In brief
For fourteen years, I did quality in small manufacturing companies. I worked in several shops and visited many more. And there's one thing I never saw, not once: a real quality management system. Not software built for the job. Everywhere, the same thing — a pile of Excel files, network folders and a forgotten module inside the ERP, that nobody designed and everybody inherited.
This article isn't about selling or about software to buy. It asks a question I started turning over the day I moved to the other side of the table: why does the quality of an entire shop rest on a file nobody chose, handed down by someone who left?
I never saw one
Let me be precise, because this is the kind of claim people take for exaggeration. Over fourteen years, in the shops where I was the quality manager and in the ones I passed through — cross-audits, troubleshooting, supplier visits — I never once met a small manufacturer running its quality on a system built for it. Never.
It's not that these shops had no quality. They had it, often good, carried by serious people. What they didn't have was a system — something someone had thought through, start to finish, to tie non-conformities to processes, measurements to instruments, actions to owners. Instead, there was always the same thing. An assembly. The homemade system.
The homemade system
You know it, because you probably have one. It's the Excel workbook with fifteen tabs, three of which nobody remembers the purpose of. It's the network folder named "Quality", with its subfolders "2021", "2021 v2", "2021 FINAL". It's the ERP's quality module, bought one day in a bundle, that does ten percent of what's needed and everyone abandoned after three weeks. Sometimes it's simply a stack of cardboard folders and a notebook.
The homemade system has one defining trait: nobody designed it. It grew. A column added one Tuesday because a customer asked. A tab created for an audit and never removed. A macro written by someone who could code, that no one dares touch since they left. It's an accumulation, not an architecture. And it holds — it really does — as long as nobody asks a hard question.
Inherited, never chosen
Here's the point I find most telling, and the one least discussed. The homemade system isn't chosen: it's handed down.
You arrive as the new quality manager. You're given the keys — that is, a file, a network folder, and half a day of explanations from the person leaving, if you're lucky. You inherit their logic, their shortcuts, their boxes whose purpose escapes you. You're not going to rebuild it all: you don't have the time, and anyway "it works". So you add your layer on top of theirs, which was on top of the one before. Strata. Archaeology in a spreadsheet.
The real problem isn't the clumsiness of the assembly. It's that the person who truly understood the system is gone. The knowledge that holds it together was never written down anywhere — it lived in someone's head, and that head changed employers two years ago. What's left are formulas nobody dares modify and conventions nobody can explain anymore.
The ISO audit doesn't test your quality system. It tests the homemade system you inherited.
What ISO actually tests
This is where certification shows its true face. The ISO 9001 audit is presented as the validation of a quality system. In a small shop, it's almost never that. It's the annual stress-test of the homemade system.
The script is always the same. As the audit nears, everyone tenses up. Files are reopened, gaps are filled, missing records are reconstructed, answers are rehearsed for the questions people dread. You pass — most of the time, you pass. The certificate goes back on the wall. And the following Monday, the spreadsheet reopens exactly as before, until next year.
ISO doesn't push the shop toward a better tool. It checks that the assembly survives an inspection once a year. It's an exercise in survival, not construction. I've described elsewhere how to run your management review without a consultant and make it something other than a formality — but let's be honest about the starting point: for most small shops, the audit is a hurdle you clear, not a mirror you look into.
Why it stays this way
If the homemade system is so fragile, why does no one change it? I had the answer in my head for a long time, because it was mine.
First, the assumed price. Quality management software, in a small-shop owner's imagination, is "a big-company thing". Expensive, heavy, built for a thousand-person plant with a full quality department. Not for a thirty-person shop. You don't even look into it: you file it, in advance, under "not for us".
Then, because it holds. The homemade system works, in the sense that the shop ships, invoices, and passes its audits. The pain is real but diffuse — time lost here, a file that can't be found there, an annual stretch of stress — never sharp enough on any given day to justify rethinking everything. You put it off. Indefinitely.
And it has to be said plainly: most quality managers, myself included, talked themselves into believing their assembly was enough. You're proud of your file. You know it by heart, you've tamed it. Admitting a tool would do better is a little like admitting you spent years carrying something that shouldn't have weighed so much.
The real cost of the assembly
The homemade system isn't free. It's just unbilled. Its bill is paid elsewhere, in several currencies.
It's paid in knowledge that evaporates. Every departure takes away a piece of the system nobody wrote down — and the shop rediscovers, painfully, what the person who left knew by instinct. It's paid in blindness: an assembly that grows in layers doesn't let you see the critical points at a glance, let alone know whether the shop is reactive or truly under control. It's paid in re-keying, when the same information is retyped into five files that don't talk to each other — the exact opposite of an ecosystem where the work order is entered once. And it's paid, quietly, in the corners the assembly never covers well: metrology kept in a spreadsheet updated "when someone thinks of it".
None of these costs shows up on an accounting line. All of them weigh on the margin, on the nerves, and on the shop's ability to grow without its quality cracking.
What I wish I'd had
I don't write this from above. I was exactly the quality manager I'm describing. I had my processes, my binders, my files, and I told myself it was enough. It wasn't — not because I was bad at my job, but because the tool that would have turned my assembly into a real system never felt within reach. I never seriously looked for it. Like everyone, I made do, and I ended up believing that was the natural order of things.
It's by building software that I understood how unnatural it was. What I built, I didn't design to replace bad software — in the shops where I worked, there was no software to replace. I designed it to replace the file nobody chose, the inherited assembly, the knowledge that fits in a single head. For the thirty-person shop that always believed, like me, that a real system "wasn't for it". If you want the constructive version of all this, it's in the guide to structuring your quality, and in the principles I stand on.
But the opening question, I'll leave with you, because it deserves a pause before we talk about tools. The system that guarantees the quality of everything you ship — who designed it? Did you choose it, or inherit it? And if the person who truly understands it left tomorrow, would a system remain, or only files?
The convictions defended in this article are the ones that guided the development of Asterion Solutions, a suite of trade-specific tools built for small manufacturers who want to structure their quality without multiplying paperwork.