to content

Fabric finishing: shrinkage, width and GSM against the buyer’s spec

Ask for a quality’s approved finish recipe, a lot’s shrinkage against tolerance, or a roll’s 4-point grade. The agent answers from your recipes, run sheets and test reports with the sources shown, and nothing is sent or written unless an admin approves it.

Works from: Finish recipe, Stenter run sheet, Sanforizer batch sheet, Shrinkage report

What goes wrong between the stenter and the cutting table

Finishing is the last process before cutting. A width, GSM or shrinkage miss found here costs a second pass; one the buyer finds costs a claim. Each problem below lives in a record your finishing floor already keeps.

  • A lot that passed your shrinkage test fails at the buyer’s lab

    The lot passed your own shrinkage test, then fails at the buyer’s lab or in garment washing. On stretch fabrics, shrinkage can change roll to roll. The playbook says such rejections almost always trace to a mix-up between AATCC and ISO methods. Its escalation rule: the QA head, the finishing manager and Merchandising within four hours. The finishing manager must find the retain sample, roll range and finishing batch.

    16% to 24%

    Shrinkage can vary roll by roll, sometimes to this level, on cotton–elastane and blends, a production GM said at a 2025 Textile Today meeting on denim washing in Bangladesh. Measured after washing, not residual shrinkage after sanforizing.

    Source: Textile Today, Shared responsibility across the value chain to integrate mills, laundries, and brands in minimizing quality issues (opens in a new tab), 2025.

  • Width and GSM that drift across a lot, pass by pass

    Width narrows from head to tail, or bulk GSM comes in lighter than the sample because it ran on a compactor with a worn blanket. Cutting plans its marker on that width, and every metre under the minimum width costs points at final inspection. The playbook says to check stenter temperature first, then overfeed, then chain tension, but the readings and the settings sit on separate sheets.

    4 points per metre or yard

    Penalty in Cotton Incorporated’s fabric grading guide for each linear metre or yard where the usable width is less than the minimum specified. The guide says seller and buyer must agree the grading system.

    Source: Cotton Incorporated, Visual Inspection and Grading of Fabrics (opens in a new tab), n.d.

  • A water-repellent finish that fails after the buyer’s washes

    A water-repellent or anti-microbial finish passes your test, then fails the buyer’s test, which used more washes or another method. The playbook warns against promising 50 washes when the recipe spec says 20, and sets 72 hours to reconcile recipe and method with the buyer’s QC. The finishing manager must pull the recipe as run, the pick-up and both test reports.

    5 washes

    After five washes, water repellency fell below the acceptable Grade 3 for all five outerwear fabrics in one washing variant of a 2022 lab study in Poland. The authors note that the test method can change the conclusion.

    Source: Materials (MDPI), via PubMed Central, Quality and Quantity Assessment of the Water Repellent Properties of Functional Clothing Materials after Washing (opens in a new tab), 2022.

  • Final inspection grades scattered across paper and printouts

    The playbook calls final inspection the last chance: after it, every defect is a buyer claim. Yet grades often sit on paper sheets or per-roll printouts, so the inspection in-charge cannot quickly say which rolls of a lot failed, or how many this week showed pin marks or calender glaze. Once the sheet is an Excel or CSV upload with a week column and one defect to a cell, each count is one question away.

How the finishing agent works with your recipes and reports

Every answer comes from a record you connected: the finish recipe, the stenter run sheet, the shrinkage or inspection report. A value that is not on record is named as missing, and a message it drafts goes out only if an admin approves it.

  • The approved finish recipe, beside what the stenter ran

    Answers with sources

    Name a quality or a lot, and the agent answers from your finish recipes and stenter run sheets: chemicals, add-on, padder pick-up, temperature, speed, overfeed and width. It sets the run sheet beside the approved recipe and lists the values that differ, and any that were never recorded. When the senior operator is on leave, the night shift still gets the recipe and its source.

    Ask“What is the approved finish recipe for quality CT-140, and how does last night’s run sheet for lot 3307 compare with it?”

  • Width and GSM spread, worked out from your rows

    Answers with sources

    Upload the GSM and width sheet as an Excel or CSV file, with head, middle and tail readings in their own columns. Ask for a lot’s minimum, maximum, average or spread in one column, such as tail width, for one stenter pass, machine or shift. The agent calculates each from the rows, the spread as maximum less minimum, and cites them. With the buyer’s spec uploaded too, it cites the rolls it reads outside the tolerance.

    Ask“What is the average tail width of lot 3307 on the second stenter pass, and which rolls on that pass read under the buyer’s minimum width?”

