An article for managers and general managers of companies that still — fortunately — employ people.
The question no business plan asks
There is a question that industrial plans have carefully avoided for the past decade, because the answer does not fit on a line of the P&L: if automation progressively eliminates paid work, who will buy the products that automated factories turn out?
This is not an ideological provocation, it is arithmetic. The world economy, as it is generally configured, works because a mass of people receives a wage and uses that wage to buy goods and services. A salary is not merely a cost to the producer: it is the demand for the producer next door. Remove wages from half the active population and you are left with perfect production lines serving a market that no longer exists.
The automation numbers are real and deserve to be looked at without denial: 542,000 industrial robots were installed in 2024, more than double the figure ten years earlier, bringing the worldwide operational stock to 4,664,000 units, up 9% in a single year (International Federation of Robotics). The phenomenon exists, it is growing, and it will not stop. Yet the debate on so-called technological unemployment has produced no consensus among economists on whether and how governments should intervene, and the most discussed proposals — such as taxing robots — remain heavily contested (working paper on robot taxation).
Translated for anyone running a company today: there is no systemic answer, so the answer has to be built inside the company. And the corporate answer is not to resist automation. It is to stop treating people as the cost line to be compressed while investing millions in machines that, without those very people, produce nothing worth selling.
No million-euro machine defends itself
Anyone who has worked in production knows it and sees it daily: a line worth several million returns exactly as much as the technician running it is worth. The same equipment, with two different operators, delivers different yields, different consumption, different downtime, and a different service life. The difference is not made by the invested capital. It is made by whoever touches it.
The cost of incompetence is documented, and it is enormous.
Downtime. According to Siemens' True Cost of Downtime report, unplanned downtime now costs the world's 500 largest companies 11% of annual turnover — nearly $1.5 trillion — with a lost hour worth an average of $39,000 in fast-moving consumer goods factories and more than $2 million in automotive (Siemens). In the most recent edition, the cost of an idle production line at a large automotive plant is put at $695 million a year, 1.5 times higher than five years earlier (Siemens, True Cost of Downtime 2024). A significant share of that downtime originates in maintenance not performed, performed badly, or performed too late — in other words, in competence.
Human error inside the process. A field study of electronics assembly attributed 17.69% of product defects to human error, with error frequency rising in tasks involving more steps and under time pressure (research on human error and production rate). This is not indiscipline: it is the absence of structure and of consolidated skill.
The skills gap as a growth constraint. The Deloitte–Manufacturing Institute study estimates the skills gap could leave 2.1 million US manufacturing jobs unfilled by 2030, at a potential cost of $1 trillion in 2030 alone (Deloitte and The Manufacturing Institute). Machines are not what is missing. People capable of running them are.
The message for a general manager is blunt: you can buy your competitor's technology in six months with a wire transfer; you cannot buy their technicians' competence in six months at any price. The only genuinely non-replicable asset you own is the one that walks in on its own legs at seven in the morning.
What Ford actually teaches (and it isn't what you were told)
It is often repeated that in 1914 Henry Ford doubled wages to five dollars a day so his workers could afford the cars they built. Historically that account is contested: several analysts argue the raise was intended to stop the haemorrhage of staff, not to create demand (Adam Smith Institute).
The documented fact, however, is even more instructive for a manager: by late 1913 turnover at Ford had reached 380% — workers disliked assembly-line methods so intensely that the company had to replace nearly four times its headcount in a year — and the move to five dollars for an eight-hour day, against the previous $2.34 for nine hours, was the lever that made the model sustainable (The Henry Ford).
The lesson is therefore twofold, and it still holds. First: process innovation without investment in people destroys itself, because nobody sustains a system in which they are merely a cog. Second: when you pay people in a way that makes staying and improving visibly worthwhile, the numbers work out anyway, because the hidden cost of the revolving door is higher than you think. Gallup puts the cost of replacing one employee at one-half to two times their annual salary, and calls that a conservative estimate (Gallup).
