The short version
- Two weeks of notice cannot transfer two years of context. The arithmetic does not work, however well the fortnight is run.
- Notice periods are also badly suited to it: the departing person is disengaging and the receiving person has their own job.
- Spend the fortnight on what only they can give: reasoning, relationships, and dead ends. Not system descriptions.
- What they wrote down over two years stays in the record. That is the only version of this that actually works, and it has to be built long before the resignation.
Somebody resigns and a knowledge transfer plan appears: sessions booked, a document started, a colleague assigned to shadow. It is a reasonable response to bad news and it recovers a small fraction of what is leaving.
"Two weeks' notice, two years of context."
The arithmetic of two weeks
Two years of work is roughly four hundred working days of accumulated decisions, attempts, conversations, and judgement. A notice period offers maybe twenty hours of transfer time once handover meetings and closing work are accounted for.
Even at a generous rate, that captures the outline. What it cannot capture is the reasoning behind four hundred days of small choices, and that reasoning is precisely what the next person will need, usually six months later when something breaks and the question is why it was built this way.
This is worth stating plainly because it changes what you do with the fortnight. If you accept that comprehensive transfer is impossible, you stop attempting it and start triaging.
Why a notice period is the worst conditions
Three reasons, all predictable and none of them anybody's fault.
The departing person is disengaging. This is normal and human: their attention is on what comes next, and the motivation to document thoroughly for an organisation they are leaving is low even when the relationship is good.
The receiving person has a job already. Shadowing is added on top of their existing work, so they attend the sessions and retain a fraction.
And there is no feedback loop. Transfer works when the receiver makes decisions and the expert corrects them. In a notice period there is no time for a cycle of attempt and correction, so what gets transferred is facts rather than judgement.
What to spend the fortnight on
| Spend time on | Do not spend time on |
|---|---|
| Why current commitments exist and what they trade against | Describing how systems work, which is in the code |
| Relationships: who needs warning, who is unhappy, who to trust | A complete list of responsibilities |
| Dead ends: what was tried and abandoned, and why | Recording long walkthrough videos nobody watches |
| The three things they worry about breaking | A comprehensive wiki nobody will maintain |
The left column is a day of focused conversation and a few pages. The right column is what most notice periods actually produce, and almost none of it is consulted.
One specific exercise is worth more than any document: sit the receiving person down and have them make three real decisions in the area, with the departing person critiquing each. Judgement transfers through correction, and that is the only mechanism available at speed.
What survives is what was written along the way
The honest conclusion is that this problem is not solvable at the notice period. It is solvable at the level of habit, and the habit has to exist for years before it is needed.
A daily brief is that habit. Ninety seconds at the end of each day, mostly pre-drafted from the work that already happened: current state, decisions made and why, open questions, what was tried and abandoned. Over two years that accumulates into the reasoning behind four hundred days of choices, in the person's own words, written when the context was fresh rather than reconstructed under time pressure in a final fortnight.
With StandIn, that record stays answerable after they go. A colleague six months later can ask why the pipeline was built this way and get the departed person's own reasoning with a source under it, rather than a guess from whoever inherited it. It never guesses, so when something genuinely was not recorded it says so, which is also useful: the team knows it is reasoning from scratch rather than reconstructing. See how StandIn works.
That does not make a resignation painless. It changes what is lost from the whole reasoning of a role to the part that was never written down.
Common Questions
How do you do knowledge transfer in a two-week notice period?
Triage rather than attempt completeness. Spend the time on reasoning behind commitments, relationships, dead ends, and the things the person worries about breaking. Have the successor make real decisions and be corrected, because that is the only way judgement moves quickly.
Should we ask a departing person to write documentation?
A few pages of reasoning, yes. A comprehensive document, no: it will not be finished, it will not be maintained, and the time is better spent on the conversations only they can have.
What is the most commonly lost thing when someone leaves?
Why decisions were made, and what was tried and rejected. Both are cheap to record on the day and effectively unrecoverable afterwards, which is why they are the highest-value things to capture as a daily habit.
How do we prepare for resignations we cannot predict?
The same way you prepare for unplanned absence: a daily record of decisions and state, kept as a habit rather than a project. It is the only form of knowledge capture that exists before anyone knows it is needed.
When you're off, your StandIn is on.
It answers your teammates' questions from work you've already done, in your words, with a source under every answer.