The hidden factory: the rework and workarounds that quietly consume your capacity

Walk into a workshop or an operations floor at five in the afternoon and you can often see it. Someone is re-checking a batch. Someone else is correcting a delivery note. A supervisor is on the phone because a figure in one system does not match the figure in another. None of this is in the plan for the day. All of it is paid for.

Quality engineers have a name for it. The hidden factory is the part of your capacity that is used up redoing work because it was not done right the first time. It is hidden because nobody budgets for it, nobody owns it, and everyone is too busy dealing with it to measure it.

Where the idea comes from, and what the numbers really are

A September 2026 article in ReliaMag by reliability practitioner Ricky Smith traces the concept to Armand Feigenbaum, who introduced it in a 1956 Harvard Business Review article. Feigenbaum's later estimate was that the hidden factory consumes between 15% and 40% of a plant's capacity in rework, repair and replacement of defective work.

The same article reports the American Society for Quality's frequently quoted rule of thumb: quality-related costs at many manufacturers run 15% to 20% of sales revenue, and as high as 40% of operations at the worst performers, against 10% to 15% for a thriving company.

It is not only a manufacturing problem

The language comes from factories, but the pattern is everywhere work is done in steps. In a services or administrative operation the hidden factory looks like this:

• Data typed into a second system because the first does not talk to it.
• Orders, invoices or reports corrected after the customer or the finance team spots an error.
• Staff chasing colleagues for approvals, information or the latest version of a file.
• Duplicate checks added after one past mistake and never removed.
• Customers contacting you again about the same issue because the first contact did not resolve it.
• Spreadsheets maintained "on the side" to compensate for what the main system cannot do.

Each item is small. Together they decide whether a team of ten behaves like a team of seven.

A practical way to find your own

1. Ask what work exists only because earlier work was wrong

This single question separates productive work from failure demand. Put it to the people who do the job and keep a running list for two weeks. Do not ask for solutions yet. Ask for examples.

2. Count and time the top five

For the five most frequent items, record how often each occurs per week and how long it takes to fix. Multiply. Even rough numbers, agreed with the team, are enough to rank the problems and to show leadership what a single correction costs over a year.

3. Find the first point where the defect could have been caught

The 5 Whys method is useful here, with one rule: stop only when you reach a cause you can change. "The customer gave the wrong address" is not a root cause if the order form accepts an incomplete address.

4. Fix at the source, then delete the check

A better form, a mandatory field, a clearer hand-off or a short checklist usually costs far less than the downstream correction. Once the source is reliable, remove the extra checks that were added to compensate. An improvement that leaves the old control in place does not release capacity.

5. Review it monthly

Make the top rework items a standing item in the operating review, with an owner, a number and a target. The hidden factory returns whenever nobody is looking at it.

Why this is often the fastest win

Reducing rework does not require new technology or new headcount. It releases capacity inside the team you already have, improves customer experience at the same time, and builds the habit of looking at how work really flows. That habit is what later makes automation and system changes succeed. For growing organisations under pressure to do more with the same people, it is frequently the highest return available.

We would add one honest note. The first measurement is often uncomfortable, because it puts a number on something people have quietly absorbed for years. Present it as a finding about the system, not about individuals, and the team will help you find more.