The real context: skills that expire and people who struggle to read
Two data sets belong on the same table, because together they define the problem every GM actually faces.
Skills expire fast. The World Economic Forum's Future of Jobs Report 2025 finds that 39% of workers' core skills will change by 2030; that skill gaps are considered the single biggest barrier to business transformation by 63% of employers; and that 85% of surveyed employers plan to prioritise upskilling their workforce (World Economic Forum). On the employment side, the projection is 170 million new jobs created and 92 million displaced by 2030 (World Economic Forum).
Many adults start from a fragile base. This is the point almost no management article has the nerve to state, and the one every plant manager recognises immediately. The OECD's PIAAC Survey of Adult Skills shows that, on average across OECD countries, 18% of adults do not reach even the most basic levels of proficiency in any of the domains measured (OECD). In Italy the picture is harsher: 35% of adults aged 16 to 65 score at Level 1 or below in literacy against an OECD average of 26%, with mean scores below the OECD average in literacy, numeracy and adaptive problem solving (PIAAC country note), and the situation remained essentially flat between 2012 and 2023 (Inapp).
This, expressed in numbers, is what we commonly call functional illiteracy. And here lies the managerial knot: you do not convey the value of training to someone who struggles to interpret structured text through a memo, a motivational speech, or a poster about corporate values. The message lands only when it becomes concrete, visible, measurable and — without hypocrisy — monetisable. If the link between "I learn something new" and "I earn more" is immediate and verifiable by anyone, the message gets through without cultural mediation. If it stays abstract, it does not.
The paradigm shift: the skill belongs to the person too
The traditional model treats competence as a company asset: the company trains, the company owns, the company fears the employee will leave and take the investment along.
The model that works inverts the premise: competence is the individual's own capital, which the company helps to build and in return puts to use. This is not philanthropy, it is incentive design. The moment an employee perceives that what they learn stays theirs — certified, documented, portable — they stop experiencing training as an imposition and start seeking it out.
The data supports the choice. According to LinkedIn's workplace learning research, 94% of employees say they would stay longer at a company that invests in their growth, and learning opportunities rank as the number-one retention strategy cited by respondents (LinkedIn Workplace Learning Report). The apparent paradox — "if I train them, they leave" — is refuted: the ones who leave are mostly the ones you did not train.
There is also specific evidence on pay systems tied to skills. A study published in the Academy of Management Journal, using 37 months of time-series data with a comparison plant, found that skill-based pay was associated with 58% greater productivity, 16% lower labour cost per part, and an 82% scrap reduction relative to the comparison facility (Academy of Management Journal). Subsequent reviews confirm positive effects on employee attitudes and performance, with a high success rate but complex causal patterns (Center for Effective Organizations, USC) — meaning: it works, but the mechanism has to be designed properly.
And it is precisely how to design it properly that makes one concrete example worth examining.
An operating model: managing skills the way you manage production
There is a platform built exactly on this philosophy. GTMC-SKILLS is a GDPR-ready system for skills matrices, online exams, training paths and HR reporting: it maps competencies across roles, teams and sites; creates, schedules and verifies exams with a full audit trail; plans learning paths while tracking attendance and materials; exports readiness, gap and compliance reports; and engages teams through skills-based internal contests and rankings, with enterprise-grade data protection and access control (GTMC-SKILLS).
In shop-floor language: it does for competence what an MES does for production. It turns it into data rather than impressions. Who can do what, at what level, with what evidence, with what expiry date, and what is missing to cover tomorrow's shift.
But the genuinely interesting part — the one that should make every general manager stop and think — is the economic mechanism built on top of it.
How the skills bonus works
The principle is a single one: the company rewards those who learn more than their own department requires. Every employee has a personal page showing what they are accruing in real time, what is worth doing to increase it, and the history of what has already been paid out.
The bonus has two components, answering two different questions:
- The share rewards how much you know beyond what is required. The company allocates a budget for the period and divides it among everyone in proportion to each person's points.
