Institutional memory

When your best person leaves, so does the process.

In most founder-led scale-ups, the real process lives in a few people’s heads, so when one of them leaves, the process leaves too. If it’s not written down, you don’t have a process – you have a person. The fix is a habit of capturing how work is done and why decisions were made, not a one-off handover document.

By Simon Wakeman, Principal Consultant & Fractional COO
· 5 minute read

The best account manager in the business hands in their notice, and the only version of how their biggest client actually works leaves with them. The notes are thin, the handover was rushed and a few weeks later the person who could explain any of it works somewhere else.

It’s one of the most common ways a founder-led business gets fragile without noticing. And it’s much cheaper to fix before someone hands in their notice than after.

Why does knowledge end up living in people’s heads?

Because, early on, that’s the fastest way to work. When there are ten of you, writing things down feels like overhead. You just ask the person who knows.

Over time, those people become the process. The account lead who knows which stakeholder actually signs things off. The ops manager who knows why invoices for one client go out on a different date. The developer who remembers why that integration was built the way it was. None of it is written anywhere, and none of it feels urgent, because the person is right there.

There’s a less comfortable reason too. Being the person who knows things brings status. People who hold knowledge nobody else has become the go-to, and asking them to write it all down can feel like asking them to give some of that up. That’s rarely deliberate, but it’s real. Sam Spurlin and Rodney Evans of The Ready explore this in an episode of At Work with The Ready on institutional knowledge.

What does it actually cost when someone leaves?

The obvious costs are the ones you can see: a gap in capacity and a few months of a new hire finding their feet. The hidden costs are bigger.

  • Clients feel it first. The new person asks questions the client thought were settled years ago.
  • Mistakes get repeated. Nobody remembers why something was tried and dropped, so it gets tried again.
  • The founder gets pulled back in. When nobody else knows, the question lands with you.
  • New starters take longer. Every new hire walks into the same gap and pieces things together from whoever has time to explain.

That last one matters more than most founders realise. In a scaling business people join all the time, so a slow start isn’t a one-off. It’s a tax on every hire.

Isn’t a good handover enough?

Rarely. A handover written in a notice period is a leaving present, and usually a rushed one. It captures what someone remembers under pressure, not what they know.

What gets lost is the stuff that never seemed worth writing down: the exceptions, the history, the reasons. And the reasons are often the most valuable part. A decision without its reasoning is just an instruction, and instructions get ignored or undone the moment circumstances change.

Institutional memory has to be a practice, something the business does every week, not something it scrambles to do when someone leaves.

What should you actually write down?

Not everything. Trying to document every task is how you end up with a wiki nobody reads. Focus on what would hurt most if it disappeared tomorrow.

How the critical work gets done. The handful of workflows the business earns its money through: winning work, delivering it, getting paid. The steps, the checklists people genuinely rely on and which tool each part happens in.

Who owns what. Which role owns each workflow, client or system, and which decisions they can make without asking. If your answer is “it depends”, that’s worth writing down too, because it’s probably where things fall through the cracks.

Why decisions were made. Not just what was decided, but the options you weighed and why you landed where you did. A simple decision log keeps that context where everyone can find it, stops agreements being quietly forgotten and makes onboarding far easier.

A good test is whether a capable new hire could read it and know where to start.

Why won’t my team document anything?

Usually because the cost to them is higher than the benefit. Writing things down feels like extra admin on top of the real work. Sharing a half-finished draft can feel risky if people get picked apart for it. And if nobody ever reads what gets written, why bother?

So change the maths. Make capturing knowledge part of how work already happens, not a separate task. Give it a regular slot in the week. Make sure leaders visibly use the shared record, and point people to it instead of answering from memory. If the founder keeps answering every question personally, the team learns that the document doesn’t matter.

It also helps to stop expecting people to write from scratch. Some of the best knowledge capture starts as a conversation: sit with the person who knows, ask them to walk you through it and record it. AI transcription makes that much easier than it used to be. The Ready’s hosts make the same case for lowering the cost of sharing until it becomes the easy option.

Where do you start before anyone resigns?

  1. Run the leaving test. List your five most critical people. For each, write down what would break if they left tomorrow. That list is your documentation backlog.
  2. Capture one workflow by talking it through. Choose one that matters commercially and lives mostly in one person’s head. Spend an hour with them walking through it and record the conversation.
  3. Start a decision log and give it an owner. One shared document, three columns: what we decided, why we decided it and who owns it. Add to it at the end of every leadership meeting, and make one person responsible for keeping it current.
  4. Stop answering from memory. Next time someone asks you how something works, find or write the answer somewhere shared and send them the link.

How do you make it stick?

Treat it like any other operating habit. A one-off documentation sprint will feel great for a month and then quietly decay. What lasts is a small, boring rhythm: decisions logged every week, workflows reviewed when they change and new starters asked to fix anything they found confusing. Our guide on how to make operational change stick covers how to keep habits like this from fading.

Knowledge capture is one part of a wider business operating system, and it works best alongside clear ownership and a regular leadership rhythm. Our guide on how to build a business operating system shows how those pieces fit together. And if you want to know where your biggest key-person risks really sit, the Find diagnostic will show you in 14 days, with a prioritised roadmap of what to fix first.

Some commonly asked questions from founders…

Why is undocumented knowledge such a big risk for a growing business?

Because it’s invisible until it walks out of the door. When how a client, a process or a decision works lives only in one person’s head, everyone else depends on that person – and the business quietly depends on them staying. Growing teams add new starters faster than that knowledge spreads, so the gap gets wider.

Isn’t a good handover enough when someone leaves?

Rarely. A handover written in someone’s notice period captures what they remember under pressure, not what they know. The context, the exceptions and the reasons behind past decisions usually get lost. Capturing knowledge as part of normal work, long before anyone resigns, is far more reliable.

Won’t writing everything down create bureaucracy?

Not if you write down the right things. You don’t need a manual for every task. You need the critical workflows, who owns what, the checklists people actually use and a record of why big decisions were made. Short, findable and kept current beats long and comprehensive.

Why won’t my team document how they work?

Usually because it costs them more than it gives back. Documenting can feel like extra admin, and being the person who knows things brings status. Make sharing easier than not sharing, give it a regular slot in the week and make sure leaders visibly use what gets written.

What should a growing business write down first?

Start with the few workflows that win, deliver and bill your work, and the knowledge held by only one person. Then add a simple decision log that records what was decided and why. Those three things cover most of the risk of a key person leaving.

How does Mallard help businesses reduce key-person risk?

Mallard’s Find diagnostic looks at 262 operational data points across people, practices and technology, including how knowledge and decisions are captured, and ends with a prioritised roadmap. An Improve sprint then puts a senior Mallard operator inside the business for three months to build the practice alongside your team.

Stop being a bottleneck in your own business

Book your free 30 minute
founder review call