Six management lessons, from a field power station to Industry 4.0

I led my first team in 1987, at twenty, in the French Army Signal Corps: a field power station. I left my last salaried position in March 2026, after six years leading the digital transformation of an automotive-safety electronics plant. Thirty-nine years, a dozen roles — and six lessons I still repeat to myself today, now that I help other organisations transform.

Lesson 1 — Obedience is not the point

French Army, Signal Corps — 1987-1992

People picture the army as a school of obedience. I was taught the opposite. Three things, in this order.

First, that law and human dignity come before orders, and that courage and honour can consist in disobeying. A hierarchy that explicitly teaches you when not to follow it teaches you, in the same breath, that obedience was never the goal — only a means, and not the most reliable one.

Second, that nothing is achieved alone: cooperating with every team engaged on the mission, including those outside your chain of command, was not a soft skill. It was the mission.

Third, that you surpass yourself for something bigger than you. Not for a chief, not for a stripe: for what was bigger than all of us put together.

Rank gives you the right to give an order. It never gives anyone the desire to carry it out properly at three in the morning, in the rain, when nobody is watching. Forty years on, what I have kept fits in one sentence: statutory authority is a credit you burn; shared intent is an income that renews itself.

Simon Sinek (Start with Why) is right about the why, and he made it accessible to everyone. But the collective work edited by Emmanuelle Duez, Ibrahima Fall and Édouard Dequeker, Le commandement ne dort jamais (« Command never sleeps », in French), is the book that best captures this mindset — precisely because it confronts the ruggedness of military command with business leaders, without complacency or folklore.

Lesson 2 — Anything can be learned, but only practice makes it stick

Retraining, IB Formation, Freecom — 1994-1999

After the army I went back to school, starting from the bottom: office-IT technician training, then systems and networks. In under three years I went from the classroom bench to systems engineer and certified trainer for the leading vendors of the day — Novell and Microsoft. I trained those vendors’ certified partners.

First lesson: you can learn anything, at any age, all the way to expertise. It is not a matter of talent; it is a matter of method and invested time. Stanislas Dehaene (How We Learn) demonstrated it: attention, active engagement, error feedback, consolidation. All four work as well at thirty as at seven.

Second lesson, a harder one: teaching a technology gives you the illusion of mastering it. I explained beautifully architectures I had never had to rescue at midnight, in production, at a customer losing money by the minute. That is exactly why I left training for the field — performance audits, migrations, server rescues at Canon, Nestlé or Hachette.

Good books accelerate. They anchor nothing. Ever since, I distrust my own competence until I have put it to the test of reality — and that is as true of management as of LDAP directories.

Lesson 3 — Excellence is proven, not declared

Bristol-Myers Squibb & UPSA — 1999-2008

The pharmaceutical industry taught me what a culture of excellence is — not as a slogan, as a system. Five things, which have never stopped serving me since.

First, meaning bears repeating. The BMS pledge — “Our mission is to extend and enhance human life” — and the strategic vision that unfolded it into ten therapeutic areas were hammered home until everyone, from the lab to IT support, could say what their work contributed to. That is where I understood that vision is not a communication exercise: it is a management tool, and it only works repeated.

Second, excellence is not reaching the state of the art. It is doing better. And above all — this is the lesson pharma teaches like no other industry — excellence is proven. Qualification, validation: everything that touched the medicine, including our software and our hardware, had to demonstrate its efficiency, its reliability and its repeatability. Not promise them — demonstrate them. Ever since, whenever someone tells me a system or an organisation “works”, my first question has stayed the same: what proves it?

Fourth principle: the state of the art applies everywhere, not only where the regulator is looking. That is the spirit in which we implemented ITIL good practice for services and PMI for projects — the same level of demand for IT as for the medicine.

Finally, deep organisational change is managed deliberately: built into projects from day one, never bolted on, and built with the users and the people concerned in the field — not merely decided at headquarters. The Field Technology Services project proved it to me at scale: defining and deploying a standard, ITIL-based support model for the sales forces of 26 EMEA countries — 5,500 medical reps. Users who by definition were far from headquarters, never in our offices, and for whom IT was a road tool, not a topic. A support model designed without them would have been a support model worked around by them.

