Made-to-order production vs. manual control in footwear

An order for 2,400 pairs arrives from a German client in February. Sizes 36 to 46, four colours, two lasts. Production starts on Monday. On Friday, the production manager does not know how many pairs are completed per reference — because the control is done on paper and Excel. The question that should be answered in 30 seconds takes a day to answer, and by then it is too late to recover.

The thesis of this article is not "digitalise everything". It is more uncomfortable: manual control fails not because people are careless — it fails because the data structure of footwear is inherently matrix-based, and any linear tool, whether paper or Excel, collapses under that complexity. A collection of 120 models with four colours and eight sizes easily generates 3,800 active references simultaneously. There is no spreadsheet designed for that.

The market context aggravates the equation. According to APICCAPS, the Portuguese footwear industry exported €1,718 million in 2025, with growth of 1.8% in volume — but the US fell 12.3% and the pressure on European margins is not easing. In an industry where market diversification is the response to risk, getting an order wrong through production invisibility has ceased to be a minor operational problem. It is a direct commercial risk.

The real problem is not Excel — it is invisibility

Most footwear factories in Felgueiras and Guimarães did not lose control of production for lack of effort. They lost it through an excess of unmapped complexity. Colour × size × last × finish: each axis multiplies the number of references. When you try to fit that matrix into a flat grid, the information does not disappear — it becomes hidden. And what is hidden is not managed.

There is a pattern we see repeated: the factory has Excel updated to the end of the previous day, the production manager knows the total pairs packed, and everyone believes the order is under control. The problem appears on Thursday, when it is discovered that the 400 pairs of size 38 in camel are still in finishing with a quality problem in the sole — and dispatch is Monday. The total was right. The distribution was wrong. And that distinction is invisible in an Excel of totals.

What you need before launching production

Before issuing any production order, seven elements are needed. Not as bureaucracy — as a starting condition. The absence of any one of them turns order management into crisis management.

First, the complete order matrix: not totals per model, but quantity per each combination colour × size × last, confirmed with the client. Second, the operations map per section — cutting, stitching, lasting, finishing, packaging — with average times per operation and per size. Times vary by size more than most planners admit: a size 46 in lasting can take 15% longer than a 39, and that deviation accumulates over an order of 2,400 pairs.

Third, the real capacity per section: shifts, operators, workstations — with historical absenteeism included, not the theoretical capacity on paper. Fourth, the dispatch date confirmed with documented tolerance. Fifth, the suppliers of critical components identified with updated lead times. Sixth, the definition of who validates quality at each control point — not just at the end. Seventh, historical production data: pairs per shift, per operator, per similar model.

This last element is the most ignored and the most valuable. Without history, planning is optimistic fiction.

Step 1 — Structure the order before launching it

Do not launch production with a total number of pairs. Launch with the complete matrix broken down. For each reference — model + colour + size — define the quantity to be produced, the date of entry into cutting, the expected date of exit from lasting, and the necessary components with confirmed available stock.

This breakdown seems bureaucratic. It is not. It is the only moment at which you can detect that size 46 in a specific model uses a different sole that has not yet arrived — before the line stops halfway through the week. In the projects we support in the footwear industry, this step is systematically ignored when delivery pressure is high. It is also the step whose absence generates the most rework downstream.

Validation checklist before launch:

• Colour × size × last matrix complete and confirmed with the client.
• Component stock verified reference by reference — not by product family.
• Orders to suppliers issued for what is missing, with confirmed delivery date.
• Production sequence defined: which model enters cutting first and why.

Step 2 — Define control points at each section transition

The most common error: the factory counts pairs at the end of the day — total pairs packed. It does not count per reference, does not count per section, does not record where work in progress is. The result is a number that seems reassuring and hides the real problem.

The production manager who only counts pairs packed is managing the past. Whoever counts pairs in progress per section is managing the future.

Define a control point at each transition: exit from cutting, entry into stitching, exit from stitching and entry into lasting, exit from lasting and entry into finishing, final packaging. At each point: quantity per reference, quality confirmation, comparison with the plan. Each point generates a record. With manual control, it is a signed sheet. With KORA Productivity, it is a scan on an industrial terminal that updates the order status in real time — without depending on anyone remembering to fill it in.

