Most succession plans I have reviewed cover about forty roles. Executive committee, their direct reports, a handful of high-potential names on a slide. Meanwhile the person who knows why that particular substation trips in the rain, and what the workaround is, retires in November and nobody has scheduled a handover.
That is the real exposure. It rarely appears on a risk register because it does not fail loudly until it fails expensively.
Age profile is a leading indicator
Run a simple query: headcount by age band, filtered to roles requiring statutory appointment, specialist authorisation or ten-plus years of plant-specific experience. In the utilities and heavy industry clients I work with, that curve is consistently top-heavy. Half the authorised switching staff over 52 is not unusual.
Now overlay the qualification pipeline. How many candidates are currently in structured development for those roles, and what is the realistic time to competence? For an authorised MV switching operator you are looking at years, not months, and the constraint is supervised exposure rather than classroom time.
When you put the two curves on one page, the conversation changes tone quickly. I have used exactly this slide to unlock funding that three years of general appeals could not.
Tacit knowledge does not transfer through documents
The instinct is to write it all down. Procedures, handover packs, wikis. Useful, but limited. The knowledge that matters most is judgement — knowing which alarm is real, which contractor's work needs checking, which valve seats badly and needs the extra quarter turn.
That transfers through working alongside someone, being allowed to make a decision, and having it critiqued. Which means the only real intervention is deliberate overlap: putting the successor in the seat while the incumbent is still there and still accountable.
Overlap costs money. It shows up as duplicated headcount on a budget line and there is always pressure to shorten it. The plants that do it properly treat the last twelve to eighteen months of a senior technical career as a formal mentoring appointment with defined deliverables — not a wind-down.
The reverse-mentoring trap
There is a fashionable version of this where younger staff teach senior staff about digital tools and everyone declares knowledge transfer achieved. It is pleasant and it is not the same thing. Showing a 58-year-old foreman how to use a tablet does not move a single piece of operating judgement in the direction it needs to travel.
Build capability where the scarcity is, not where the enthusiasm is
Development budgets drift towards people who ask for them. Ambitious, articulate staff in corporate functions submit good motivations. Meanwhile the artisan cohort that carries your statutory risk submits nothing, because nobody asked and the forms are in English on a system they never log into.
A short audit will tell you where your spend actually went last year by job family. In most organisations it is badly misaligned with where the operating risk sits.
Practical sequence
- Identify roles where the loss of one or two individuals materially degrades operating capability or statutory compliance. Usually fewer than thirty.
- For each, record the realistic time to competence and the current pipeline depth. Two deep is comfortable, one deep is a risk, zero deep is a live issue.
- Set formal overlap periods for the top exposures and fund them explicitly, with named mentor and mentee.
- Convert the knowledge that can be documented — switching schedules, plant quirks, contractor history — during the overlap, not after the exit interview.
- Review quarterly with operations, not annually with HR.
The uncomfortable part
Doing this properly means telling some capable people that they are not the priority this year, because the risk sits elsewhere. It means paying for two people to do one job for a period. And it means accepting that some knowledge will be lost regardless, so you had better be deliberate about which.
None of that is comfortable. It is considerably less uncomfortable than the alternative, which I have watched twice: an organisation discovering, in the middle of a fault, that the only person who understood a system left eight months ago and the documentation was never true in the first place.