My last working day at the agency was 28 February 2025. Fifteen years — from intern through consultant to leading a seventeen-person engineering team. The decision had been made in late 2024; the start as a solo consultant came a few months later. This is not a settling of scores. It is the inventory I did for myself before starting out: what do I take with me, what do I leave behind — and why.
Why I left
The short version: in my role, hardly a day could be planned. Jumping into a project at short notice, preparing a pitch in the afternoon, learning a technology overnight and representing it as an expert in front of a client the next day — that was the norm, not the exception.
Technically I always managed. What wore me down over time was something else: my own standard never to claim anything in front of a client that I had not understood myself. Keeping that standard up under constant time pressure costs more energy than is visible from the outside. With a small daughter at home, the wish for more predictability simply became more important — and at some point it was clear I could honour it better in a different setup.
I am not telling this to talk the agency down — the years there made me into the person who now works independently. I am telling it because the reason behind it is the point at which a text like this becomes credible, or does not.
What I take with me
Three things from agency life proved themselves so well that I kept them unchanged.
A serious estimation method. At the agency, nothing was estimated by gut feeling — it was structured: effort broken down, assumptions made explicit, risks named. The real value is the side effect: a clean estimate is already half the project plan. Anyone who estimates this way has run the project through their head once beforehand.
Structured work and project-management craft. Out of pitches and delivered projects, a repertoire grew over the years that is now basic equipment: how to cut scope, how to make dependencies visible, how to run an escalation before it turns into a crisis. That is not certificate knowledge — it only emerges in real projects.
Always something tangible on hand. Before every first client conversation at the agency, I brought a clear structure of how I would approach the problem — no glossy slide battle, but something prepared. I kept that habit. It forces me to think before the conversation, not first during the meeting.
What I now do deliberately differently
Just as important is the other list — the things I often chafed against and now solve differently.
I choose the technology stack freely. As part of an agency, you also recommend what partner contracts, existing licences or the currently available skills suggest. Solo, I propose what genuinely makes sense for the problem — and I also say when the expensive enterprise solution is the wrong one for a mid-sized company. That degree of freedom has become one of the most important parts of my work.
Status meetings are lean. Reporting routines can take on a life of their own until the reporting costs more time than the work it reports on. My updates today are short, written, decision-oriented. That is not only more effective — it is simply cheaper for the client.
Availability is a decision, not an expectation. Reachability after hours still exists — but deliberately, agreed, and transparently billed. The difference is not the number of hours. It is that today I know what I invest them in, and for whom.
What surprised me
One thing I had misjudged: acquisition. I had expected tough cold-calling. In fact, a rather casually written LinkedIn post unexpectedly took off — and the first lead from it became the first client directly. More surprising than the lead was how smoothly the collaboration ran from the first minute. In hindsight that is not luck: anyone who has spent fifteen years saying what they actually think attracts the people looking for exactly that.
Who I do this for
My work fits best with mid-sized companies and smaller IT departments that actually want to change things — not just talk about it. What they get can be named concretely: a single point of contact they know and who answers. Solution proposals ad hoc in the meeting, not in a loop two weeks later. Few to no intermediate layers. And assessments from fifteen years of real end-customer projects — not from a slide deck.
This is not a promise of "more pragmatism". It is a different build of collaboration: less apparatus, more direct responsibility.
If that sounds like the way you want to work — Technical Consulting is the area where I move the most: from architecture decisions through system integration to performance. The simplest first step is a non-binding 30-minute call: you describe what is stuck, and we clarify whether and how I can help.