Discover more from Build To Scale
What Went Wrong with the OpenAI Board?
And how to build an effective startup board
This weekend’s drama at OpenAI raises a lot of questions about what makes an effective board. While we don’t know all the details behind Sam Altman’s ouster, the poor communications, board wavering and disastrous outcome indicate a high degree of dysfunction. There have been a few excellent posts by investor Chris Yeh and Hubspot founder Brian Halligan on the specifics of OpenAI so I’ll focus on the lessons for startup founders and CEOs on how to build an effective board.
First of all, it’s worth spending a moment to reflect on the primary purpose of the board of directors. A board does not run the organization; that’s the CEO’s job. Nonetheless, the CEO has a boss and that’s the board of directors. The board’s role is to ensure the proper governance of the organization and is responsible for hiring and firing of the CEO. That’s it.
Thanks for reading Build To Scale! Subscribe for more posts on how to scale your startup.
In the case of a typical venture-backed startup, the board has a fiduciary responsibility to shareholders. This was not the case for OpenAI, which had an unorthodox structure as a non-profit with a for-profit division. Surprisingly, OpenAI’s board had no representation from its outside investors Microsoft, Sequoia and Thrive Capital and seemingly no obligation toward them.
Building A Balanced Board
In a typical for-profit startup, the board consists of the founders and some of the investors. Usually, early stage seed investors don’t take board seats and if they do, they don’t keep them for long. That makes sense given the large number of investments they undertake. Generally early Series A and B venture investors take board seats in order to provide guidance to the founders and keep a watchful eye on things. Later stage investors may or may not get board seats or observer (non-voting) seats.
I have been on boards with investors like Bill Gurley and Peter Fenton from Benchmark, and there is no doubt that top-notch VCs provide tremendous strategic value to the companies they invest in. That said, VCs with operational experience are fairly thin on the ground.
If you’re the CEO of a startup, while you work for the board, you’re also responsible for the composition of the board. Venture investors (as opposed to seed investors) won’t usually abdicate a board seat and often their investment agreement may preclude such a change.
There is usually a provision for the CEO or founders to add one or two independent board members, unaffiliated with the investors. Ideally, independent board members add experience and skills that complement the founders and investors.
Typically you are looking for people with operational expertise that is relevant to your business and the challenges you’re facing. Unlike investors, independent board members usually have a four-year agreement, though that’s not a hard cutoff date.
A Board Must Be Helpful
I first decided to become a board member after participating in board meetings at MySQL. The MySQL board was a good one, and had a mix of investors Kevin Harvey (Benchmark), Danny Rimer (Index) and operational executives Bernard Liautaud (BusinessObjects), John Wattin (TietoEnator), Dana Evan (Verisign) and Tim O’Reilly (O'Reilly Media) as well as the founders.
My inspiration came from one particularly complex discussion around our technology strategy where it seemed the input was all over the place. I thought to myself: in this kind of situation, I would like to be the helpful board member.
I have always believed that the board is there to support the CEO. The CEO has a difficult job and the issues that are brought to the board are inherently difficult. A good board can provide help when navigating choppy waters.
In the case of the OpenAI board, there had been three resignations in 2023 (Reid Hoffman, Will Hird, Shivon Zilis) over the last year leaving the board unbalanced.
The board had also clearly not thought through the implications of firing Altman. They waffled over the weekend and the result has left the company in a vastly weakened state. They jeopardized an active funding round, angered their largest investor and technology partner, alienated employees and showed an extreme level of incompetence.
Big Moves Should Be Carefully Deliberated
When a founder or C-level executive is moved out of a key role in a company, this is not to be taken lightly. In a good board, such a move would be the subject of weeks if not months of discussion and planning. This includes succession planning, contingency plans, internal and external communications plans and more. It doesn’t appear that the OpenAI board did any of this.
In extreme cases where criminal behavior or workplace misconduct is suspected, there may be a need for an emergency removal, but even in such a case there must be careful consideration to operational details to ensure they don’t lose the trust of employees.
The OpenAI board indicated that they “lost confidence” in Altman. Further there were rumors of rifts between co-founder Ilya Sutskever and Altman around a reduction in Sutskever’s responsibilities and Altman’s focus on rapid commercialization over safety.
It’s not the board’s role to resolve management squabbles. That’s the CEO’s job. I may be reading too much into the reporting but it does appear Altman’s rift with Sutskever became a board level issue. Since Sutskever was a co-founder with a board seat, perhaps it was inevitable. Still, I have seen rifts between founders and CEOs get resolved with a lot less drama than what we saw this week.
When a shit show like this emerges, there’s enough blame to go around for everyone. While I would largely fault the OpenAI board, Altman has admitted he should have done a better job managing the board.
Final Thoughts for Founders and CEOs
When there is a fundamental disagreement on strategy or direction, whether at the board or management level, it is up to the CEO to get everyone on the same page. Neither the board nor the executives should ever be confused about who is running the show. That is the CEO’s job. The board provides input, but it is the CEO who steers the ship.
So if you’re a founder or CEO of a startup, what lessons should you take from this train wreck?
In the unlikely event that you have an unorthodox board structure serving multiple masters, you’ve got to make sure all constituencies are represented. Similarly, you must create mechanisms for resolving conflicts that will likely arise, whether the issue is balancing commercialization and security, openness and censorship, content moderation etc.
You should strive for a balanced board with diverse views. If your board consists exclusively of investors, consider adding an independent board member with operational experience that can provide guidance as you scale.
Don’t leave board seats empty for too long. Not because you might need the votes, but because you’re missing out on potentially valuable input.
Resolve operational issues and management rifts before they become board-level problems. With co-founders this is sometimes easier said than done. If executives cannot work together, they must be removed.
Keep your board apprised of what you’re doing and what you’re thinking. A board relationship must be high trust. There should be no surprises. Whether it’s good news or bad news, share it.
If there is conflict, get it out into the open. You must be able to discuss complex strategic issues with the board and get to a shared understanding of the best way to proceed.
I have found it useful to schedule regular communications with board members in order to stay in sync. I provided updates at least every six weeks (mid-quarter, end of quarter) in addition to the regular board meeting. For sensitive issues, I would reach out to them as soon as I was aware of a problem, let them know my plan and ask for additional input.
No doubt we’ll learn more in the days and weeks ahead about the fate of OpenAI, Sam Altman and Microsoft. Few companies go through this level of public drama. And perhaps with more thought to board composition we can avoid similar blow ups in the future. What lessons do you take from this? Post a comment below.
Thanks for reading Build To Scale! Subscribe and share with others.