- Valuable skills reward one specific competence the company wants to spread: whoever reaches it at the stated minimum level receives a dedicated amount.
The point calculation is deliberately simple, because it has to be understood by anyone:
- Each level is worth points — Basic 1, Autonomous 2, Expert 4, Instructor 7.
- The points of the level required by the department are subtracted from the points of the level held; the remainder is multiplied by the weight of the competence, from 1 to 5.
- If the department does not require that competence, the required level is zero and all its points count. This is why a new competence generally pays more than one extra level on a competence already required of you.
- Below the required level nothing is earned: first you reach what is due, then you start accruing.
A concrete example: if the department requires "Use of calipers" at Autonomous level (2 points) and the person is Expert (4 points), 2 points remain, multiplied by the weight; the same competence, in a department that does not require it, would yield 4.
The system is designed to be transparent even about its own uncertainty. The estimate for the current period updates every night, with a progress bar and the days remaining until closing, and it is explicitly declared an estimate rather than a guaranteed amount: because the share depends on everyone's points, if colleagues acquire competencies in the meantime your own share can fall. The figure becomes final only at period close, when HR calculates the bonus and submits it to the administrator for approval; only after final approval does it appear in the payment history and the employee receives an email with the details in their own language.
Two functions deserve particular attention, because they are what turn a bonus into a steering instrument:
- Highest-return moves, calculated on the individual case: which competencies, if acquired or raised to the next level, would increase the bonus most, each paired with the concrete action — the study material, the available exam, the open contest.
- The simulator, which lets anyone test any competence in the catalogue, not just those already held: choose a competence and a level, press Calculate, and see how much the period's bonus would rise. The list separates "Your competencies" from "To acquire", is searchable by code or description regardless of case and accents, marks valuable skills with a star showing the amount and the triggering level, and flags with a warning symbol those with prerequisites not yet held. Where a competence requires others first, the simulator does not display an unreachable number: it lists the steps in the order they must be tackled and shows the gain for the entire path. If prerequisites are misconfigured — forming a loop, for instance — it says so and shows no amount at all, rather than showing a wrong one.
The operational advice the system gives the employee has a clarity that many corporate incentive schemes never achieve: cover the mandatory requirement first, then target the competencies your department does not require because they carry all their points, at equal effort choose those with high weight and high levels, look at the valuable skills, and use the simulator before signing up for an exam.
Why this design is smart: five principles worth stealing
Beyond the product itself, the mechanism embodies five design choices that work as a general model.
1. Below the requirement, nothing is earned. The bonus does not pay for what is owed. This protects compliance — safety, mandatory qualifications, certifications — from becoming a bargaining chip, and puts coverage of requirements ahead of everything else.
2. It rewards versatility, not vertical accumulation. Valuing competencies the department does not require pushes precisely where the company's exposure is greatest: substitutability. An absence, a peak, an urgent order are handled with people who can do more than one thing. This is operational resilience bought with a bonus rather than with overtime and temp agencies.
3. Fixed budget divided into shares: the cost is predictable. This is the feature that makes the scheme sellable to a CFO. The company decides how much to spend in the period; the split rewards relative merit without opening an unlimited tap. In exchange you must accept — and communicate honestly, as the system does — that an individual share depends on others too.
4. Valuable skills are the steering wheel. The general share rewards learning as such; the dedicated amount directs effort toward what matters now: a new technology, a certification a customer demands, a competence held by a single person in the entire plant. It is a strategic lever management can move quarter by quarter.
5. The simulator removes suspicion. In opaque incentive schemes the typical reaction is "it's decided upstairs anyway". Being able to calculate in advance what an exam is worth turns the bonus from a lottery into a rational economic decision — and this works particularly well with the people least inclined to trust corporate promises, who are often the ones with the most to gain.
What the company gets, in P&L terms
- Less downtime and less damage to equipment, because autonomous maintenance and correct operation depend on certified competence levels rather than on who happens to be on shift.
- Fewer defects and less rework, in the order of magnitude suggested by the skill-based pay literature (Academy of Management Journal).
