New Zealand’s trades and technical workforce is ageing, with experienced people retiring faster than the industry replaces them. Meanwhile new entrants to workplace-based training fell 16 percent in 2025 to 16,780.
That combination has a predictable consequence, and it is one of the few workforce problems a business can see coming years in advance.
Know your own exposure
Most businesses have never done the arithmetic. It takes an afternoon:
- List everyone whose departure would create a genuine capability gap.
- Note their likely retirement horizon.
- For each, identify what they know that nobody else does — technical judgement, client relationships, undocumented process, the history of long-running projects.
- Ask what recovery would actually take if they left next month.
Most businesses find several uncomfortable answers. That list is what to work on, and it is a business continuity issue as much as a people one.
Knowledge transfer is harder than recruitment
When an experienced person leaves, the drawings and files remain. The reasoning does not.
Why a particular detail was chosen, what was tried and rejected, which client cares about what, where the buried problems are on a long-running job — none of that is in the file, and a handover document written in the final fortnight does not capture it.
What works better:
- Decision records. A short note with each significant technical decision recording what was considered and why. Written at the time it takes minutes; reconstructed later it is impossible.
- Paired working on hard problems. Two people on a difficult problem is the most reliable knowledge transfer mechanism available, and it looks inefficient on a timesheet.
- Structured mentoring with allocated time rather than goodwill.
- Technical review as a conversation, not a signature. Ten minutes explaining the reasoning teaches more than a marked-up drawing returned by email.
- Client relationship mapping, so no relationship is held by one person only.
- Standard details and internal guidance capturing how your business does things, maintained rather than written once.
Phased retirement is under-used
The binary of full-time to nothing suits almost nobody. Many experienced people would work reduced hours for several more years if offered, and that period is exactly when knowledge transfer happens best.
Arrangements that work: reduced days, a mentoring or review role rather than production, project-based engagement after leaving, or a consultancy arrangement for specific expertise.
Note that a genuine contractor arrangement after retirement needs to look like contracting in substance — not the same job with a different label — or it risks being characterised as continued employment.
Supporting registration keeps capability in the business
For engineering practices, Chartered Professional Engineer registration matters commercially, since producer statements and council acceptance reference it.
The habit worth instilling in every graduate from year one is a running record: project, role, problem, decisions, reasoning, outcome. Reconstructing six years of evidence at application time is genuinely difficult, and it is the most common reason capable engineers stall.
Firms that fund the process, allocate time for it and provide mentoring retain people who would otherwise leave to get it elsewhere.
The retention drivers are not money
Exit interviews consistently identify the same things: no visible progression, work that stopped being interesting, poor management by someone promoted for technical ability with no preparation, sustained overload, and inflexibility.
A technical progression path matters particularly. Not everyone wants to manage, and forcing progression through management loses good technical people and creates poor managers. A senior technical track with genuine status, pay and autonomy retains people who would otherwise leave for it elsewhere.
Succession in an owner-operated business
Where the retiring person is the owner, this becomes a business succession question as well.
The work that adds most value is slow: reducing owner dependence, documenting processes, building a management layer, formalising customer contracts, and getting financial records to a state a buyer or lender would accept.
Five to ten years is realistic. Successions attempted in twelve months, usually triggered by a health event, are where the expensive mistakes cluster.
Education Counts publishes workforce and training pipeline data under an open licence, and Engineering New Zealand publishes competence standards and CPD requirements.
Figures: Education Counts workplace-based learner data, 2025. General information only.