  • Shrinkage and functional tests read against the buyer’s method

    Reads your documents

    Upload your shrinkage and functional test reports, and the buyer’s reports and spec. For a lot, the agent sets warp and weft shrinkage after the first and fifth wash, or a water-repellent finish’s spray rating and wash count, beside what the buyer requires. It names the method each report used, such as AATCC 135, ISO 5077 or AATCC 22, and cites every document.

    Ask“Is lot 1874 inside the buyer’s shrinkage tolerance? Show warp and weft after wash 1 and wash 5, and the method each report used.”

  • The lot release summary, from your inspection sheet

    Creates files

    Upload the final inspection sheet as an Excel or CSV file, one row per roll. The agent counts a lot’s rolls, its accepted rolls, or its rolls with a defect type you name, from the actual rows, and cites them. It builds the release summary as a document with the shrinkage, GSM and functional test results on record, and lists every result still missing.

    Ask“Make the release summary for lot 3388, with the number of its rolls accepted at final inspection.”

  • The night shift’s stenter settings, sent only if approved

    Acts in your tools

    Name the lots for the night shift, and the agent drafts a WhatsApp Business or Slack message to the stenter in-charge. Each lot gets the temperature, speed, overfeed and width from its approved recipe, with the recipe named. An admin approves or declines the message, and it is sent only if approved.

    Ask“Draft a WhatsApp to the night-shift stenter in-charge with the approved settings for lots 3410 and 3412.”

A week on the stenter, the sanforizer and the inspection frame

  • Width narrows after the second stenter pass

    Stenter in-charge

    Readings on lot 3307 show the width narrowing toward the tail after the second pass, and cutting is waiting for the fabric.

    1. Asks

      “What is the average tail width of lot 3307 on the second stenter pass? Then compare its stenter run sheet with the approved recipe.”

    2. The agent

      Works out the lot’s average tail width on the second pass from the tail-width column of the GSM and width sheet you uploaded, and cites the rows. Reads those rows against the buyer’s spec you uploaded and cites the rolls it finds under the minimum width. Sets the temperature, speed, overfeed and width on the run sheet beside the recipe, and quotes the playbook’s order for drift: temperature first, then overfeed, then chain tension.

    3. Result

      The in-charge has the second pass’s average tail width, the narrow rolls found and the recorded settings that differ from the recipe, with every row and sheet cited.

  • A buyer’s lab fails a lot on weft shrinkage

    Finishing QC in-charge

    The buyer’s report arrives: lot 1874 failed weft shrinkage. Your own report, and your retest of the retain sample, passed.

    1. Asks

      “Put the buyer’s report on lot 1874 beside ours and the sanforizer batch sheet. Then draft an email to their QC asking for the method, water temperature and dryer temperature.”

    2. The agent

      Reads both reports and lists each one’s method, wash count, temperatures and warp and weft results, showing where they differ. Adds the roll range and sanforizer batch from your batch sheet. Builds the comparison as a spreadsheet, then prepares an email to the buyer’s QC through Gmail or Outlook asking for the method, water temperature and dryer temperature, as the playbook names them.

    3. Approval

      The finishing manager, as department admin, reads the email and approves or declines it. It is sent only if approved, and the audit trail shows who decided.

    4. Result

      A comparison sheet for the four-hour escalation the playbook sets, with both reports cited, and a request to the buyer that goes out only if approved.

  • Releasing a finished lot to cutting

    Inspection in-charge

    The last rolls of lot 3410 come off the inspection frame, and the garment unit is asking for the fabric.

    1. Asks

      “Prepare the release summary for lot 3410 from the inspection and test reports. List anything still outstanding.”

    2. The agent

      Reads the final inspection sheet you uploaded as an Excel or CSV file, the shrinkage and GSM reports and the functional test report. From the inspection sheet’s rows it counts the lot’s accepted rolls, cites them, and reads each roll’s grade and recorded defects. It builds the release summary as a document: quantity, grades, test results and open items, such as a shrinkage result not yet on record.

    3. Result

      A release summary with every figure traced to its report, and the gaps named before the lot leaves finishing.

Sizing the cost of re-runs and rejected rolls

These are published figures on stenter passes, reprocessing in one finishing department and finished-fabric inspection, cited at the source. They are not Mavick results. Baseline the scorecard below before a pilot, then read it again on your own lots.

  • Fewer unplanned passes on the stenter

    2–3 times

    How many times fabric is treated on a stenter in finishing, on average, by a European estimate (2003). These are planned passes, not re-runs. The 2010 US guide that quotes it says production bottlenecks tend to occur at stenters.

    How the agent helps

    Each unplanned re-run takes a slot on that bottleneck. The agent does not run or set the stenter. When you ask, it sets a lot’s run sheet beside the approved recipe and lists each setting that differs. Asked once the first roll is tested, it builds the playbook’s pre-run checklist from what is on record, with each missing result left open.

