Repeat.
Think in loops.
The wing that carries the other three. Not a look back but a method: every completed loop has three beats, what I did, what it taught me and what the next loop changed. The newest hangs at the front.
- 2026 The learning center: 16 accounts, 0 bookings I built a learning center with daily courses to teach people a piece of what I learn about marketing, data and AI every day. In the end it showed 16 accounts, all test or bot accounts, and 0 bookings. As lessons go, a learning center could hardly ask for a clearer one. On 29 October 2026 I shut it down. How long it ran and what it cost is still open and will be added here.
- 2018/19 Business and tax coaching: a holding and an operating company I got business and tax coaching. It led to a holding and an operating GmbH.
- Year to follow A clear niche, the first client on Upwork I set up a profile on Upwork with a clear offer for a clear niche: tracking and data. About two weeks after the profile was approved, I had the first client.
- Year to follow Start-ups: the ZenMate exit, employee shares at zero I worked in start-ups and held equity there. At the ZenMate exit the employee shares went to zero. At least the maths was quick.
- Year to follow Corporate: Telekom, ended early I was in corporate, at Telekom, and ended that time early.
01
Think in loops
I work with two statements. The first is Learn. Think. Do. Repeat. This site is organised by it. The second explains why the fourth word is not a look back but the most important one: think in loops.
Thinking in loops means: never think in one-time shots, always in long-term, strategic loops. A one-time shot is a project, an attempt, an idea you try once and then tick off, whatever the outcome. A loop is the same, plus a step most people leave out: the review. What happened? What did it teach me? What do I do differently on the next run? And then there is a next run.
I believe that over ten years someone who thinks in loops beats someone who thinks in one-time shots, and not by a little. The reason is simple. A loop has a review step, so the second run is better than the first, the third better than the second. A one-time shot never learns. Whoever fires ten one-time shots ends up with ten attempts at the level of the first. Whoever runs a loop ten times is someone else by the tenth.
For you, if you want out and do not yet know what to sell, this may be the most important idea on this site. Your first offer will not be the right one. It does not have to be. It only has to be the first run of a loop you review honestly.
02
What went wrong
What goes wrong when you think in one-time shots shows in my entries below. The learning center was such a shot: built, launched, and only then asked whether anyone wanted it.
The opposite mistake is staying in something without ever reviewing it. That is a one-time shot too, only a very long one.
The entries below are my completed loops, newest first. Each has three beats: what I did, what the loop taught me and what the next loop changed. Where a year is missing, it stays open until I add it.
03
The loops
2026
The learning center: 16 accounts, 0 bookings
What I didI built a learning center with daily courses to teach people a piece of what I learn about marketing, data and AI every day. In the end it showed 16 accounts, all test or bot accounts, and 0 bookings. As lessons go, a learning center could hardly ask for a clearer one. On 29 October 2026 I shut it down. How long it ran and what it cost is still open and will be added here.
What the loop taught meNobody wakes up in the morning thinking: I need a course today. You wake up thinking: why are no enquiries coming in? Knowledge sells badly when it sits in every chat window, and nobody buys a solution to a problem they do not feel. I had not kept my own first rule, research demand first.
What the next loop changedThis site is content only again. That loop is running right now, and what it yields goes in the next entry once it is complete.
2018/2019
Business and tax coaching: a holding and an operating company
What I didI got business and tax coaching. It led to a holding and an operating GmbH.
What the loop taught meCoaching is worth something when it leads to a decision you would not have made without it.
What the next loop changedSince then I have control over my own time and my own money. I decide what I work on, and I can run experiments like the ones on this site.
A clear niche, the first client on Upwork
What I didI set up a profile on Upwork with a clear offer for a clear niche: tracking and data. About two weeks after the profile was approved, I had the first client.
What the loop taught meWhoever solves one thing for one group gets found. I was not the best at everything but the right person for one thing.
What the next loop changedSince then the niche is the rule I build every offer with. Today it sits at the top of how I work, under Do.
Start-ups: the ZenMate exit, employee shares at zero
What I didI worked in start-ups and held equity there. At the ZenMate exit the employee shares went to zero. At least the maths was quick.
What the loop taught meEquity is a promise, not money. You should not mistake a promise for a salary.
What the next loop changedSince then I look very soberly at offers that want to pay in equity instead of money, and I started working on my own account.
Corporate: Telekom, ended early
What I didI was in corporate, at Telekom, and ended that time early.
What the loop taught meA safe place is no reason to stay if you do not learn there what you want to learn.
What the next loop changedThe next loop was start-ups, with everything the entry about them tells.
Nobody wakes up in the morning thinking: I need a course today.
04
What I know now
What I know now from these loops: none of them was a success or a failure on its own. Leaving Telekom early, shares that went to zero, a learning center without sign-ups: as single events you would call them failures. Read as loops, each one made the next one better.
I also know that the review step does not happen by itself. It has to be planned, with a date, or it falls away. That is why every page here has a changelog and every thesis a date. That is the review step, built into the form.
And I know that loops need time. A loop you break off after three weeks is a one-time shot with a good conscience. That is why I test a service for three months before I review it. Long enough to see a pattern, short enough to start the next loop in time.
05
Open question
What is open for me is how long a loop should be. Three months for a service, a year for a structure, several years for a career? I believe every kind of decision has its own length, but I do not yet have a rule I can give you with a clear conscience.
06
Changelog
First version. Nothing here is rewritten silently: every later change gets its own entry in this changelog.