Business · Technology and Product · Organisation

Running the business. Building what comes next.

From plant projects to vertical software: I have brought series products into manufacturing, led a business unit with its own P&L statement, and today I am responsible for international software development.

This site is a portfolio: selected work, the domains I operate in, and the convictions that came out of both.

Sectors
Vertical software for energy & water utilities Connected MedTech — patient monitoring Drives, actuators & control electronics Consumer goods (FMCG) Machine vision & inline inspection Speciality & security paper
International reach
Today: development units under my lead in Bulgaria, North Macedonia, Romania, Pakistan and India Before that: manufacturing and development built up and led in China — series ramp-ups, qualification, customer acceptance on site Plant projects across Europe, China, Russia, India and Pakistan — commissioning and acceptance at the customer
Scope

Scope of responsibility

Six figures from four companies. Each sits with its own station and period — they do not add up.

~190
people, led through five directors
Software organisation · today
6
countries with development units
Software organisation · today
2
platforms running in parallel, one being replaced
Software organisation · today
4
product lines built from zero
Business unit · 2018–2021
20+
plants and 200+ lines in the serialisation programme
Operator side · 2015–2017
14
years on manufacturing lines, across four stations
Plants and series · 2003–2017
What I lead

Business, technology and organisation

Three fields, of equal weight. The customer connects them: they decide which offer holds, what can be delivered, and where investment continues.

Developing the business

Which offers deserve further investment? What does the customer actually need? And how does the transition stay economically viable? These questions connect portfolio work with responsibility for a business unit result.

Turning technology into products

My experience spans industrial systems and drive systems through to connected products and vertical software. What interests me is which technical decision improves an offer, and how that becomes a product you can deliver reliably.

Making the organisation capable

I have built units and led international development. Change has to land in the organisation: with clear ownership, priorities people can follow, and room for those who know the existing business.

From customer value to decision

Where technology becomes a business decision

Three places where a technical question is a commercial one.

What the customer actually signs off

In plant projects, acceptance at the customer site decides. Technical clarification, integration and commissioning therefore have to fit together from the start.

What can be delivered reliably

A series product needs more than a design that works. Target costs and qualified sources of supply help decide whether an offer can hold its own in the market.

Where investment continues

When a platform is replaced, the existing business keeps running. The sequence has to bring together customer needs, technical dependencies and available capacity.

Selected work

Proof, across industries

Four cases, each with its own job to do: a business unit, a software organisation, a plant project, a series product. Confidential clients are described by sector. Detail on request.

From device business to connected product business

Electronics & drives
Product & site P&L · business built from zero

Turned a pure device business into a connected product business. Four product lines cut and built from zero, each carrying its own share of the revenue:

  1. Sensing built into the bed, with analysis in an app of our own
  2. Voice control as an IoT module and retrofit adapter, on a cloud of our own
  3. Production maintenance with ERP integration for medical and furniture manufacturing
  4. An after-sales business in the US for connected beds

Together the four carry a recurring share of revenue, on multi-year agreements with manufacturers rather than one-off sales.

How it was built

What held the cost line was a platform: fixed electronics building blocks and drive elements reused across product families, so a new variant no longer meant new electronics. Software followed the same move — a white-label app platform with some fifteen to twenty customer-specific variants, native on iOS and Android, built, designed and shipped to both app stores by our own team.

On the voice-control programme the cloud was ours. I steered certification across FCC, CE, RED and Bluetooth SIG, executed through the development partner in Texas and the sister company in China. A quality lab of our own in Hamburg tested every build — nothing went into series without passing it.

Outcome
own P&L for the business unit, site built from zerocustomer NPS 25 → 40series ramp-up China · sales and after-sales US

VP Product & R&D — enterprise software

Vertical software
Two product lines · ~190 people · five directors

Product strategy, roadmap and portfolio for two software product lines inside a global software group, as a member of the management team reporting to the CEO. Alongside it a platform-replacement programme of €45–50M over five years that I own.

Portfolio governance across two platforms at once: phase-in of the new one, phase-out of the old one, investment priorities in between. The hardest question in a move like this is not the target architecture. It is the order in which things come out of the legacy platform.

To do that I rebuilt substantial parts of the organisation: product management and the roadmap process, product design, platform and DevOps, and international development units.

Alongside portfolio and delivery I am responsible for the budget, investment priorities and the commercial steering of the unit, together with the multi-year transformation programme.

Organisation, cadence and ways of working

Development in Germany, plus fully integrated development units under my lead in Bulgaria, North Macedonia and Romania as well as in Pakistan and India. The North Macedonia unit is one I identified and set up myself.

Moving grown platforms to cloud-native development meant building a platform team and a cloud/DevOps team inside my remit, staffed with new hires and people from the legacy side. That is where the product knowledge sits. The people who built the old system are the ones who have to replace it.

Release cadence in the new product now runs weekly instead of four times a year, at two weeks of lead time to the next increment. Alongside that, the AI-first transformation from strategy into delivery: spec-driven development, where the specification is the leading artefact and code is generated and verified from it with AI. And an LLM-agnostic architecture that turns switching AI provider into a configuration decision rather than a product rebuild.