    Track: Lots re-run on the stenter or sanforizer, by cause and quality

    Source: Lawrence Berkeley National Laboratory, Energy-Efficiency Improvement Opportunities for the Textile Industry (opens in a new tab), 2010.

  • Less fabric reprocessed or rejected in finishing

    0.5 million metres a month

    Fabric reprocessed on average each month at one Pakistani home-textile unit producing about 14 million metres a month. 55% went back for re-equalising on the stenter and 45% for re-finishing, a 2025 study of its finishing department found.

    How the agent helps

    Upload your reprocessing and rejection log as an Excel or CSV file, with a column for the month. The agent sums the metres reprocessed or rejected for the month, quality and recorded cause you name, such as shrinkage, width, GSM or hand-feel, from the actual rows. Each figure names the log rows it came from.

    Track: Metres reprocessed or rejected, by cause and quality

    Source: Journal of Design and Textiles (University of Management and Technology, Lahore), Improvement of Sigma level in a Finishing Department of Home Textile Unit Industry (opens in a new tab), 2025.

  • More rolls through final inspection the first time

    30%, then 50%

    Share of a finished batch checked on the inspection machine after the stenter or compactor at some Bangladeshi knit mills, then the share checked when the lot misses its quality requirements, per a 2024 peer-reviewed study.

    How the agent helps

    From the uploaded final inspection sheet, with a column for the week, the agent counts the rolls accepted in the week you name, or the rolls showing one defect type, citing the rows. The playbook ties pin marks at the edges to pin chain alignment, and glaze seen only on saturated colours to calender pressure, so a defect count points to a machine to check.

    Track: First-pass final inspection %, by quality

    Source: Heliyon (Elsevier/Cell Press), via PubMed Central, An approach to automatic fault detection in four-point system for knitted fabric with our benchmark dataset Isl-Knit (opens in a new tab), 2024.

What to baseline before a pilot

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

Finishing: the KPIs to baseline before a pilot, and why each matters
KPI to baselineWhy it matters
Rolls within ±1.5% of the buyer’s width, % of rollsA narrow roll changes the marker at cutting and costs points at final inspection. The playbook checks width on every roll.
Rolls within ±5% of target GSM, % of rollsThe playbook’s band, tighter for some buyers, with GSM read at three points across the width. Drift across a roll points to stenter temperature, overfeed or chain tension.
Tested batches inside the buyer’s shrinkage spec, %Warp and weft after one wash and after five, by the buyer’s method. By the playbook, a rejection at the buyer goes to the QA head, the finishing manager and Merchandising within four hours.
First-pass final inspection %The last check before the lot leaves finishing. The playbook targets 92% or more and reads it daily beside a defect Pareto.
Hand-feel acceptance by the buyer, per shipmentThe playbook tracks it on every shipment, and a buyer complaint about hand-feel sends the finish recipe back for review.

What the agent works from

Records it reads

  • Finish recipe
  • Stenter run sheet
  • Sanforizer batch sheet
  • Shrinkage report
  • GSM and width report
  • Functional finish test report
  • Buyer’s test report
  • Final inspection report

In your ERP

Reads Odoo production orders and stock transfers, each with the time fetched. Drafts a note on a held lot’s transfer, or a production order’s planned start or assignee; an admin approves or declines, and Odoo changes only if approved. Recipes, test results and grades come from your files and sheets. The demo shows whether your Odoo holds finishing’s orders.

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 a buyer’s QC email in Gmail or Outlook, or a GSM log in Google Sheets, when you ask. Prepares an email to the buyer’s QC through Gmail or Outlook, a WhatsApp Business or Slack message with a lot’s approved settings, or a hold-register row in Google Sheets. An admin approves or declines each; nothing is sent or added unless approved.

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

Never made up

  • Finish recipes, chemical add-ons or pick-up percentages
  • Stenter or sanforizer settings: temperature, speed, overfeed or width
  • Shrinkage, width, GSM or hand-feel test values
  • Inspection grades or defect points for any roll

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

Belongs to another department

What the Finishing playbook covers

  • SOP: setting the stenter

    Width is set to the buyer’s spec within ±1.5%, and overfeed controls shrinkage in the length. Cotton finishes typically cure at 140–180°C; synthetics and elastane blends heat-set at 190–210°C, and elastane is never heat-set below about 190°C. Too hot for the recipe shifts the shade and makes the fabric feel harsh. GSM is checked at three points across the width on every roll.

  • Decision rule: a shrinkage rejection at the buyer’s QC

    Retest the retain sample by the method the buyer used, since AATCC 135 and ISO 5077 give different results. If the retain passes and the shipment failed, ask for the buyer’s test report with the method, water temperature and dryer temperature. If the retain fails too, find the roll range and the finishing batch.

