Tacit knowledge, the asset nobody inventories

The Monday after

One Friday in June, an engineer retires after twenty-two years in the laboratory. A book as a present, a farewell drink, a promise to see each other again. Her protocols are on the server, her notebooks neatly filed. Nobody is worried, and everyone tells her to enjoy it. On Monday, the experiments carry on. The results come in as usual.

In November, the first problem: the yield of a protein purification falls threefold. The reagents are suspected. A new batch is tried, without success. In January, a PhD student tries to bring a frozen cell line back into culture. The cells die. In March, someone finally makes the connection: these procedures all relied on the same engineer. She knew the small movements that made the difference. Nobody had thought to ask her about them.

It will take more than a year to get back to the previous results. Dozens of attempts.

The story is invented, but it should ring true for many research directors. Knowledge is not necessarily missed on the day its holder leaves. Its absence is discovered much later, when an experiment stops working. And the first thing people look for is a technical fault.

Knowledge held in the hands

Tacit knowledge is what we know how to do without always being able to explain it. The philosopher and chemist Michael Polanyi put it this way: “we can know more than we can tell”. In the laboratory, it sometimes comes down to almost nothing. A technician can tell the state of a culture from the colour of the medium. A PhD student hears that a tube is badly balanced in the centrifuge. A researcher adjusts an incubation to the humidity in the room. An engineer waits ten minutes before using an instrument that drifts every morning. None of this necessarily appears in the protocol.

These details could be mistaken for quirks. Yet they can decide whether an experiment succeeds. In one survey, more than half of researchers said they had already failed to reproduce their own results. Tacit knowledge does not explain everything, but it counts: even the person who holds it sometimes struggles to put it into words.

What the notebook never says

The lab notebook mostly describes what worked. Unsuccessful attempts take up less room in it. It sets out the steps.

What is missing? The attempts abandoned because they seemed pointless. Judgements such as “fairly cloudy” or “nice and dry”. The known little faults in the equipment: a pipette that leaks, a water bath that is warmer on one side. The reasons for choosing a supplier or a concentration. And the useful contacts: the person who repairs the instrument, the colleague who knows another method. This is not necessarily a deliberate omission. After repeating a movement so many times, the expert no longer thinks to describe it. What they master best ends up seeming obvious to them.

Spotting a critical skill

To spot a critical skill, there is no need to wait for an audit. Sometimes all it takes is listening to what is said in the team.

“Ask Claire.” The phrase comes up often. An experiment waits until one particular person is available. Results vary depending on who does the bench work. All newcomers are trained by the same colleague. While that colleague is on holiday, some experiments fail. As for that instrument, nobody else dares touch it. Behind these ordinary situations lies the same problem.

The rule of three

How many people should know how to carry out an essential technique?

With only one person, everything rests on them. A departure or a long absence is enough to bring the work to a halt. With two, the risk falls but does not disappear. If one leaves, the other must both keep the experiments going and train someone. They do not always have the time. With three, the team still has two competent people after a departure. One can carry on the work while the other helps the next person take over. It is not an absolute guarantee, but it is a far more comfortable safety margin.

In summary: how many people for an essential technique

Why look at three years

Three-year retention gives a first benchmark: of the people who were there three years ago, how many still work in the team?

Three years is the usual length of a PhD. It is also a period during which one can learn a demanding technique and begin to pass it on. If departures follow one another too quickly, knowledge circulates without always having time to settle. This figure also raises a fundamental question: are we training researchers who stay, or researchers who move on elsewhere? Both have their value. In the second case, you simply need to make sure that the knowledge acquired does not depart with those who leave the team.

Before a planned departure: three steps

Retirement, the end of a contract, a thesis defence: some departures can be foreseen.

1. Have them do it rather than show them. The successor takes over, with the expert watching. It is by trying that they discover what is missing from the instructions. Every hesitation deserves an explanation.
2. Film the technique and talk it through. Watching the expert at work, hearing them explain what they observe and why they correct a movement: video preserves details that are hard to put into words.
3. Reproduce a known result. Before the departure, the successor attempts a reference experiment alone. If it works, that is encouraging. If not, there is still time to understand what they are missing.

When the departure comes as a surprise

A sudden resignation, an illness or an accident can turn everything upside down. You then need to act quickly to identify the knowledge at risk and try to preserve what can still be passed on.

Start by listing the projects, instruments and contacts that only this person had mastered. If the situation allows, keep in touch with them for targeted questions. Go back to the raw data, the files and the samples: they sometimes hold valuable clues. Finally, contact your partners. One of them may know the technique. You also need to plan for a difficult period. Research on the co-authors of scientists who died suddenly shows a drop in their output. Rebuilding working habits takes time.

But the best way to face an unexpected departure is still not to wait for it to happen.

What fails, and why

Two solutions seem obvious. In practice, they often disappoint.

First, the procedure sheet. Written by the expert, it leaves out the movements they find obvious. It explains what to do when all is going well, less how to react when something unexpected happens. Have a beginner test it: their questions will reveal the passages that need reworking. Then the exit interview. Often too short and too late, it sometimes boils down to: “What do you know that we should know?” It is hard to answer such a broad question. We rarely become aware of our automatic reflexes until someone stumbles over them.

Both approaches share the same limitation: they try to lock practical knowledge into a document.

The reading that stayed with me

In 2013, the biologist Mina Bissell published a piece in Nature on the difficulty of reproducing certain experiments. Her point caught my attention: results can depend on movements learned over the years. She describes how researchers had to come and work in her laboratory, alongside her team, in order to reproduce them.

This reading reminded me of something essential. Papers pass on results and methods, but they cannot always replace the presence of an experienced person.

Grandmother's recipe

Almost all of us have known that family recipe, written on a slightly stained card. You follow the instructions, and yet the cake does not taste the same. What exactly does “a pinch” mean? At what point does the dough “hold together”? And “a good hot oven”.

The recipe is still there. What is missing is the knack, those small movements and adjustments that experience has made instinctive and that nobody thought to pass on.

In the laboratory, the same is sometimes true. Protocols keep the recipe, not necessarily the touch. Preserving your scientific capital also means making sure that someone can pass that touch on.