Key-Person Dependency: What Stops if a Specialist Disappears?

Every manufacturing business has at least one person whose absence would cause more disruption than the organisation chart suggests.

It might be the programmer who understands a legacy machine no one else wants to touch. The setter who can hear when a process is drifting before a measurement confirms it. The quality manager who knows why a customer approval was written in a particular way. Or the commercial director whose personal relationship has kept a major account stable through years of change.

These people are often described as invaluable. Operationally, that can also mean the business has allowed a dependency to form around them.

Key-person dependency is not simply an HR or succession issue. It is concentration risk in people. If one person becomes unavailable, the effect can pass through production, quality, customer relationships, cash and contractual performance at the same time.

The useful question is not, “Could we eventually recruit someone else?” It is:

Which specialist person would create the greatest operational or financial impact if unavailable for 30 days?

This is concentration risk in people

A name on an organisation chart can hide several different dependencies.

One person may hold the permissions needed to access a system, the practical knowledge needed to set a machine, the authority needed to release work and the customer confidence needed to explain a delay. Their formal job description captures the role. It may not capture what the business actually relies on them to do.

That distinction matters. A vacancy can be filled. A dependency has to be understood before it can be transferred, duplicated or controlled.

The exposure is rarely limited to senior leadership. In manufacturing, it often sits with technical specialists, maintenance engineers, estimators, schedulers, quality personnel, toolmakers, process owners and experienced operators. The person may not manage anyone, yet their absence can stop other people from working.

What actually stops?

The first impact is not always an empty workstation.

Production may continue until the next difficult setup, tool failure or process deviation. A quotation may wait because no one else understands the assumptions behind the estimate. A customer complaint may take longer to resolve because the history sits in one person’s memory. An apparently routine system change may expose the fact that only one employee has administrator access.

That creates a dangerous delay between the person becoming unavailable and the business recognising the full consequence.

The dependency may include:

• undocumented machine settings, tooling choices and workarounds;
• knowledge of why a specification, tolerance or inspection step exists;
• passwords, permissions or control over a critical digital system;
• personal relationships with customers, suppliers or technical approvers;
• the judgement needed to distinguish an acceptable variation from an emerging defect;
• ownership of maintenance, calibration, certification or compliance routines; and
• an understanding of which shortcuts are harmless and which could create a failure later.

The important question is therefore not only who covers the job. It is whether the replacement has the information, access, authority and practical competence to keep the dependent activity operating.

Tacit knowledge is not owned until it is transferable

Manufacturers often say that critical knowledge belongs to the company because it was developed at work. That may be legally correct, but it does not mean the knowledge is operationally available.

If it has never been written down, demonstrated, tested or transferred, the business may own knowledge it cannot use without the person carrying it.

This is particularly important in a workforce shaped by long service, retirement and persistent technical-skills shortages. Make UK reports that manufacturers continue to face skills constraints and an ageing workforce, while its skills research highlights the risk that retirement removes institutional memory as well as headcount.

Documentation helps, but a folder of procedures is not proof of resilience. Some knowledge can only be transferred through supervised practice: seeing the unusual failure, completing the difficult changeover or handling the customer conversation when the standard process no longer fits.

A credible control therefore combines written instructions with access, training, rehearsal and evidence that another person can perform the task without the original specialist beside them.

The replacement-time assumption

Boards can underestimate the exposure by treating recruitment time as recovery time.

Hiring a replacement may be only the first stage. The business may also need to allow for notice periods, machine or product familiarisation, supervised experience, internal authorisation, customer acceptance and the rebuilding of trusted relationships. The real recovery period will vary significantly by role, process, location and labour market.

There is no defensible universal timetable. A manufacturer should establish its own answer by mapping what the person controls and asking how long each dependency would take to restore—not merely how long an advert would run.

This is also why the financial effect can exceed salary and recruitment cost. During the recovery period, the business may carry lost output, overtime, scrap, delayed deliveries, external support, management distraction or customer concessions.

Insurance is only one part of the response

Key-person insurance may provide liquidity following an event covered by the policy. Whether it is suitable, what it covers and how much it pays depend on the actual terms and the business’s circumstances.

It cannot transfer tacit knowledge, create a qualified machine setter, restore a customer relationship or give a colleague the permissions they never had. Treating insurance as the whole answer confuses financial relief with operational recovery.

The stronger approach is layered: understand the dependency, reduce it where practical, prepare continuity arrangements and then consider whether financial protection is appropriate for the residual exposure.

Warning signs

• Only one person can set, programme, maintain or troubleshoot a critical machine or process.
• Work instructions describe the normal routine but not the exceptions that experienced staff manage by judgement.
• A customer or supplier relationship exists mainly through one individual’s phone, inbox or personal credibility.
• Essential passwords, permissions, files or contact histories are not accessible through controlled company systems.
• Holiday cover works only because the key person remains available by phone.
• A named deputy exists, but that person has never completed the work independently.
• Retirement, absence or resignation is discussed as an HR matter without testing the effect on production, quality, cash or customers.
• A recent acquisition still depends on the former owner or a small group of legacy employees.
• Training records show attendance but not proven competence.

Six questions for the board

1. Which one person would create the greatest operational or financial impact if unavailable for 30 days?
2. What production, quality, customer, supplier, system or compliance activity depends on that person?
3. Who can perform each critical task independently today—not after training begins?
4. Are the necessary knowledge, permissions, records, contacts and decision rights accessible through controlled company systems?
5. What would the first day, first week and first month of absence cost in output, margin, cash and customer confidence?
6. Have we tested the cover arrangement in practice, including a difficult or unusual situation rather than only the routine process?

Five actions to take now

1. Map the dependency. Choose the person with the highest potential impact and list every process, machine, customer, system, approval and decision that relies on them.
2. Separate the controls. Identify which dependencies require documentation, shared access, delegated authority, customer introduction, formal training or practical rehearsal.
3. Prove the deputy. Ask the nominated cover person to complete the critical activity independently while the specialist observes rather than intervenes.
4. Run a 30-day scenario. Work through the operational and financial effect at day one, day seven and day thirty, including quality, delivery, cash and stakeholder communication.
5. Set a transfer deadline. Give each unresolved dependency an owner, action and review date. Knowledge transfer without a date is an intention, not a control.

The InduX view

The aim is not to make experienced people feel replaceable. It is to stop their value becoming a single point of failure.

Strong manufacturers preserve specialist knowledge, give capable people room to progress and make continuity part of everyday work. That protects the business while reducing the burden on the individual who is otherwise never fully able to switch off.

Start with one person and one 30-day scenario. Then follow the dependency through the six pillars: what stops operationally, which customers or suppliers are affected, what it costs, which systems and approvals are involved, how the workforce responds and whether the business could defend the decisions it makes.

If the answer exists only in someone’s head, the dependency still exists.

---

Sources and further reading

• Shape of British Industry 2026 — Make UK
• Industrial Strategy Skills Commission Report 2025 — Make UK
• Business impact analysis guide — City of York Council

This article provides general information and prompts for management discussion. It is not legal, financial, insurance, employment or other professional advice. Appropriate specialist advice should be obtained for the circumstances of the business.