The Automattic Board Fight: A Control Lesson for New Founders

Here's the compressed timeline.

September 9: Automattic's board votes to put founder and CEO Matt Mullenweg on a paid leave of absence. CFO Mark Davies is named interim CEO.

September 11: Mullenweg posts in the company Slack accusing board members of conspiring against him, removes other admins from the workspace, and tells employees he's back in control.

September 12: Automattic confirms it. A spokesperson says Mullenweg is chairman and CEO "with full support of the board."

September 14: Reporting indicates the board members who voted for the leave are out. So is Chief Legal Officer Andy Missan. So is Davies, who never really got to be interim CEO.

Five days. A founder removed and restored, a board replaced, two senior executives gone.

You can read this as palace intrigue at a company most founders will never work at. That's a waste. It's a live demonstration of how corporate control works when the polite fiction breaks down, and the mechanics are the same whether your company has nine employees or nineteen hundred.

A lot of first-time founders don't register this until it's happening to them: the CEO works for the board. Not the other way around.

The board hires and fires the chief executive. That's one of its core functions, along with approving major transactions and compensation. When you incorporate and issue yourself shares, you own a piece of the company. Owning shares makes you a shareholder. It does not make you unfireable.

The connection runs through director elections. Shareholders elect directors. Directors appoint officers. So if you control enough votes to elect the board, you effectively control who can fire you, because any board that tries can be replaced by the shareholders who put them there.

That's the loop that decided the Automattic situation. The board acted. Shareholder power then reshaped the board. This is not a loophole. It's the basic architecture of a corporation.

The founders who get surprised by it are the ones who never mapped their own version of the loop.

These get conflated constantly and they are not the same.

Voting control is about shares. What percentage of votes do you cast at a shareholder meeting? Dual-class structures, where founder shares carry ten votes each and everyone else's carry one, exist to keep this number high while ownership economics get diluted.

Board control is about seats. Who sits at the table and how do they vote? A typical seed-stage board is three seats: one founder, one investor, one independent. Series A often moves to five.

You can have one without the other, and each creates a different vulnerability. High voting control with a hostile board means you can eventually fix the board, but you might eat an ugly week getting there. Board control with weak voting power means a shareholder coalition can remove your directors.

Ask yourself two questions and write down the actual numbers. If every investor voted against me tomorrow, what percentage of shareholder votes do I control? And if every director voted against me tomorrow, do I still have a seat?

Most founders can't answer either without opening three documents. That's the problem.

Founders negotiate hard on the investor seat and then wave through the independent seat like it's a formality. It isn't. On a three-person board with one founder and one investor, the independent director is the deciding vote on everything contested.

Pay attention to three things in the docs. Who nominates the independent? Who has to approve them? And what happens if the seat sits empty?

The strongest version for a founder is mutual consent: you and the investor both have to agree, which means neither can install an ally. The weakest is language letting the preferred shareholders designate the seat if it's vacant past some deadline. That clause looks like housekeeping. It's a trapdoor.

Also: fill the seat. Founders leave it empty for eighteen months because the search is annoying, and an empty seat shifts the arithmetic in ways nobody modeled.

When you do fill it, pick someone who will disagree with you in the room and back you outside it. An independent who just nods is a wasted seat. One who leaks to investors is worse than an empty one.

You will not win everything. Pick the items that matter when things go wrong.

Board composition at each stage, written into the docs. Not "we'll figure it out at Series A." Specify how the board expands and who nominates each new seat.

Protective provisions you can live with. Investors get veto rights over certain actions. Read the list carefully. Vetoes over selling the company or issuing new preferred stock are normal. A veto over hiring decisions or annual budget approval hands over operating control, and founders sign these more often than you'd think.

Removal mechanics for your own seat. Is your board seat tied to being CEO, or to your shareholding? If it's tied to the CEO role, losing that job costs you the seat too, and you're out of the room precisely when you most need to be in it.

Information rights that run both ways. You should know what investors are told about you.