An operational detail that the manuals do not mention: the most critical control point is not packaging — it is the exit from cutting. That is where quantity deviations per reference appear first, still with time to recover. Factories that only control at packaging discover the problems when there is no longer any room to manoeuvre.

Validation checklist:

• Control points defined and communicated to each section supervisor.
• Person responsible for recording identified at each point — one person, not "everyone".
• Recording frequency defined per shift, not per day.
• Deviations reported to the production manager within the very shift in which they occur.

Step 3 — Calculate the real progress, not the declared progress

"We are at 60% of the order" is declared progress. Real progress is: "we have 60% of the pairs packed, 15% are still in finishing with a quality problem in the sole, and 8% are stuck in cutting because the leather for size 44 has not arrived." The difference between the two is the difference between an on-time dispatch and a difficult call to the client.

Calculate progress per reference and per section, daily. The formula is simple:

Progress (%) = Pairs completed in section X / Pairs planned for section X × 100

Apply this for each section, for each reference. The result is a progress matrix that shows exactly where the bottlenecks are — before they become unrecoverable delays. Deviations above 10% in a critical reference require a recovery plan the same day, not at the weekly meeting.

Validation checklist:

• Progress calculated per reference — not only per total order.
• Progress calculated per section — not only at the end of the line.
• Deviations above 10% with a documented recovery plan.
• Daily 15-minute meeting with section supervisors to review the matrix — no more, no less.

Step 4 — Treat traceability as production data, not as bureaucracy

Batch traceability in footwear has ceased to be optional. The European ecodesign product regulation (ESPR), in force since July 2025, increases the requirements on materials and components. Brands such as Inditex, Decathlon or Mango already contractually require their Portuguese suppliers to be able to answer the question: "this batch of soles, which supplier did it come from and in which pairs was it used?" With manual control, this answer takes hours — or does not exist. With an integrated system, it is a 30-second query.

Record at each control point the batch of raw material used (leather, sole, lining, laces), the supplier and date of receipt, the operator responsible for the operation, and the date and shift of the record. This record serves two purposes simultaneously: it answers the regulatory requirement and identifies the origin of a quality problem in minutes, not in days.

There is a systematic error we see at this stage: factories record the batch on entry into the warehouse but do not make the batch → product reference association at the moment of cutting. When the traceability request arrives, they know the leather arrived — but they do not know in which pairs it was used. The association has to be made at cutting, not afterwards.

Validation checklist:

• Raw material batches recorded on entry into the warehouse.
• Batch → product reference association made at cutting — not retrospectively.
• Operator record at each critical operation.
• Ability to respond to a traceability request in under one hour.

Step 5 — Close the order with data, not with intuition

When the order goes out for dispatch, the formal closing is not optional. Quantity dispatched per reference versus quantity ordered. Deviations documented with identified cause — missing pairs, substitutions, quality rejections. Real production time versus planned time, per section. Real production cost versus budgeted cost.

Without this closing, the next order from the same client starts with the same planning errors. The closing data feed the history that will improve the estimate of the next similar order. With the MULTI ERP, this history is automatically available for the next estimate — without depending on the memory of the production manager, who may have changed in the meantime.

Closing done on the very day of dispatch has a value that closing done a week later does not: the causes of the deviations are still in the memory of those who lived through them. A week later, what remains is the official version — which is rarely the complete version.

Validation checklist:

• Order closing done on the very day of dispatch.
• Deviations documented with identified cause — not only with quantity.
• Real time data per section recorded.
• Real cost calculated and compared with the budget.

Manual control vs. production software: what actually changes

The errors that cost the most — and how to avoid them

Launching production without confirming the stock of critical components. The line stops halfway because the soles for size 45 have not arrived. The confirmation has to be made component by component, reference by reference, before issuing the production order — not by product family, not by estimate.