- Lower turnover and lower replacement cost, in a range Gallup places between one-half and two times annual salary per departure (Gallup).
- Peak coverage without external cost, thanks to incentivised versatility.
- Audits and certifications prepared in a day rather than a month, because the matrix, exams, attendance and certificates are already tracked with an audit trail and exportable as readiness, gap and compliance reports (GTMC-SKILLS).
- An edge in tenders and supplier qualifications, where documented proof of workforce competence is increasingly demanded.
The external calling card
There is a return that management-control systems never measure but customers register instantly. A company that takes a prospective client on a plant visit and shows them, alongside the machines, an up-to-date skills matrix, active training paths and a system that pays employees to learn, communicates something no brochure can replicate: here, process outcomes do not depend on luck.
It is the same principle as production transparency toward the customer: you stop declaring quality and start demonstrating it. With a further advantage in the labour market, where the competitive differential of a mid-sized manufacturer is almost never base pay — it is prospects.
The four objections you will hear, and the answers
"I train them and they leave for the competition." Companies that do not train lose people anyway, only later and on worse terms: learning opportunities rank as the leading stated retention lever (LinkedIn). And a single replacement costs up to twice annual salary (Gallup). Someone who leaves with one extra competence has cost you less than someone who stays without knowing how to use it.
"I have no budget for another bonus." The fixed-budget, shared-allocation model exists precisely for this: the amount for the period is decided in advance. Compare it with one hour of unplanned line stoppage in your sector (Siemens).
"My operators don't want to learn." Often what they don't want is to listen to speeches. With 18% of adults across OECD countries below basic proficiency in every domain measured (OECD), a system that displays on screen "this exam is worth X to you" communicates more effectively than any internal campaign. The issue is not motivation: it is the channel.
"It's just another across-the-board bonus." Quite the opposite: below the required level the bonus is zero, points are computed as a difference against departmental requirements, and weights steer effort where it is needed. It is the antithesis of an equal-for-everyone payout.
How to start in 90 days
- Map departmental requirements before mapping people. Which competence, at which minimum level, for which workstation. Without this, "above what is due" does not exist.
- Weight competencies from 1 to 5 using explicit criteria: risk, quality impact, internal scarcity, customer demand.
- Photograph the current state through assessments and exams, accepting that the first photograph will be embarrassing. That is normal, and it is the real baseline.
- Pick two or three valuable skills for the first period, tied to a real industrial objective that can be stated in one sentence.
- Open the simulator to everyone from day one, even before the first bonus is paid. Trust is built by showing the mechanism, not the results.
- Pay the first period on time, whatever the amount turns out to be. The credibility of the entire system rests on the first payment.
- Measure the control layer, not just the payout: requirement coverage by department, number of people per critical competence, time to cover an absence, scrap and downtime by department before and after.
Conclusion: back to the opening question
None of us will stop the robots and the drones, and there would be no point in trying. But a general manager can decide what role to assign people inside their own company: that of a variable cost to be squeezed until the last possible automation, or that of the one asset which, unlike every machine, does not depreciate but appreciates with use.
The first road has a terminus no business plan ever shows: perfectly efficient factories producing for consumers who no longer have income. The second builds companies where technology returns three times as much, because there are competent people making it return — and where those people stay, because what they learn is theirs as well.
Human capital is not the most important resource because it reads well in the letter to shareholders. It is the most important because it is the only one that multiplies when you treat it as such.
To explore the tool referenced here: the platform for skills matrices, exams, training paths and HR reporting described in this article is available to trial at GTMC-SKILLS.eu.
Sources: World Economic Forum Future of Jobs Report 2025; OECD Survey of Adult Skills (PIAAC) 2023 and Inapp country notes; Deloitte and The Manufacturing Institute; Siemens True Cost of Downtime; Gallup; LinkedIn Workplace Learning Report; International Federation of Robotics World Robotics 2025; Academy of Management Journal; Center for Effective Organizations, USC; The Henry Ford. All links appear inline.