  • KPI: first-pass final inspection

    The playbook targets 92% or more, read daily beside a defect Pareto. Final inspection uses the same 4-point system as greige inspection. The defects to look for at this stage are stenter pin marks, uneven width, calender glaze, oil stain and stain carried over from the previous lot.

  • Heuristic: sign off hand-feel on sanforized fabric

    The playbook says finishing is where shade and hand-feel claims are decided. So get hand-feel signed off on a sanforized swatch, not on greige, and match it to the approved buyer sample on the pre-run checklist before every bulk run.

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.

  • Approved finish recipe for a quality

    What is the approved finish recipe for this quality, and what is currently set on the stenter or sanforizer?

  • Shrinkage against buyer tolerance

    What shrinkage was recorded for this lot in warp and weft, and is it within the buyer’s tolerance? Cite the test report.

  • Width and GSM variation in a lot

    Show the width and GSM readings across this lot and tell me where the variation appeared and at which machine pass.

  • Final inspection grade for a roll

    What did the final 4-point inspection grade this roll, and what defects were recorded against it?

  • Functional finish test results

    What are the test results for the functional finish on this quality — water repellency, anti-pill, or peach — against its required standard?

  • Stenter settings actually run

    What temperature, speed, overfeed and width were actually run on the stenter for this lot, and how do they compare with the recipe?

  • Lot release summary

    Prepare a release summary for this finished lot from the inspection and test reports: quantity, grade, test results, and anything outstanding.

  • Shrinkage rejections this month

    How much fabric was rejected or reprocessed for shrinkage this month, and which qualities does it concentrate in?

  • Fabric pending finishing

    What fabric is waiting for finishing, how long has it been waiting, and which orders does it belong to?

  • Chemical add-on against recipe

    Show the chemical add-on and pick-up percentage recorded for this lot against the approved recipe.

Finishing: questions before a demo

Our GSM cutter readings go into a hand-written register. What can the agent use?

Only what is typed or exported. A photo of the register, or an image-only PDF of it, has no text to read. Once the readings are in a Google Sheet you connect, the agent reads and cites them. Upload them as an Excel or CSV file, with the readings on the first sheet, and it also works out a lot’s minimum, maximum or average from one reading column, such as middle width, for one pass or shift if you name it. Recipes, shrinkage reports and inspection reports that are already PDF, Word or Excel files can be uploaded as they are, up to 10 MB each.

Can it read the settings straight from our stenter or an online GSM gauge?

No. Mavick does not connect to stenters, sanforizers, compactors or gauges. If a machine’s run data is exported to a file you upload, or typed into a Google Sheet you connect, the agent reads it there and compares temperature, speed, overfeed and width with the approved recipe. It answers only when someone asks, and if a reading is not in your connected data, it says so.

Can the agent note a held lot in our ERP?

Only in Odoo. There it can draft a note on the lot’s stock transfer, such as “3 rolls held for re-inspection”; the note informs, and it does not block or move the stock. It can also draft a new planned start on a re-finishing production order already in Odoo. A department admin approves or declines each draft, and Odoo changes only if it is approved. It writes no recipes, test results or inspection grades to any ERP: those stay in your reports and sheets. With another ERP, it can add the hold as a row to your hold register in Google Sheets instead, with the same approval.

Will it tell us if our shrinkage test and the buyer’s used different methods?

Yes, when both reports record it. Upload your shrinkage report and the buyer’s test report, and the agent lists each one’s method, such as AATCC 135 or ISO 5077, with the wash count, the temperatures and the warp and weft results. If a report does not name its method, it says so. It does not test fabric or decide which lab is right. The playbook’s next step is a retest of the retain sample by the buyer’s method. Questions about the test methods themselves belong to Quality (QA/QC).

Can Finishing read the dye house’s recipes or the buyer specs Merchandising holds?

No. The finishing agent sees only Finishing’s own files and tools, files shared across the factory, and the ERP records its work concerns. The separation is built into how Mavick stores and retrieves everything. Asked about a dyeing recipe or a shade match, it says Dyeing & Processing owns that; asked about buyer tolerance standards and approvals, it names Merchandising. To check lots against a buyer’s shrinkage or width tolerance, upload that spec to Finishing, or share it across the factory.

Will it tell us how to change the recipe when a lot fails?

It does not write new recipes or settings, and it never makes up an add-on, a temperature or an overfeed. It reads the approved recipe and the run sheet, lists what differs, and quotes the playbook’s rule for the case: for width or GSM drift, check temperature first, then overfeed, then chain tension; for pin marks only at the edges, check the pin chain alignment. The change itself stays with your finishing manager.