The Varnish Nobody Could Replicate

VISPAICO Journal — Issue 003

---

For nearly two centuries, violin makers have tried to reproduce the varnish Antonio Stradivari used in Cremona in the early 1700s. Chemists have analysed it. Physicists have X-rayed instruments layer by layer. Entire careers have been spent trying to reverse-engineer a formula that, by any reasonable modern standard, should not be this difficult to copy. It is, after all, just varnish.

And yet no one has fully succeeded, because the varnish was never really the point. Stradivari's workshop held something more elusive than a recipe: a set of judgments about wood density, humidity, ageing, and application technique, refined over decades by a master craftsman who trained his apprentices by feel rather than by formula, and who died before anyone thought to write the whole of it down. When he died, the workshop did not lose a chemical formula. It lost a body of experience that had never existed anywhere except inside him and the small number of people he had personally trained.

This is the story companies tell themselves is unique to their situation whenever a long-tenured employee retires. It is not unique. It is one of the oldest patterns in economic history, and businesses have been re-learning it, at great expense, for centuries.

---

The Difference Between Losing a Person and Losing a Decade

When a company loses an employee, the language used internally is almost always about headcount. A seat needs filling. A req goes up on the careers page. HR runs the numbers on time-to-hire. This language, however practical, obscures the actual loss, which is rarely about the seat at all.

Consider a plant manager who has run the same manufacturing line for eighteen years. On paper, her value is captured in her title and her salary. In practice, her value lies somewhere else entirely: in the fact that she knows, without needing to check anything, which machine tends to drift out of calibration in humid weather, which supplier's raw material batches occasionally run slightly off-spec in ways that don't show up on the certificate, and which combination of settings the previous plant manager discovered, through trial and painful error, fifteen years ago.

None of this is written down anywhere, because none of it ever needed to be. It lived in her, the way a language lives in the people who speak it rather than in any dictionary. When she retires, the company will conduct an orderly handover — a few weeks of shadowing, a folder of notes assembled in a hurry. And within eighteen months, in nearly every case like this, some version of a problem she would have caught instantly will surface, cost real money, and take a new team most of a year to fully diagnose and prevent from happening again.

The company did not lose an employee. It lost a decade of accumulated pattern-recognition that took that long to build and disappeared in a single afternoon of exit paperwork.

---

What Oral Traditions Understood That Companies Forgot

Long before anyone wrote anything down, entire civilisations preserved extraordinarily complex bodies of knowledge — legal codes, medical practices, genealogies stretching back centuries — through oral tradition alone. The griots of West Africa, for instance, served for generations as living archives: trained from childhood to memorise and recite the history, law, and lineage of entire kingdoms, word for word, because the alternative was for that knowledge to simply cease to exist the moment the previous griot died.

What is striking about these traditions is not that they relied on memory — every pre-literate society did that, by necessity. What is striking is how seriously they treated the transfer of that memory as a discipline in its own right. A griot's apprenticeship could last decades. The role was not casual. It was understood, explicitly, that the kingdom's continuity depended on this transfer happening correctly, and enormous institutional effort went into making sure it did.

Modern companies, by contrast, treat knowledge transfer as an afterthought — a checklist item during offboarding, squeezed into someone's final two weeks alongside returning a laptop and cancelling access badges. The griot's civilisation understood that losing an elder without a properly trained successor was a genuine institutional crisis. The modern company treats the same event as a line item in a departure email.

This is not a criticism of any individual company. It is an observation about how casually an entire economic system has come to treat something oral traditions once guarded with real seriousness: the fact that expertise does not automatically survive the person who holds it. It has to be deliberately carried forward, or it simply stops.

---

The Guild's Answer, and Its Limits

Medieval guilds understood this problem well enough to build an entire economic structure around solving it. The apprenticeship system — years of unpaid or low-paid labour under a master craftsman, followed by years more as a journeyman before one could be certified a master oneself — existed almost entirely to solve the transfer problem. A guild could not survive if its accumulated expertise in metalwork, or weaving, or masonry died with each generation of masters. So it built a structure, rigid and slow, that forced knowledge to move deliberately from one person to the next before the first person was gone.

It worked, in the sense that certain crafts and their standards did survive for centuries. But it worked slowly, and it worked narrowly. An apprenticeship might take seven years to produce one trained successor, and even then, that successor absorbed only what their specific master happened to know, filtered through whatever that particular relationship allowed to be taught. The guild system preserved expertise, but at enormous cost in time, and with real limits on how many people that expertise could ever reach.

