Real-time Barcode & RFID Bundle Tracking on Sewing Floors

Every sewing floor loses bundles, and the loss is rarely theft. A bundle sits behind a machine, or under a pile at finishing, or in the wrong size trolley, and nobody notices because the only record of its position is a paper ticket that travels inside the bundle itself. The production manager asks where order quantity stands, the supervisors count what they can see, and the answer is a guess rounded to the nearest hundred pieces. That guess is the reason shipments go short, and the reason the shortfall is discovered at packing instead of at the station where the bundle stopped moving.

Paper tickets fail in a specific way: they report the past. Tickets collected at the end of a shift tell the manager what happened yesterday. A bundle that stopped on Tuesday morning appears in Wednesday report, and by then the line has already lost a day of output it cannot recover. Real-time tracking changes the tense of the information. A scan at handover says where the bundle is now, and a bundle that has not scanned for a shift is a visible problem during the shift, while help can still change the outcome.

The workflow starts at cutting, not at sewing. When fabric is cut, each bundle receives one ticket carrying order number, style, size, colour, quantity, and a unique barcode. That ticket travels with the bundle through every department, and every handover becomes a scan. The bundle is scanned out of cutting, into sewing, out of sewing, into finishing, and into packing. Five scans per bundle is enough for most floors, because five points describe the whole journey: what left cutting, what entered sewing, what left sewing, what entered finishing, and what reached packing.

Scan ownership matters more than scanner hardware. Each scan point needs a named role, on every shift, for whom scanning is part of the handover rather than an extra task afterwards. When scanning is an extra task, busy operators skip it exactly when the floor is busiest, which is exactly when visibility matters most. When scanning is the handover, the bundle cannot change hands without a record, and the record stays honest even under pressure.

RFID tags cost more than printed barcodes, and the extra cost is only justified where reading without handling saves real time or solves a real problem. A bundle passing through a reader gate at a wash plant, where bundles move on conveyors and stopping to scan each one would choke the flow, is a genuine RFID case. A high-value garment where each piece carries a tag through finishing and packing, and where the tag also serves as an anti-counterfeit record, is another. Reading many bundles at once, for example a full trolley passing a gate, is a third.

For ordinary handover points, where an operator already holds the bundle in hand, a barcode scan takes about four seconds and costs almost nothing. Replacing that scan with RFID adds tag cost, reader cost, and a new failure mode in tag loss, without adding information. The honest comparison is not barcode versus RFID in general. It is the cost and reliability of one read at one point, against the value of the decision that read enables. Where the decision needs a bundle counted without stopping, RFID wins. Everywhere else, barcode wins.

Managers often resist tracking because they imagine an army of scanners. The arithmetic usually surprises them. Take an order of 10,000 shirts with a bundle size of 25 shirts. Dividing 10,000 by 25 gives 400 bundles. With five scan points, 400 bundles multiplied by 5 gives 2,000 scans for the whole order. At 4 seconds per scan, 2,000 scans multiplied by 4 seconds gives 8,000 seconds of scanning labour. Dividing 8,000 seconds by 3,600 seconds per hour gives about 2 hours and 13 minutes for the entire order, spread across every operator who handles bundles over the full production run.

That figure is the whole cost of visibility for the order: a little over two hours of scanning, against days of production. It also sets the design constraint. If a proposed design needs twenty scan points, the same order generates 8,000 scans and nearly nine hours of scanning labour, and operators will start skipping scans. Five points is not a compromise; it is the design that survives contact with a busy floor.

Scan data becomes management information through one report: bundle ageing. The report lists every open bundle by current location and hours since its last scan. A bundle that scanned into sewing four hours ago is normal work in progress. A bundle that scanned into sewing two shifts ago is stuck, misrouted, or short, and each possibility needs a different action. A stuck bundle needs bottleneck help. A misrouted bundle needs finding. A short bundle needs the cutting balance checked before packing promises a quantity that does not exist.

Daily reconciliation closes the loop. Scanned bundles are reconciled against cut quantity every day, and every missing bundle is investigated before it becomes a short shipment. The reconciliation habit matters more than the dashboard, because a dashboard nobody reconciles becomes decoration within a month. The factory that reconciles daily learns its true work-in-progress position, and the factory that learns its true position stops promising shipment quantities it cannot meet.

Fix bundle size per style and print tickets at cutting with a unique barcode per bundle. Choose five scan points plus one at the bottleneck, and name the scan owner at each point for every shift. Build the ageing report before buying hardware, because the report defines what the hardware must deliver. Run a pilot on one line for two weeks, measuring scan compliance, ageing accuracy, and the time from a bundle stopping to a supervisor noticing. Only then extend line by line, moving scan points when bottlenecks move, and retiring scan points that produce no decision. Tracking that nobody acts on is worse than no tracking, because it creates the illusion of control while the floor keeps losing bundles the old way.

A factory curious about RFID should pilot it on one line for four weeks before committing to a floor-wide rollout, because RFID performance depends on the physical environment in ways no brochure captures. Week one establishes the baseline with barcodes: scan compliance, bundle ageing accuracy, and the time from a bundle stopping to a supervisor noticing. Week two installs tags and readers on the pilot line only, with the reader gate placed where bundles already pass without stopping, such as the entry to the wash plant or the finishing department door.

Week three runs both systems in parallel, and the comparison must be honest. Measure read rate at the gate, meaning the share of passing bundles the reader captures without a manual rescan, and measure tag loss, meaning tags that arrive damaged, detached, or unreadable. A read rate below about 98 percent in the pilot is a warning, because every missed read recreates the blind spot the system was bought to remove. Week four decides on numbers, not enthusiasm: if the gate reads reliably, if tag loss stays low, and if the pilot line shows faster bottleneck detection than the barcode line beside it, RFID earns extension to the next use case. If any of those three tests fails, the factory keeps barcodes and revisits RFID when the use case changes, because a technology that works in the pilot is an investment, and a technology that merely impresses in the pilot is a cost.