Map how these interact before you sign. A spreadsheet handles it. So does Notion, or a planning tool like Foundra that keeps your ownership assumptions next to the rest of your plan. What matters is one place where the picture lives.

Mullenweg kept the job. Look at what it cost anyway.

The chief legal officer left. The CFO left. Multiple board members are gone. The whole thing played out on TechCrunch across four articles in six days, plus hundreds of comments on Hacker News from developers who use the company's products. Automattic is separately in high-profile litigation with a WordPress hosting company.

Every departure represents institutional knowledge that walked out. Every article is something a recruiting candidate finds when they search your company. Every enterprise customer with a renewal coming up now has a question about stability.

Winning a governance fight is closer to winning a lawsuit than winning a deal. Better off than losing, meaningfully worse off than never having the fight.

Which is the actual lesson. The goal is not to be well positioned for the boardroom battle. It's to build a board where the battle never becomes the available move, because the people in the room trust each other enough to argue in it.

Structure buys protection. It does not buy alignment.

Ask these out loud and write down the answers.

1. What does the board look like after this round, and after the next one?
2. Who nominates the independent director, and who has to approve?
3. What happens to that seat if it's vacant for six months?
4. Is my board seat tied to my title or to my shares?
5. What's on the protective provisions list, specifically, and which of those do I expect to do in the next two years?
6. What vote threshold removes the CEO, and what threshold replaces a director?
7. If the three largest shareholders agreed on something tomorrow, what could they do without me?

Then ask your lawyer the same questions and compare. If the answers don't match, you misunderstood a document you already signed, which is worth finding out now rather than during a five-day news cycle about your company.

None of this assumes bad faith from your investors. Most investor relationships never come close to this. It's about knowing the shape of the thing you're inside of, which is reasonable to know about a company you're spending a decade on.

• Automattic's board voted on September 9, 2026 to put Mullenweg on leave. By September 12 he was confirmed back as chairman and CEO. By September 14 the directors who voted for the leave were reported gone, along with the CFO and chief legal officer.
• The CEO reports to the board. The board is elected by shareholders. Whoever controls the shareholder vote eventually controls the board.
• Voting control and board control are separate. Know your number for each.
• The independent seat decides contested votes on a three-person board. Negotiate how it gets filled, then fill it.
• Read protective provisions line by line. Vetoes over budget and hiring hand over operating control.
• Tie your board seat to your shareholding rather than your title.
• Winning a governance fight still costs executives, trust, and press. Structure protects you. It doesn't replace alignment.

Can my investors fire me if I own a majority of the company?

The board can remove you as CEO. But if you control enough shareholder votes, you can replace the directors who did it, roughly the loop that played out at Automattic. Majority ownership makes removal temporary rather than impossible, and the interim can still be destructive.

Should I push for a dual-class share structure at seed stage?

Almost never, and most seed investors will decline. Dual-class typically shows up later, often near an IPO, for founders with real bargaining power. At seed, spend your negotiating capital on board composition and protective provisions instead.

What's a reasonable board structure for a seed round?

Three seats is common: one founder, one lead investor, one independent chosen by mutual consent. Expanding to five at Series A is normal. What matters more is who nominates each seat and what happens when one sits vacant.

How often should a founder actually think about this?

At each financing, and once a year otherwise. Set a reminder to reread your charter and voting agreement. It takes an hour and most founders have never done it once.

Is it a red flag if an investor pushes back on these questions?

Asking is normal and good investors expect it. Vagueness is the concern. "We'll sort that out later" about board composition is not an answer, and later is always a worse time to negotiate.

1. TechCrunch: Automattic's board forces CEO Matt Mullenweg into leave of absence
2. TechCrunch: Mullenweg tells Automattic staff in Slack he's back in control
3. TechCrunch: Automattic confirms Mullenweg has returned as CEO
4. TechCrunch: Sources say Automattic's board is out after failed attempt to oust CEO
5. Hacker News discussion: Mullenweg has returned as CEO after attempted board ouster