Outcome
portfolio governance across two platforms4 releases a year → weeklycloud-native move, teams builtAI-first transformation

Camera-based inspection & inline marking

Plants & metrology
Project lead, plant engineering · order intake to final acceptance

International projects in camera-based vision, laser inspection and inline ink-jet marking, for FMCG and packaging and for makers of speciality and security paper. The scope was never the camera system alone but the complete installation: machine frames, unwind systems carrying the inspection heads, marking systems, control cabinets with operator stations, integrated into existing production lines and their safety concept.

The whole path there

From technical clarification with the customer through contract review and internal acceptance to leading the installation team on site, customer acceptance and operator training. Standing at the machine in a customer plant changes how you decide about products later.

Outcome
order intake to final acceptance in one handFMCG, packaging, speciality & security papercommissioning and acceptance on site

A complete drive system from concept to high-volume series

Electronics & drives
Product management, series production · development → series · target cost

Took a complete drive system into series production, almost entirely for the American market and to the standards that apply there: not only the drives, but wiring, RF hand controls and control electronics, plus certification, test runs, releases and quality assurance.

Mechanical engineering, electronics engineering and the software for that electronics sat under one roof. The remit covered target cost and programme margin.

Series ramp-up and proof of capability

Manufacturing and series ramp-up in China, built and steered together with suppliers, from supplier selection through first-article approvals to documentation. Added a stage-gate process from idea to series and a cross-site control process for international development requests.

The PPAP requirement came from furniture-industry customers, some of whom come out of automotive themselves — so the evidence was produced to automotive standards. Orders never simply arrived. First you proved the system could be built to price, quality and volume; then came the order. This is the base the connected business was later built on.

Outcome
drives and system components in high-volume seriescertification, releases and series assurance (APQP/PPAP)supplier management and series ramp-up in Chinaseries releases with customers in the US, Taiwan and China

Two further stations

Medical technology

Patient monitoring business unit

Led a business unit for connected patient monitoring across two German sites and China. Sensing, firmware and app logic in one hand, together with the documentation for approval. Business-unit result 8 → 14%.

Consumer goods manufacturing

Serialisation programme on the operator side

Two years on the other side of the table: traceability against a statutory deadline, across 20+ plants and 200+ lines, with the data chain from the production line into the ERP. Delivered without stopping production.

Measures

What I was measured by

Responsibility came with different measures depending on the task. These were mine.

Business unitresult and development of customer satisfaction
Building a sitebudget adherence and pace of build-up
Plant projectsacceptance at the customer site
Series productsunit cost and series release
Serialisation programmedelivery by the statutory deadline
Planningmonthly variance analysis against forecast

Experience across material worlds

Plants & metrology2003–2013
Electronics & drives2013–2021
Connected medical technology2021–2022
Vertical softwaresince 2023

The technical conditions differ considerably. What recurs is the task of bringing together customer requirements, commercial priorities and the organisation that has to deliver. In between lie two years on the operator side.

Motivation

Why this translation

Because that is where the interesting mistakes happen. Software teams underestimate what series production, tolerance and site acceptance actually mean. Hardware teams underestimate what release cycles and data models do to a business model. Both sides are right, and they talk past each other. I have worked on both sides of that misunderstanding, first on machines, then on electronics, then on software. I stayed because little releases as much value as a product that finally does both. This is not enthusiasm for technology. It is the appeal of getting a translation right that nobody else takes on.

Perspectives

Four convictions from practice

Each of them cost me time or money before it became clear. And each is visible in the work above.

01

Hardware companies rarely fail at software.

They fail at the business model. Recurring revenue demands different pricing logic, different service, different sales incentives and a different org design. Adding a cloud connection changes nothing on its own.

02

AI-first is an operating question, not a tooling question.

Handing out licences is the easy part. The difference appears where decision rights, quality gates and throughput measurement get redesigned. Otherwise it stays a collection of pilots.

03

Portfolio discipline beats feature enthusiasm.

The hardest margin gains almost never came from the next feature. They came from disciplined phase-outs, platforming and Design-to-Cost. From the willingness to stop things.

04

Regulation is a roadmap problem, not a legal one.

MDR, EU serialisation, energy-market rules, the Cyber Resilience Act. In three companies I have built portfolios against deadlines nobody can move. Delegate that to legal and you find out too late that the requirement sits in the design, not in the contract. The date belongs in the roadmap. The evidence duty belongs in engineering hygiene.

In print Modern software interaction concepts are user-centred Handelsblatt online · sponsored company feature · 2023
Notes

What I am working through

Short pieces from practice.

All notes
Stefan W. Kehrein
Get in touch

Start a conversation

Based in the Hamburg region, Germany · working internationally, in German and English (both to negotiation level).

MBA NORDAKADEMIE · Engineering degree in electrical engineering and information technology

I read every message myself and normally reply within two working days.