Lesson 4 — “The SLA is met” never consoled anyone

Carrier Transicold Europe, IT Service Delivery & Supplier Relationship Manager — 2008-2014

Arriving at Carrier Transicold: 1,200 users in 17 countries, largely outsourced support, SLAs everywhere. Green indicators — and unhappy users. The dashboard measured my contractual peace of mind, not their experience. W. Edwards Deming (Out of the Crisis) had warned us: running an organisation on visible figures alone means running it on what is easy to count rather than on what counts.

I answered with three undertakings. First, creating the service catalogue: describing what IT delivers in the words of the people who use it — named services with their commitments, not availability rates. You cannot judge a service you cannot name, and an SLA only means something against a promise the user understands. Second, leading ACE for the European IT cell — ACE, for Achieving Competitive Excellence, the lean operating system of UTC, the group Carrier then belonged to¹: continuous improvement pulled by the voice of the customer, not by internal metrics. Applying to IT the discipline the group applied to its factories.

That left the third actor: the suppliers, who delivered most of the support. I applied the same logic to them with VIRTUE, a supplier balanced scorecard — the idea, and the acronym, came from Gilles Hueber, my boss at the time, who entrusted me with making it real: weighted criteria across six pillars — Value for money, Innovation, Relationship, Time, Utility & Warranty, Ethics & Environment. Meeting the SLA is one criterion out of thirty-three. End-user satisfaction, measured by survey, is another — with equal standing. Assessing your providers on that, and co-building the improvement plan with them at every quarterly review, changes the nature of the conversation: you no longer manage a contract, you manage a relationship. That method served me well enough that I made it, fourteen years later, the foundation of SupPerf, the supplier-performance platform I am building today. And perhaps that is the lesson within the lesson: the best bosses entrust you with their best ideas.

An IT department that hits every target and loses its users’ trust has picked the wrong targets.

Lesson 5 — Managing without formal authority is the real job

Carrier Transicold Europe — 2014-2020

Then responsible for the EMEA application portfolio — ERP, CRM, EAM, with 85% of support outsourced — and later leading the front-office digital transformation. In none of those roles did I have any formal authority over the people my results depended on. What worked best comes down to two complementary constructions: a joint governance to align the top, key-user communities to listen to the ground.

Governance began, in late 2018, with a three-way alignment meeting — the EMEA Sales & Marketing, Services and IT directorates — whose first exercise set the tone: everyone openly assessed the strengths and weaknesses of our organisation as it stood, on the pillars inherited from ACE — quality, on-time delivery, relationship, responsiveness. Starting from a shared diagnosis rather than a target org chart means showing up with a common problem instead of a finished solution.

Then a monthly board where everything played out. Around the table, alongside the directorates, sat the Sales and Services Business Relationship Managers — the business-to-IT bridging role codified by the BRM Institute — carrying their directorate’s voice into IT, and IT’s voice back into their directorate. The mission that brought us together was not decreed there: it was debated, meeting after meeting — it changed three times in eighteen months before settling. Trade-offs were visible: every request went through a value/execution matrix unapologetic down to its vocabulary — proceed, investigate, consider, kill — decided by business and IT together, not in a corridor. And in early 2020 we did collectively the most useful exercise of all: listing the year’s missed or abandoned objectives, and why. The projects that had worked all shared the same ingredients — time genuinely dedicated by managers, the right skills on the business, IT and supplier sides, firm deadlines, clear objectives and real change management. The others had none of it. Saying that together, out loud, without hunting for a culprit: that is publicly accepting you were wrong — and putting all our energy into finding solutions.

Governance aligned the directorates; the ground still had to be heard. That was the role of the key-user communities we created: in every country, volunteer users, champions of the solution on their territory, with genuinely dedicated time — negotiated with their manager, written into their objectives — to support their colleagues in their own language, test evolutions with business knowledge, and send ideas and irritants back up. In exchange they received what matters: recognition, real influence over the roadmap and priorities, and a community of peers. Governance decided top-down; key users fed bottom-up. Either one without the other falls short: governance without the ground arbitrates in a vacuum; the ground without governance exhausts itself in orphan initiatives.

The result goes beyond IT, and it is the point I care most about: this organisation accelerated the transformation of the business itself — servitization. Selling availability and service, no longer just machines: maintenance contracts, telematics, service-offer configuration, subscription billing. It is that very shift that had decided me, as early as 2013, to go back to school (Neoma) for a specialised master’s in service marketing — an echo of my second lesson: when the ground changes, you learn again.

