Here is a scene every founder will recognise. A client relationship goes wrong in a way that feels horribly familiar — the same slow disengagement, the same renewal that never quite closes, the same month-nine drift. Someone in the room says: didn't this happen with that account a few years back? And it had. But the person who managed it has left, the post-mortem was never written down, and the lesson the firm paid for in real money has to be bought all over again.
This is not carelessness. It is the default condition of a growing organisation, and it has a structure worth understanding, because understanding it is the first step to fixing it.
Memory is the first thing growth costs you
Neustadt and May, studying how governments use and misuse their own history, reach for a phrase coined by Hugh Heclo that travels straight into a scaling agency: a "government of strangers." Turnover means that the people making decisions are forever new to the institution they are running. They inherit the consequences of choices they did not make and cannot remember, and so they remake the same mistakes with perfect sincerity. An organisation with high churn is, functionally, a stranger to itself.
Their more striking insight is what they called placing organisations — the discipline of reading an institution through its own history, because institutions, like people, behave in ways their past has trained them to. They found this the least natural habit to teach and the highest in pay-off. A firm is not a blank slate that responds rationally to present incentives. It is a creature of its own story: the founding crisis that still shapes what it fears, the early client who set the template for how it sells, the one bad year that quietly governs every decision about cash.
Marc Bloch's defence of history rested on a phrase that names exactly what is at stake: history, he said, is the science of men in time. A firm read only as a present-tense snapshot — its current numbers, its current org chart — is being read in the one dimension that conceals the most. The forces actually driving its behaviour run backwards through time, and they are invisible to anyone looking only at the dashboard.
The knowledge that cannot be written down
There is a tempting answer to all this, and it is worth naming why it only half works. The obvious fix for a forgetful firm is documentation: write the lessons down, build the wiki, capture the playbook. Do it — but do not believe it is sufficient, because the most valuable thing a departing person takes with them is precisely the part that does not fit in a document.
Michael Oakeshott drew a distinction that every advisor should keep close. There is technical knowledge — the kind that can be written in a manual and looked up: the process, the template, the checklist. And there is practical knowledge, which, in his phrase, exists only in use. It is the judgment that knows when the rule applies and when it doesn't; the feel for which client is about to churn before any metric says so; the instinct for how hard to push in a negotiation. It cannot be formulated, only acquired — by doing the work alongside someone who already has it. Oakeshott's complaint against the "rationalist" was exactly this: the belief that the manual is the whole of the knowledge, that you can hand someone the crib and dispense with the apprenticeship.
This is why a firm can have immaculate documentation and still forget everything that matters. The wiki holds the technical knowledge. The practical knowledge walked out of the door with the person who had it, and no amount of writing-down would have caught it. Institutional memory, properly understood, is not a database. It is the deliberate transfer of practical knowledge from the people who hold it to the people who will need it — which happens through overlap, mentoring and shared work, not through a handover note.
How the firm forgets without noticing
There is a subtler failure than simply losing memory, and it is in some ways more dangerous: remembering the wrong thing, and ceasing to know that you are.
Carlota Perez, describing how technological revolutions settle in, shows how a new way of working hardens over time into what she calls common sense — embedded in the rules, the structures, the assumptions, until it is no longer experienced as a choice at all. What was once a deliberate response to specific conditions becomes simply "how things are done here," its origins forgotten, its continued fitness untested.
Tom Holland makes the same observation at the scale of a civilisation, and it is the most useful warning of all for a firm. The assumptions that shape us most powerfully, he argues, are the ones we have stopped being able to see — the inheritance so deep that we mistake it for human nature rather than history. A firm's deepest habits are like this. The way it prices, the way it pitches, the kind of work it instinctively says yes to: these feel like the firm's permanent character, when they are very often the fossilised solution to a problem the firm no longer has. The firm has not lost this memory. It has done something worse — it has converted memory into instinct, and instinct does not get reviewed.
So the firm that forgets is not only the one that loses its records when people leave. It is also the one that remembers so thoroughly, so invisibly, that it can no longer tell which of its habits are wisdom and which are merely old.
What it changes on Monday
The practical programme that follows is modest, and deliberately so — institutional memory built as bureaucracy will be ignored, and should be. A few habits, in a twenty- to fifty-person firm, do most of the work.
Write the decision down at the time, not the lesson after the fact — a short, honest note of what was decided, what was known, and what was merely assumed, dated and kept. A contemporaneous record resists the flattering rewrite that hindsight performs.
Protect the overlap. The single highest-return act of memory is keeping a departing person long enough, and close enough to their successor, to pass on the practical knowledge that no document will hold. Treat that overlap as an investment, not an inconvenience.
Periodically ask of any settled habit: what was this the answer to, and is that still our problem? This is the same question that opens an inquiry into a commercial model, turned on the firm's own instincts. Holland's discipline of seeing one's own assumptions is hard precisely because the assumptions feel like facts; asking the question on a schedule is how you make the invisible visible again.
The caveat
The risk here is nostalgia, and it is real. Memory is for making better decisions, not for venerating how things used to be done. An organisation can fail by forgetting its hard-won lessons; it can fail just as surely by remembering them too well, defending a habit long after the conditions that justified it have gone. The same faculty that preserves wisdom preserves superstition, and from the inside they are very hard to tell apart.
That is the whole point. The aim is not to remember more, but to remember deliberately — to know which of your instincts you chose and which you merely inherited, and to be willing to re-examine both. A firm that can do that is not at the mercy of its own past. It is in conversation with it.
So the question to leave a growing firm with is not "what should we write down?" It is the harder one underneath: of everything this firm now does by instinct, how much could anyone here still explain — and what has it cost us, quietly, to have forgotten the reasons?
This is a companion piece to the argument at the argument. If it resonates — or if you want to push back on it — I'd be glad to hear from you: chris@christhurling.com.