It is six in the evening and the sewing supervisor on line three photographs a handwritten tally sheet and posts it in the production group. Forty other messages sit above it: a voice note about a shortage of zips, a photo of a broken needle guard, someone asking whether order 1142 has left cutting yet. The next morning the planner opens the group, scrolls, and retypes the numbers into the production program on the office PC — a Windows application installed years ago by a vendor nobody can reach now. By the time the numbers are in, the floor has moved on a full shift.
Nothing in that picture looks like a crisis, and that is exactly the problem. A factory running this way does not lose production time in one visible failure. It loses it in dozens of small delays, every day, that nobody records because the tools in use cannot record them.
If you run or manage a garment factory and suspect your lines are losing hours you cannot see, what you want is to find those hours, put a number on them, and get them back without stopping production to do it. This article is built around that: where the time goes, how to measure it in your own factory in a week, what line tracking has to do if it is going to replace the group rather than sit beside it, and how to change over one line at a time.
Why WhatsApp won in the first place
It is worth being fair to the group before criticising it, because any replacement that ignores why it won will fail the same way the last attempt did.
WhatsApp is on every supervisor's phone already. It costs nothing, needs no training, works in Arabic, and takes a photo, a voice note or a number in seconds. When a bundle is short or a machine stops, telling everyone takes one tap. No factory system of the last twenty years has matched that for speed of reporting, and most of them asked supervisors to walk to a terminal and type.
The old Windows system usually won its place for similar reasons: it was bought once, it runs without an internet connection, and the one person who knows it can produce the reports the owner asks for. Both tools are rational choices. The cost is in what happens between them.
Where the time actually goes
Lean manufacturing has names for most of what follows: waiting, correction, and processing that adds nothing for the customer. Naming them helps, because each has a known remedy. What a chat and an old program add is that none of them can be seen.
A problem is seen at once and acted on late
A garment line moves at the pace of its slowest operation. When the bottleneck operation stops — a machine down, a missing trim, an operator absent — every operation after it runs out of work, and the line's output for that period is gone unless there is enough work-in-progress buffered to cover it.
In a group chat, the problem is usually visible within minutes: someone posts it. But visible is not the same as received. The message sits between a greeting and a photo of a sample. The maintenance technician is not in that group, or is in it with notifications muted. The planner reads it at the end of the hour. The gap between the moment a problem is posted and the moment someone who can fix it acts on it is where most of the lost time lives, and in a chat that gap is invisible, because nothing records when the right person saw the message.
Everything is typed twice
Numbers that start on paper, travel through WhatsApp and end in the Windows program are entered at least twice, often three times. Each entry costs someone's hours, every day, and each one is a chance to swap a digit or a size. Worse, the re-entry always runs behind the floor. The system in the office describes yesterday, so nobody makes today's decisions from it, and the owner stops trusting its reports.
Nobody can answer "where is order 1142?"
When a buyer or a merchandiser asks for the status of an order, the answer is assembled by hand: scroll the group, call the cutting supervisor, walk to finishing, count. That can take an hour for one order, and it is asked several times a day. The same question from the owner gets the same treatment. None of it produces a single garment.
Handovers between stages leave no trace
Cutting hands bundles to sewing, sewing to finishing, finishing to packing. If each handover is a message in a group, or nothing at all, then a shortage discovered at packing cannot be located. Was it short at cutting, lost in sewing, or rejected at finishing and never replaced? The only way to find out is to recount the order stage by stage, and the argument about whose count is right takes longer than the count.
Quality problems repeat for the rest of the shift
A defect reported as a photo in a chat is not tied to the order, the line, the operation or the operator. By the time the quality supervisor connects the photo to its cause, the same operation has produced the same defect on every piece since. Rework found at final inspection costs far more than a defect stopped at the operation that made it.
Planning is done from memory
Without a live view of work-in-progress per line, the planner loads lines from what they remember and what the supervisors said this morning. One line waits for cut panels while another is buried. An order is promised a ship date based on a line that is already behind. The planner is not careless; they are working without the one thing planning needs.
The Windows program is a single point of failure
The old system usually lives on one PC. Only one person can use it at a time, and only in the office. It cannot be reached from the floor or from home. Backups, if they exist, are on the same disk. The vendor has moved on, the version cannot be updated, and nobody dares touch it. To share anything, someone exports it to Excel — which creates another copy that will disagree with the first by next week.
A worked example of the arithmetic
The figures below are an illustration to show how the losses add up, not a measurement from any particular factory. Replace them with your own.
Take a line with a target of 600 pieces in a ten-hour shift — sixty an hour. Its bottleneck operation stops at ten in the morning because a machine fails. The supervisor posts it in the group at once. The technician sees it forty minutes later and the machine is running again at eleven thirty. With little buffered work after the bottleneck, the line loses close to ninety minutes of output: about ninety pieces. If something like that happens twice a week on each of four lines, that is roughly 720 pieces a week that were planned and never made — before counting re-entry, status-chasing, recounts and rework.