Yves Morieux (BCG, Six Simple Rules) recommends reinforcing the “integrators” — the people who bridge departments — and rewarding cooperation, rather than piling up structures. By that yardstick our record is honestly mixed: we did add a governance body. But what actually produced the cooperation was the rest — bridges recognised and valued, Business Relationship Managers and key users alike, all the silo-breakers, and trade-offs made visible. Managing without formal authority is not a constraint: it is a demanding teacher — it forbids every shortcut.

Lesson 6 — Knowing how is not enough; you must know how to make it understood

Veoneer France, 4.0 transformation — 2020-2026

An automotive-safety electronics plant, where traceability is the heart of the reactor. My team lived by a simple, demanding vision: deliver the right information, of the right quality, to the right person or system, at the right time and in the right format — enabling the best possible decision. Repeated until understood by all: the pharma lesson, twenty years on.

On paper, those six years were the most technical of my career. Eliminating obsolescence — more than 10% of temporary production-line stoppages avoided —, pushing data availability beyond 99.99%, turning a twenty-two-year-old text-mode ERP into modern mobile warehouse apps, moving from batch to event-driven architecture — information by the second rather than by the day — all the way to our first artificial-intelligence pilot. Agile, open source, and with apprentices trained in-house.

If that team dared so much — dismantling a twenty-two-year-old ERP, betting on open source, trying AI before everyone else — it is because it lived in what Amy Edmondson named psychological safety: the freedom to propose, to contradict and to be wrong without paying for it. The meta-analysis by Frazier et al. — 136 samples, more than 22,000 individuals — makes it one of the most robust predictors of team performance². It is not decreed in a meeting; it is demonstrated the day someone contradicts you. My favourite proof is a simple fact: employees and apprentices achieving things they believed unrealistic, even impossible for a small team like ours. I should say the model that inspired me is Yves Caseau’s, CDIO of Michelin (The Lean Approach to Digital Transformation): aim for the state of the art with small autonomous teams, from customer to code.

And yet, honestly: we achieved digital adaptation more than genuine transformation. It was not the know-how that was missing — governance, key users, change management: Carrier had taught me all of it. That is where I received the most expensive lesson of my career: knowing how is not enough; you must know how to make it understood. At Carrier, the management consensus on the need to change existed — not always comfortable, but real, and backed by the group. Here, it never did: as if being the best exempted you from changing in order to remain the best. I never believed that analysis. But I did not manage — or was not able — to get my vision shared.

Was it the legitimacy of an IT manager speaking about business transformation? The perceived risk of touching methods that, on the surface, worked? An approach — breaking silos, building cross-functional agile teams — too far from the house culture? Probably a bit of all three. A digital transformation remains a human project disguised as a technical one, and the sixty-year review of quantitative studies by Oreg, Vakola and Armenakis says precisely that: genuine participation, quality of information, trust in management — the how weighs more than the what³. The best tool deployed without the people loses to an average tool deployed with them. And the best know-how without consensus remains know-how without ground.

That is precisely why I made it my profession. And true to my second lesson, I keep practising what I advise: I learn AI by building with it, and I develop myself the SaaS platform born of my supplier years. The sixth decade is handled like the first — in the field.

What has not changed

The technologies I administered in 1999 are gone. NetWare, the NT servers, the 3,000 PCs to remote-deploy: none of it exists any more. The six lessons, though, have held since the field power station.

They even draw a circle: my last battle — getting an intent shared — was already the first one, the young team leader’s of 1987. Only the scale has changed.

Perhaps that is the real surprise of a career in IT: management goes obsolete far more slowly than technology.

And I would be leaving out the essential if I did not say it: none of this — not at BMS, not at Carrier, not at Veoneer — would have been possible without genuinely committed, volunteer teams, who refreshed their skills at the speed of technologies and good practice. Without them, nothing. The lessons are mine; the successes are theirs.

A few books from my library

• Le commandement ne dort jamais — ed. E. Duez, I. Fall & É. Dequeker · obedience is not the point
• Start with Why — Simon Sinek · intent before orders
• How We Learn — Stanislas Dehaene · attention, engagement, error, consolidation
• Simply Managing — Henry Mintzberg · what managers really do
• The No Asshole Rule — Robert Sutton · target the behaviour, not the person
• Out of the Crisis — W. Edwards Deming · measure what counts
• Six Simple Rules — Yves Morieux · reinforce the bridge-builders
• Drive — Daniel Pink · autonomy, mastery, purpose
• The Lean Approach to Digital Transformation — Yves Caseau · customer to code, in small teams