Hiring · August 7, 2026 · 10 min read

Manufacturing hiring: widen the pool, keep the standard

Manufacturing hiring struggles less with talent than with a screen that rejects it. How skills evidence widens the pool, why speed matters, what tests miss.

← Part of The five pillars of hiring: what assessments measure

If you are the plant or operations manager who needs to seat a line, a maintenance crew, or a quality bench and keeps hearing there is nobody out there, this guide is written for you — and it opens with a mild heresy. The manufacturing hiring problem is real and structural, but a large part of it is self-inflicted at the point of screening. Deloitte and The Manufacturing Institute's 2024 workforce study projects 3.8 million manufacturing jobs opening between 2024 and 2033, of which roughly 1.9 million could go unfilled if the talent gap is not addressed. That is a genuine shortfall, driven by retirements and new capacity. But sitting on top of it is a screen — the degree line, the tidy CV, the continuous employment history — that rejects capable people the sector cannot afford to lose. This guide argues that the fastest lever most plants can pull is not a bigger advert but a better filter: one that measures what the job needs and stops discarding aptitude it never saw. It also stays honest about what none of that will fix.

How bad is the manufacturing labour shortage, really?

Structural and long-running, not a passing squeeze. The headline figure is Deloitte and The Manufacturing Institute's: 3.8 million jobs to fill across 2024 to 2033, with about half — 1.9 million — at risk of sitting empty. This is a decade-long gap fed by an ageing workforce leaving and new plants opening, and no single employer can advertise its way out of it.

The employers themselves keep confirming it, though the ranking moves around. In the National Association of Manufacturers' quarterly Manufacturers' Outlook Survey, over 65% of firms named attracting and retaining a quality workforce their single biggest challenge in the first quarter of 2024. By the fourth quarter of 2024 it had eased to fourth place, at roughly 55%, behind healthcare costs, the wider economy and trade uncertainty — not because hiring got easy, but because other worries grew louder. The same NAM Q4 2024 survey put the average share of manufacturing jobs sitting unfilled at 4.2%. Read those two facts together and the shape is clear: a persistent, structural gap that never quite leaves the top of the worry list, even when it is not at the very top. Averages like these also flatter the plants doing well and hide the ones stuck; the way a headline number conceals a handful of chronically open roles is the same effect mean time to hire unpacks for time-based metrics.

Why does the usual screen reject the people you need?

Because the filters most plants inherited — a degree requirement, a polished CV, an unbroken work history — are weak proxies for the actual work, and they fail hardest on exactly the candidates the sector is short of. The people who can hold a tolerance or diagnose a fault often learned it in ways that leave a thin paper trail: an apprenticeship, the armed forces, a family workshop, a job they lost when a plant closed.

Think about who a CV-first screen quietly discards. The career changer coming out of a trade with real mechanical aptitude but no manufacturing job titles to list. The machine operator who is brilliant on the floor and hopeless at writing about it. The person with a two-year gap that says nothing about their hands and everything about a caring responsibility or a redundancy. A screen built to sort office applications treats all of these as noise, filters them out before anyone assesses what they can do, and then the plant concludes there is nobody out there. There is somebody out there. The filter is standing between you and them. I once sat with a maintenance lead who told me his best hire in a decade had been rejected twice by the applicant-tracking system before someone met him at a jobs fair and watched him strip and rebuild a gearbox in twenty minutes — the paperwork had said no, the work said yes.

How does skills evidence widen the pool without lowering the bar?

By replacing a proxy with the thing itself. Instead of inferring capability from a degree or a job title, you measure it: aptitude, attention to detail, and situational judgement, assessed directly and on identical terms for every applicant. That simultaneously opens the door to non-traditional candidates and holds a firmer line than a CV ever did, because it tests what the role actually needs.

For most production and skilled-trades roles, three things carry the signal. A light cognitive skills test covers the reasoning and numeracy behind reading a drawing, working a gauge, or following a multi-step instruction. Attention to detail covers spec adherence and quality — the difference between a part that ships and a part that comes back. And situational judgement covers the safety and priority calls a shift throws up: the machine behaving oddly, the instruction that conflicts with what the operator can see, the choice between hitting a number and stopping the line. None of these requires a degree to demonstrate, and none of them is easier to fake than to do. A candidate either reasons through the scenario or does not. That is the whole appeal: the test is blind to the background that a CV screen fixates on, and sharp about the capability that a CV screen never sees. It is the practical face of skills-based hiring — judge the work, not the paperwork around it.

Why does speed matter when candidates are scarce?

