Work on your business, not in it. The choice most founders keep postponing.

Most founders I meet work 70 hours a week. And 65 of those hours go into doing the job — not running the company.

That’s the trap Michael Gerber documented in The E-Myth Revisited in 1986. The book has sold over two million copies.

Forty years later, AI hasn’t changed it.

The chief everything officer

The pattern shows up in almost every company I see — funded startups, post-Series A scale-ups, SMBs doing several million in revenue. The cast changes, the script doesn’t.

The founder runs sales. The founder reads every contract. The founder fixes product bugs. The founder hires. The founder books meetings. The founder manages the board.

Officially, they’re called CEO. In practice, they’re chief everything officer. They founded a company, and they hired themselves as employee one, two, three, four, and five.

The trap isn’t that they do too many things. The trap is that they do too many operational things. And while they’re doing them, nobody is actually running the business.

What Gerber wrote in 1986 — and that hasn’t aged

The thesis of The E-Myth fits in one sentence: most entrepreneurs don’t run a business. They self-employ in the craft they know how to do, while believing they’ve built a company.

The mechanic who goes solo doesn’t create a mechanics business. They create a mechanic’s job, plus the worries of a boss. It’s harder, riskier, and the day they stop, the “business” stops too.

That logic applies fully to tech. The developer who launches a SaaS and spends their days coding hasn’t built a SaaS. They’ve built a senior developer’s job — more lonely, and worse paid than what they’d have at a real company. The salesperson who launches an agency and spends their days closing deals hasn’t built an agency. They’ve built a salesperson’s job, plus the responsibility of paying everyone else.

Forty years later, the diagnosis still holds. It’s probably the main reason 90% of companies that take off stay capped at a plateau they never get past.

AI hasn’t changed the underlying problem

In 2026, the dominant story is that AI will solve everything. A single founder can now code, sell, script their emails, generate their content, automate their workflows. Tooling has never been this accessible.

True. And irrelevant to the problem.

Better tools don’t change the fact that you decide where your energy goes. The same instinct that pushed the 1986 founder to run sales themselves pushes the 2026 founder to spend their days in Cursor fixing bugs a hired dev would have closed in two hours. The same instinct that pushed the 1986 CEO to validate every contract pushes the 2026 CEO to manually approve every LinkedIn campaign an AI agent generates.

AI saves time on execution. It doesn’t save time on the question that matters: should this task be done by me?

That question, no tool asks for you.

What I figured out — badly, and too late — at Sortlist

For the early years of Sortlist, I was the absolute chief everything officer. I read every client contract. I validated every hire. I ran sales demos. I weighed in on product topics I had no operational reason to touch. I was on every front at once.

In the moment, it felt like being an engaged founder. In reality, I wasn’t running a company — I was burning out stacking tasks others could have done better than me.

The turning point wasn’t a revelation. It was several quarters in a row where I saw the same problems coming back because nobody had the time to look at them from above. Not me. Not my co-founders. Not my managers. Everyone was in the weeds.

I took too long to understand that my job as CEO wasn’t to be the best executor. It was to build the system that produces results when I’m not in the room.

An organization — human, AI-augmented, or both — that executes without me. A strategy that guides trade-offs without me being consulted every time. A culture that transmits the right reflexes. A brand that produces pipeline independently of my personal contacts.

Without that, you don’t have a business. You have a badly-paid job, with payroll taxes and venture capital to repay.

The two profiles I meet over and over

These days, I see two types of entrepreneurs who, deep down, have exactly the same problem.

The first is in a golden cage. Their business runs. It’s profitable. They earn very well. But they can’t leave for three weeks without things starting to unravel. The day they unplug, revenue drops. The day they stop, the business stops. At sale time, their company is worth the EBITDA multiple of a standard SMB — not the multiple of a scalable business. Because no serious acquirer pays a premium for a job dressed up as a company.

The second is in the hamster wheel. They’re chasing time. They know they should get out of the operational weeds but they never manage to free up the weeks for it. Every quarter, they tell themselves “once we close this deal, I’ll get to it.” And every quarter, a new deal absorbs the weeks the previous one freed up.

Both profiles share the same thing: they don’t have a system. They have a workload.

Working on the business — what that actually means

Moving from in to on isn’t a slogan. It’s a calendar.

It means blocking, before anything else, two full days per month — not per quarter — where you do only these things:

Look at the company’s numbers (P&L, cash, pipeline, cohorts) cold.
Identify the current bottleneck and decide who owns it — not you, except in real exceptions.
Re-question the ICP, the segments, the product priorities. Not to change them every month. To check they still hold.
Take stock of the team: who’s underused, who’s under-equipped, who should be replaced.
Explicitly decide what you will stop doing in the next 30 days.

It’s uncomfortable because it produces nothing visible immediately. No deal signed, no feature shipped, no email sent. That’s precisely why most founders don’t do it.

And it’s precisely why the ones who do end up outrunning the others.

The right tools for this discipline

Four things that help, in order of impact:

An external partner. A demanding board, an investor who challenges you, a senior advisor, a coach, or a peer founder who asks you the questions your team won’t. That’s the thing I missed most at Sortlist during the years I was wrestling with this alone. It’s also the role I play today for the founders I work with through MetSaaS.

An AI coach for retrospection. You can use Claude or GPT as an external coach to ask yourself the right questions about the business. You give it your context, you ask it to challenge your strategy, your priorities, your time allocation. It’s not a replacement for human conversation — an LLM will never create alignment in a team — but it forces a level of structure most founders can’t impose on themselves alone.

One book per year on the operating system. The E-Myth Revisited (Gerber) remains the obvious one. Traction (Gino Wickman) for the EOS framework. Built to Sell (John Warrillow) if you’re aiming for an exit. Start with Why (Simon Sinek) to re-anchor direction when the daily grind drowns out the why. Read one a year and actually apply it.

A quarterly retreat. One day per quarter, out of the office, no calls, no Slack, with a notebook. You write down three questions: what do I stop doing in the next 90 days? Who picks it up? On what terms?

What this actually changes

Once you start really working on the business, four things shift.

You make decisions earlier. Pivots, restructurings, pricing breaks, team changes — they all happen further upstream, when they still cost little. Decisions made under pressure always cost more than the same decisions made calmly.

You attract better candidates. Strong people don’t want to join a founder who micromanages execution. They want to join someone who steers, who defines, who arbitrates — and leaves them the room to do their job.

You improve your valuation. Investors, acquirers, industrial partners — they all look at the same thing: how much the machine runs without you. That’s what they buy. Not your hours.

You get mental bandwidth back. Not to do more. To think better. And that’s probably the thing that has the highest impact on a company’s trajectory over ten years.

The short conclusion

If you end your week feeling like you worked your face off without moving the needle on what really matters — not on revenue, on the trajectory of your business — you probably didn’t work on the business this week.

You did its job.

The difference between the two is everything that matters over ten years.

Thibaut Vanderhofstadt is co-founder of Sortlist — Europe’s leading B2B matchmaking platform, active in over 140 countries. He now works with SaaS founders, marketplace operators, and ambitious SMB and mid-market executives through MetSaaS. Book a diagnostic →