Counting pairs by total order, not per reference. The total may seem reassuring while there are critical references running late. Always break it down by colour × size × last. A good total with a reference in collapse is an incomplete dispatch.

Recording progress once a day, at the end of the shift. A deviation that appears at 10am and is only recorded at 6pm has lost eight hours of recovery margin. Recording per shift — at least twice a day — is the difference between managing and reacting.

Treating traceability as an audit and not as a process. When the brand requests the traceability of a batch, there is no time to reconstruct the history. Traceability has to be recorded at the moment of the operation, not reconstituted afterwards. Factories that try to do traceability retrospectively produce documentation that does not withstand a serious audit.

Not formally closing the order. The factory that does not close the order with real data is financing the errors of the next order without knowing it. The next estimate will be made with the same wrong premises — and the margin will continue to be lower than planned.

What the matrix structure of footwear demands of a management system

A generalist ERP treats an order as a line with quantity and deadline. In footwear, an order is a three-dimensional matrix: model × colour × size, with variations of components, operation times and suppliers per axis. A system that does not model this structure natively forces the factory to keep Excel in parallel — and it is that parallel Excel that fails.

Production management by order in footwear is not a problem of operational discipline. It is a problem of the fit of the tool to the real structure of the data. When the tool is right, discipline follows naturally — because recording is easier than not recording, and visibility replaces the anxiety of not knowing.

The question that the production manager should be able to answer in 30 seconds — "how many pairs of size 38 in camel have already passed lasting?" — is not a difficult question. It is a question that requires the right data to be in the right place, at the right time. That is the work that the five steps above structure. What happens when they are not, everyone who works in this industry already knows by heart.

Frequently asked questions

What is the difference between manual and per-order production control in footwear?

Manual control counts totals of packed pairs, hiding distribution problems per reference. Per-order control breaks down the colour × size × last matrix, recording quantity in each section. It allows you to detect that 400 pairs of size 38 in camel are late in finishing — while the overall total seems correct.

Why does Excel not work for footwear management?

A collection of 120 models with four colours and eight sizes generates 3,800 simultaneous references. Excel is linear; footwear is matrix-based. When you try to fit that matrix into a flat grid, the information becomes hidden and ceases to be managed. What is invisible in Excel is not controlled.

How many elements are needed before launching an order?

Seven elements are essential: complete colour × size × last matrix; operations map with times per section and size; real capacity per section with absenteeism; confirmed dispatch date; supplier lead times; quality validators at each point; historical data from similar production.

How does size affect production time in footwear?

A size 46 in lasting can take 15% longer than a size 39. This deviation accumulates over an order of 2,400 pairs. Most planners ignore this variation, creating unrealistic production plans. Without historical data per size, planning is optimistic fiction.

What is the most critical control point in footwear production?

The exit from cutting is the most critical point, not the final packaging. That is where the first quantity deviations per reference appear, still with time to recover. Factories that only control at packaging discover problems when there is no longer any room to manoeuvre to correct them.

What does "production invisibility" mean in the context of orders?

It is when the total number of pairs seems correct, but the distribution per reference is wrong. The production manager does not know that 400 pairs of size 38 in camel are held up in finishing with a problem in the sole. The information exists, but it is hidden in the total figures.

How is the Portuguese footwear industry affected by production errors?

Portugal exported €1,718 million in footwear in 2025, with pressured margins. The US fell 12.3%. In an industry where market diversification is the response to risk, getting an order wrong through production invisibility has ceased to be a minor operational problem — it is a direct commercial risk.

Sources

• APICCAPS (Portuguese Association of the Footwear, Components and Clothing Industries) — Export Statistics 2025 and market analysis (data cited: €1,718 million, growth of 1.8% in volume, US contraction of 12.3%)
• INE (National Statistics Institute) — Classification of Economic Activities (CAE) and industrial production statistics in Portugal, relevant to the footwear industry context
• ISO/IEC 27001:2022 standard — Information Security Management Systems, applicable to production control and traceability systems
• ISO 9001:2015 standard — Quality Management Systems, relevant to the definition of control points and validation in production processes