The point is not the number. It is that the ninety minutes was mostly waiting, and waiting is exactly what a better channel can remove.
Measure it before you change it
You do not need a new system to find out how much time you are losing. You need one week and a notebook, or a simple shared sheet kept by one careful person.
- Problem to response. For every stoppage, write down when it was first reported and when the person who could fix it started work on it.
- Re-entry hours. Ask everyone who types production numbers into any system how long it takes them each day.
- Status questions. Count how many times a day someone asks where an order is, and how long each answer takes.
- Count disagreements. Record every time two stages disagree about a quantity, and how long it took to settle.
- Late defects. Note every defect found at final inspection that came from an earlier operation.
A week of that gives you a figure in hours and pieces that belongs to your factory. It is the only honest basis for deciding whether to change anything, and it is the figure any software company should be asking you for before it quotes.
What line tracking has to do instead
If a system is going to replace the group, it has to beat the group at what the group is good at. That rules out most of what factories have been sold before.
- Recording happens at the point of work, in seconds. The supervisor or operator records a completed quantity or a problem on a phone or tablet with a few taps, in Arabic, without typing paragraphs. If it takes longer than sending a photo, it will not be used.
- It keeps working when the network does not. Factory Wi-Fi drops. Entries made while it is down have to queue and send when it returns, not disappear.
- Handovers are recorded by both sides. The stage that hands over states the quantity; the stage that receives confirms it. A shortage then has a location and a time instead of an argument.
- Problems are raised from the work itself. A stoppage is reported against the order, line and operation it belongs to, goes to the person who can act on it, and escalates if nobody does.
- Everyone sees the same live picture. Work-in-progress, completed and remaining quantities per order and per line, and what is stuck, visible to the planner, the owner and the merchandiser without a phone call.
- Cost follows the work. When each stage is tied to its cost and to the order it belongs to, an overrun shows up while the order is still on the line, not at the end of the month.
- Nothing is typed twice. The floor's entries are the record. Anything accounting or payroll needs is exported or connected, not re-keyed.
- It runs in a browser. Not installed on one PC; reachable from the floor, the office and outside, on hardware the factory already owns.
None of that requires sensors on every machine. Machine data can be connected later where it is worth it, but a factory gets most of the benefit from people recording the right things at the right moment.
Changing over without stopping the floor
The factories that make this work do not switch everything on one Saturday.
- Start with one line and one kind of order. Pick the line with the most visible problems, so the difference is easy to see.
- Start with handovers and stoppages. They produce the most lost time and are the easiest to record. Piece-rate, detailed operation times and machine data can follow.
- Train the supervisors first. If they find it faster than the group, the operators will follow. If they do not, nothing else matters.
- Keep WhatsApp for conversation, not for records. People will still talk in the group. The rule is only that a number or a problem is not real until it is in the system.
- Load open orders from the spreadsheets you already have, rather than asking anyone to type them in again.
- Decide what the old Windows program is for. Take its data out while someone still knows how. If it still does a job well — accounting, for example — it can stay and be connected. If its only job was production tracking, retire it once the new system has run in parallel long enough to trust.
When you do not need any of this
A single-line workshop where the owner stands on the floor all day can run perfectly well on a group and a notebook. The owner is the live view.
The signs that you have outgrown that are specific: more than one line or more than one shift; several orders on the floor at once; buyers asking for status updates you cannot answer quickly; recurring arguments about counts between stages; and an owner who can no longer see the floor from where they sit. If two or three of those are true, the week of measurement above will almost certainly show the loss is worth recovering.
If you are weighing a ready-made package against a system built around your floor, how to decide between Odoo and a custom system sets out the trade-off, and what a custom system costs explains how the price of the second option is built.
Going further
Three calculations sit behind the losses described here, and each has its own article: how garment factories calculate production efficiency, finding the bottleneck on a sewing line, and tracking fabric waste in the cutting room. If you are not sure which of these problems is costing you most, factory production tracking problems: symptoms, causes and where to start is a one-week diagnosis.
What we build for production floors
Softwiro builds and operates production tracking for factories. Production Manager follows each product through an unlimited set of stages with a handover-and-receiving step between them and a worker account for each person on the line. ProTrack ties the stages of a job to its cost and its contract, ships with a garment factory template, and raises an alert when costs pass their expected level while the job is still running. MESwiro gives operators a tablet screen that works offline and lets them report a problem from the step itself. If your lines run on a group chat and an old program, tell us how your floor works and we will tell you where we think the time is going.
Sources
- Lean Manufacturing Techniques for the Ready-Made Garments Industry — International Labour Organization, Cairo office. Defines the bottleneck and line balancing for garment lines.
- What are the 7 wastes in lean? — Lean Enterprise Institute. Waiting, correction and other wastes on a production floor.
- Theory of Constraints — Theory of Constraints Institute. Why every system's output is limited by its constraint.
- ILO promotes industrial engineering as a catalyst for productivity improvements in factories — International Labour Organization. Line balancing and productivity on cutting and sewing lines.