Somewhere in your organisation there is a task that exactly one person can perform, and that person is probably the real limit on everything the business promises it can deliver. Cross-training gets sold as development, a pleasant thing to offer people who want to grow. It arguably works better as fault removal. A team with one irreplaceable operator has already set its ceiling, and the ceiling is that operator's diary.

Start with the arithmetic of absence, because the arithmetic is not seriously in dispute. UK workers lose an estimated 148.8 million working days to sickness or injury every year, roughly 4.4 days each. The figure that should interest anyone running an operation sits further down the same release. Process, plant and machine operatives recorded the highest absence rate of any occupational group at 3.3 per cent, while managers, directors and senior officials recorded the lowest at 1.0 per cent. The people whose absence physically stops the line appear to be away more than three times as often as the people who write the continuity plan. That gap alone might reasonably have settled the argument years ago.

Coverage Is a Data Question Before It Is a Training One

Knowing who can cover what tends to be a records problem before it becomes a teaching problem. Most sites still keep the answer in a spreadsheet that one supervisor updates when they remember, which is a poor foundation for a decision you may have to make at six in the morning with a shift running short. Manufacturers that have moved past that often run coverage through manufacturing learning management software, where each operator's qualifications sit against each job and each expiry date, visible well before anything becomes urgent. The tool doesn't create the skill. It stops you being surprised by the absence of one.

Whatever holds the records, the harder question is what "qualified" is being taken to mean. Britain's safety regulator defines competence as the ability to perform to a recognised standard on a regular basis, built from skills, experience and knowledge together, and its guidance on training and competence is blunt that training alone does not get you there. Practice consolidates it. The regulator also expects refresher training for tasks that are infrequent, complex or safety critical, which is precisely the category most cross-training coverage falls into. A name on a matrix beside a job someone last touched two years ago is a record rather than a capability, and the two are easy to confuse when the spreadsheet only has room for a tick.

The Case Against Spreading Everyone Thin

Now the objection, which deserves more respect than cross-training advocates usually grant it. Spread skills too thinly and you may end up with a workforce that can do everything adequately and nothing exceptionally. Depth is real and expensive to rebuild once traded away. The setter who has run the same press for eleven years hears a bearing going before any sensor flags it, and no rotation programme will manufacture that ear in a fortnight. She also knows which of the three moulds warps in July, which is written down nowhere and probably never will be. Push coverage too hard and you risk swapping a rare failure mode for a permanent one, and the permanent one may well cost more across a year than the rare one costs in a week.

So the target is probably not universal interchangeability. A more defensible aim is a second and third person able to perform the task that genuinely stop everything, with honest acceptance that the rest will stay one deep. Two deep on the critical few. Single cover elsewhere, documented as such, so that nobody discovers it during an incident. Deciding which jobs belong in which category is a twenty-minute conversation most operations teams have never actually had.

Where Handovers Break, and Who Gets Blamed

A manager who has never thought carefully about how much control to keep when handing work over will often either hover until the learner loses confidence entirely, or vanish until something expensive breaks. Both outcomes get recorded afterwards as evidence that cross-training doesn't suit this environment. Neither looks like much more than unexamined management habit wearing a data point as a disguise.

Three things settle this before the first shift rather than after the first mistake. Agree what the learner may decide alone and what must come back to you, in words, not implication. Agree how the learner signals that they are out of their depth, because the failure mode is a proud person guessing rather than asking. And agree when the training wheels come off, with a date, because open-ended supervision curdles into permanent supervision. Guidance on balancing workloads and bringing team members on treats this as ordinary management rather than a special programme, which is roughly the right register for it.

What Financial Regulators Were Forced to Learn

Financial services were dragged into this discipline ahead of most other sectors, and not voluntarily. British regulators now expect firms to name the services that genuinely matter, set a tolerance for how long each can be interrupted, and demonstrate through testing that they can stay inside it. The supervisory framework that made this mandatory was written about payments and settlement rather than injection moulding, and the language is thoroughly bureaucratic.

The underlying question travels well though. Identify the activities that must continue, decide how much interruption is tolerable, and test whether the organisation can maintain them without relying on one individual.

The Window Most Organisations Walk Straight Past

The timing question answers itself the moment you look at when organisations actually attempt any of this. The usual trigger is a resignation letter, which is close to the worst available moment, because the person leaving now has a few weeks and very little reason to teach well. The genuinely useful window is the first month of a new starter's induction, when nobody yet expects them to be productive and a second qualification is simply more of what they're already doing.

Attaching a new hire to two lines rather than one costs almost nothing at that stage, and it sets an expectation before anyone has grown territorial about a machine. It costs a great deal eighteen months later, when they have become the only person who understands the labeller and the rota has reorganised itself around that fact without anyone deciding to. The same logic argues for a standing rule rather than a project. Any operator holding a job for two years with nobody else signed off on it is a flag, and it should surface automatically rather than relying on a supervisor's memory.