This is roughly the position most companies are still in today, four hundred years later, whether they realise it or not. Expertise passes, when it passes at all, through informal apprenticeship: a junior analyst absorbing judgment from a senior one through years of proximity, a new operations lead learning the unwritten rules by watching someone more experienced make dozens of small decisions over time. It is real. It does work, when it is allowed to happen. But it is exactly as slow, exactly as narrow, and exactly as fragile as the guild system was — dependent entirely on the master staying long enough, and the apprentice being patient enough, for the transfer to complete before something interrupts it.

Companies rarely allow that much time anymore. People change roles every few years. Reorganisations sever the mentoring relationships that would have completed the transfer. The guild's seven-year apprenticeship has, in most modern businesses, been quietly compressed into a two-week handover document, and everyone involved knows, without quite saying so, that this is not really enough.

---

Why More Documentation Was Never the Answer

The instinctive corporate response to this problem has always been the same: write it down. Build the manual. Create the wiki. Document the process. This impulse is not wrong, exactly, but it consistently underestimates what is actually being lost.

A process can be documented. Judgment, mostly, cannot — not in the form documentation usually takes. The plant manager could, in principle, write down the specific temperature range at which a particular machine tends to drift. What she cannot easily write down is the accumulated instinct that made her check that machine first, out of forty possible machines, on a humid Tuesday in October, before anything had visibly gone wrong. That instinct was built from thousands of small observations she was never fully conscious of making, and it resists being reduced to a bullet point, because it was never stored as a bullet point to begin with. It was stored as pattern-recognition, built through repetition, largely below the threshold of language.

This is precisely why documentation efforts inside companies so often disappoint everyone involved. The manual gets written. It captures the procedures. It misses almost entirely the judgment about when the procedures don't quite apply — which is, inconveniently, the exact situation in which experienced judgment matters most and novice judgment fails most often.

---

What Actually Changes Now

The genuine shift underway has nothing to do with replacing the plant manager, or the senior associate, or the client-facing partner whose instincts took twenty years to build. Nothing currently available, or plausibly on the horizon, replicates that kind of embodied judgment. The shift is narrower, and in some ways more modest, than the language usually applied to it suggests: for the first time, it is possible to capture the reasoning behind decisions — not just the decision itself, but the specific considerations that led to it — in a form that outlasts the person who made it, and make that reasoning available to someone who was not there to absorb it firsthand.

This is a meaningfully different thing than automation. Automation replaces a task. What is being described here preserves a way of thinking — the actual chain of considerations a plant manager runs through, largely unconsciously, when she decides which machine to check first. Captured properly, that reasoning becomes something a successor can draw on directly, rather than something they must spend eighteen months and one expensive mistake rediscovering on their own.

It is worth being precise about what this does not do. It does not make junior staff as good as senior staff overnight. Nothing does that, and any framing suggesting otherwise deserves the scepticism a careful executive would naturally bring to it. What it does is shorten, meaningfully, the distance between an organisation's accumulated experience and the people who need that experience today but did not have twenty years to acquire it themselves. It is the difference between an apprentice starting from nothing and an apprentice starting from everything their master already learned, made available at the moment it is needed rather than doled out slowly across a seven-year term.

---

The Real Inheritance

Every company, if it is honest with itself, is a repository of decisions made under pressure, judgments refined through repeated failure, and instincts built by people who rarely had time to explain why they did what they did — only that it worked, and that they had learned, the hard way, not to do it the other way.

This inheritance has always been the most valuable thing a business possesses, more valuable in most cases than its brand, its client list, or its balance sheet, because all of those things can be rebuilt with enough time and capital, while genuine experience cannot be rebuilt at all — it can only be relived, at whatever cost the world chooses to charge the second time around.

For most of business history, this inheritance was mortal in a very literal sense. It lived inside people, and when those people left, retired, or simply forgot details they hadn't used in years, the inheritance diminished with them, and the next generation started measurably poorer than the last, whether anyone noticed or not.

What changes now is not that experience becomes immortal — nothing makes it immortal, and no company should believe otherwise. What changes is that, for the first time, the gap between one generation's hard-won judgment and the next generation's starting point can actually narrow instead of widening by default. The griots understood that a kingdom's memory required deliberate, disciplined effort to survive. The guilds understood that expertise needed a structure built specifically to outlast any one master's lifetime.

The companies that understand this next will not be remembered for adopting a piece of technology. They will be remembered for being the first, in their industries, to take seriously an old and uncomfortable truth: that losing an employee was never really the loss that mattered. What mattered, all along, was everything that employee had learned — and whether anyone thought to carry it forward before it was too late to ask.