Because a genuinely skilled candidate in a shortage market is not choosing between your offer and unemployment — they are choosing between your offer and two others. When the labour is scarce, the person is running the process, not you, and a slow pipeline hands them to whoever decided faster. The screening delay you can absorb in a buyer's market is a lost hire in this one.

This is where the widened pool and the clock meet. Picture the plant manager who does the CV reading in the twenty-minute lulls between shift handovers, because that is the only quiet the day offers. A maintenance requisition has drawn 180 applications over three weeks — the number is invented, because credible applications-per-opening data for the trades barely exists, so treat it as a device for the arithmetic and nothing more. Four minutes of honest reading each comes to twelve hours for the one role, and those handover gaps never add up to twelve uninterrupted hours. So the stack gets read in fragments across a fortnight, the strong applicant from day one is not phoned until week two, and by then the two rival plants that assessed everyone at once have already put an offer in front of them. The delay does not merely postpone the shortlist; it strips the best people out of it, because the candidates with options accept elsewhere first. A test that ranks the whole pool in one pass turns that twelve-hour read into a shortlist ready the same day — the only speed that competes when talent is scarce. To see where your own days actually leak, the time-to-hire calculator will show you the shape with your numbers.

The AI-era wrinkle: polished applications hide the gap both ways

Generative tools now write a clean, confident application for anyone who asks, and in manufacturing that cuts in two directions at once. It flatters a weak candidate whose paperwork used to expose them, and it does nothing at all for a strong one whose ability was always in their hands rather than their prose. Either way, the written application has lost most of the signal it once carried.

The temptation is to read applications harder, hunting for tells. That is a losing game — spotting AI-generated CVs is an arms race a busy manager will not win at four minutes a document. The better move is quieter: stop asking the application to carry weight it can no longer bear. You cannot prompt your way through a gauge reading or a fault-diagnosis scenario. A short, role-shaped test is one of the few filters that a polished application cannot game, precisely because it measures the doing rather than the describing. That is also why it rescues the candidate the polish would have flattered you past — the one whose competence never made it onto the page. This is a lighter touch than the AI question dominating some sectors, but the direction is the same: move the evidence from the document to the work.

What a test will not fix in manufacturing hiring

A fair amount, and saying so is the point. An assessment widens and sharpens the pool and lets you move faster on the strong candidates in it. It does not touch the three forces that most often decide whether a manufacturing role gets filled and stays filled: pay, shift patterns, and the sector's standing with younger workers. Those are real, and no screening tool reaches them.

• Pay — if the wage sits below what the plant with the newer canteen is offering for the same shift, a better screen just delivers rejections sooner. Testing sharpens who you chase; it does not make an uncompetitive offer competitive.
• Shifts and conditions — rotating nights, hot lines, long commutes to industrial parks. A test cannot soften any of that, and for many candidates it is the deciding factor, not the assessment.
• The image problem — manufacturing still fights a perception among younger workers that the work is dead-end or dirty. Widening your screen helps you catch the people who apply; it does not fix why fewer of them apply in the first place.
• Retention — a good hire who leaves in six months over the rota is still a loss. Better fit at the front helps; it does not compensate for the reasons people walk.

So be clear about the division of labour. The wage bands and the shift design usually need executive air cover and time to change. The screen is the one lever a plant's hiring team can usually redesign this quarter, on its own authority, and it happens to be the one that is actively working against you right now. Fixing it does not solve manufacturing's workforce problem — that is a decade-long, sector-wide story the Deloitte figures capture. It removes the part of the problem you are inflicting on yourself, and it does so faster and cheaper than anything else on the list. The broader case that a strong hire justifies the effort, and a weak one costs far more than the salary, sits in the cost of a bad hire; and the general playbook for filling many seats against a hard clock — including the mechanics of learning what a role needs by watching people do it — is in the high-volume hiring guide.

The way a role-shaped assessment does this in practice is unglamorous and quick: a test is built from the actual job — this line's tasks, this bench's tolerances, this shift's safety calls — reviewed by a person before it goes out, and run across every applicant on the same rubric. Historically that kind of tailored test was priced out of reach for a single hard-to-fill role, which is why plants settled for generic aptitude batteries that measured something adjacent to the work. That constraint is what has changed. To see a manufacturing screen built from one of your own job descriptions and scored across a whole pool, book a demo.

Somewhere on the far side of your applicant-tracking system there is a maintenance hand who can strip and rebuild a gearbox in twenty minutes, rejected because the paperwork said no. Measure the work — the aptitude, the attention, the judgement on the line — and that hand walks back through the door the screen once shut, along with the rest of the pool you thought was empty.