The Retention Argument That Lands on the Shop Floor

There is also a case here with nothing to do with resilience, and commercially it may be the stronger of the two. Workforce research in the sector suggests employees are 2.7 times less likely to leave within twelve months when they believe they can acquire skills that will matter to their future, set against a projection that as many as 1.9 million American manufacturing roles could sit unfilled by 2033. Cross-training appears on the risk register as continuity and reads on the shop floor as something closer to a promotion path. Workers rarely resent being taught a second job. They resent being stuck in the first one for a decade.

Which makes the direction of travel in the UK data awkward. In the a national survey of employers, manufacturing came out among the least likely sectors to have trained anyone at all, with 51 per cent of employers providing any training and only 48 per cent of staff receiving any, both near the bottom of the table. Training spend per employee across the UK fell to around £1,700, down from roughly £1,960 two years earlier and down almost 30 per cent in real terms since 2011. Meanwhile skilled trades carried the highest skill-shortage vacancy density of any occupational group at 48 per cent. Spending is falling fastest in the places where replacements are hardest to buy, which is a difficult position to defend on either resilience or recruitment grounds.

Your Most Capable Operator May Be Your Worst Teacher

Who does the teaching matters more than how the sessions are scheduled, and this is where good intentions tend to die. The most capable operator on a line is often among the least able to explain what they are doing, because a decade of practice has compressed the reasoning into instinct. Ask how they know and you may get a shrug and the word "feel", which is accurate and completely useless to the person standing next to them. Anyone asked to pass expertise on benefits from some working grasp of the difference between teaching, coaching and mentoring, plus tolerance for a learner who will be slower and worse for months.

Two practical habits help more than a course. Ask the expert to narrate a routine job out loud while doing it, and have someone write down every decision point they mention, because the gaps in that transcript are the tacit knowledge nobody has captured. Then have the learner run the job while the expert watches in silence, noting interventions rather than making them. Pair the wrong expert with the wrong learner and you tend to produce two frustrated people and no transferred skill, which then gets filed as proof that the idea was never going to work here.

Aviation dealt with a version of this decades ago. After a run of accidents in which junior crew failed to challenge a captain, the industry built training around the principle that authority should not sit unchallenged in one seat, and the history of that shift carries a warning worth reading twice. Researchers found that acceptance of the core concepts decayed over time, even among crews receiving recurrent training. Manufacturing borrowed the checklists and left most of the rest on the shelf.

A Coverage Audit You Can Run This Week

Take your three most consequential processes and write down every name qualified to run each without supervision. Any list with a single name on it is the thing to fix this quarter. A list carrying three names where two of them last performed the job in three years ago is fiction, and arguably a more dangerous fiction than the honest single name, because it will pass an audit and reassure a board. Currency matters as much as coverage. It appears to be the part that rots first, since almost nobody schedules refreshers for a skill the records already show as held, and the record itself gives no hint that anything has faded.

Then put a number on the thing you are protecting. Unscheduled downtime costs more than most plant managers assume, and one widely cited study priced the annual loss across the world's 500 largest companies at $1.4 trillion, around 11 per cent of their revenues, with an idle automotive line running to $2.3 million an hour. Your own figure will be smaller and considerably more useful in a budget meeting. Divide contribution per shift by hours, multiply by the realistic length of a coverage gap, and the number that belongs in the business case falls out of it.

Nobody should pretend the cost of building cover is zero. It means a second body at a station doing nothing billable for a fortnight, output dipping while a learner finds the rhythm, and a supervisor spending attention they do not have. Finance departments see that cost immediately and see the avoided cost never, because the breakdown that did not happen leaves no invoice and no photograph. That asymmetry is largely unfixable, which is probably why so many coverage plans exist on paper and nowhere else. An argued estimate beats an unargued silence.

The Bargain Nobody Puts in the Business Case

Cross-training also has a habit of exposing things an organisation would rather not see. Move a second person onto a job and you may discover the first has been running it on undocumented workarounds for years, compensating without ever mentioning it for a machine that has not been right since a repair years ago. That discovery is often worth the whole exercise, and it invariably arrives at an inconvenient moment.

None of which touches the hardest part, which is neither scheduling nor software. The person who cannot be replaced usually knows it. That knowledge has been worth something to them for years. The call at midnight. The reasonable confidence that a downturn will find someone else first. Cross-training asks that person to hand over the one thing that makes them structurally safe, and offers in exchange a certificate and an employer with slightly less reason to keep them. Most organisations never say this out loud. They approve the programme, watch the sessions get postponed a fourth time, and conclude the problem was diary pressure.

Naming the problem is the part you can actually do something about this week. Write the three lists. Find the single names. Then go to those people first, before the programme is announced to anyone else, and be straight about what is being asked and what they get for it, whether that is a pay band, a title, first refusal on the next supervisory role, or simply the freedom to take annual leave without a phone call. An expert who senses they are being dismantled will slow the process down for years, and will be entirely justified in doing so. An expert who has been offered something real for the knowledge will hand it over in a fortnight. The programme rarely fails on the training plan. It fails on the conversation nobody scheduled.

---

Final Thoughts

Cross-training is not about making everyone interchangeable. It is about identifying where dependence on one person creates unnecessary risk, then using communication, coaching, delegation and deliberate knowledge-sharing to build genuine depth.

---