The org chart is the easy part. The people are the whole game.
In Part 1, I made the case that splitting the chair and CEO roles - now the global default - has a surprisingly weak link to performance. Boards tick the box and still walk off cliffs.
Here’s why. Governance codes regulate titles. They can’t regulate temperament. And the chair–CEO relationship - the single most important relationship in any company - lives or dies on temperament.
When it breaks, it almost always breaks in one of three ways. All three are versions of the same sentence: “my way or the highway”.
The Steamroller - “my way, from the corner office”
The classic failure: a dominant CEO and a chair who can’t, or won’t, push back. The board goes blind.
Uber is the US archetype. Travis Kalanick built a colossus - and a “toxic culture” to match, from harassment claims to regulatory evasion to a CEO filmed berating one of his own drivers. Super-voting shares made him almost untouchable. It took a hand-delivered letter from five major investors - “Moving Uber Forward” - to force him out in June 2017, and even then he clung to a board seat and a lawsuit. [1]
WiseTech is the Australian one. Founder Richard White stepped down as CEO in 2024 amid allegations about his conduct, and never lost control. When four independent directors, including the chair, resigned in February 2025, roughly A$10bn was briefly wiped. White, who owns about a third of the company, returned as executive chairman, and reportedly told analysts that customers don’t much care about “governance matters in Australia.” [2]
Same pattern, different hemispheres: Theranos’s board of statesmen who never tested the technology; Tesla’s board, which a Delaware judge found lacked the independence to rein Elon Musk in. [3][4] The titles were separated. The power was not.
The Backseat Driver - “my way, from the chair”
The opposite - and more genteel - failure: a chair who can’t let go, reaches down into management, and hollows out the CEO.
Nike. When founder Phil Knight stepped back to the chairmanship in 2004 and hired the company’s first outside CEO, William Perez, the experiment lasted 13 months. Perez left “citing differences regarding leadership” with Knight, who stayed an “active chairman” and judged the company was running at “80% efficiency” under his own hand-picked successor. [5] An outsider can’t lead when the founder is still steering from the chair.
Harvey Norman is the Australian version, scaled up to a family fiefdom. Gerry Harvey is executive chairman; his wife, Katie Page, is CEO; his son sits on the board; and barely a third of directors qualify as independent - well short of the ASX recommendation. Shareholders have repeatedly hammered the remuneration report (a 47.5% protest vote in 2020) and Harvey’s response is characteristically blunt: he won’t be “bullied,” he owns a third of the company, and the critics are “deluded.” [6] When the chair owns the room, “independent oversight” is a line in the annual report and little more.
In Harvey Norman’s case I would argue that the relationship actually works for the benefit of shareholders. In FY25 (to June 2025) Harvey Norman reported profit before tax of A$753.1m, up 39%, and net income of A$518m, up 47% year-over-year. Aggregated franchisee sales grew 6.1% to A$6.43bn, with the franchising margin improving to 5.36% from 4.52%, and the stock jumped about 10.7% on the result and the dividend was lifted. The numbers speak for themselves.
The Civil War - “two highways, head-on”
The most spectacular failure. Two strong personalities, neither willing to yield, turning the boardroom into a battlefield.
Volkswagen, 2015. Patriarch-chairman Ferdinand Piëch used a magazine interview to publicly disown his own CEO, Martin Winterkorn - the very confidant he had anointed as successor. The supervisory board split, the controlling families turned on each other, and within days it was the chairman who was forced out. VW spent the run-up to its emissions scandal consumed by a palace war. [7]
Disney. Michael Eisner ran the company as a combined chairman-CEO until Roy Disney and Stanley Gold launched the “Save Disney” revolt. In 2004, 43% of shareholders withheld support for Eisner’s re-election; the board stripped him of the chairmanship that same day, and he was gone within the year. [8] When power is concentrated and challenge is treated as betrayal, the only release valve left is open warfare.
The common thread
What none of these is about is the org chart. Uber, WiseTech, Tesla and Theranos all had - or were told to adopt - “proper” structures. Harvey Norman files a compliant governance statement every year.
The variable is power and personality. A CEO too dominant to challenge, a chair too attached to let go, or two egos who’d rather fight than align.
So what actually works?
If structure can’t fix it, what can? Four things not on a checklist:
- Hire the chair for the willingness to challenge, not just the independence box. “Independent” on paper and “willing to say no” in the room are different qualities. Reference-check for the second one.
- Build in release valves. Regular board sessions without management present, plus a genuinely empowered lead independent director, give dissent somewhere to go, before it becomes a resignation or a press leak.
- Write the rules of engagement. The best chair–CEO pairs agree explicitly, and early, on who owns what - strategy, the external voice, the board agenda - so the boundary is set before it’s tested under stress.
- Treat chemistry as a governance issue. Boards diligence a CEO’s CV exhaustively and then ignore whether that person and the chair can actually work together. The relationship deserves the same scrutiny as the appointment.
And the hardest discipline of all, know when it’s unfixable. A chair–CEO relationship that has curdled into domination or war rarely heals. The boards that survive are the ones that act early, and accept that sometimes the most important governance decision is simply deciding which of the two leaves.
The provocation
When the two most powerful people in your company disagree - who actually yields, and what happens to everyone else while they fight it out?
If the honest answer is “nobody yields,” you don’t have a structure problem. You have a personality problem.
And those are the ones that sink companies.
Part 2 of a two-part series; Part 1 - “The Chair–CEO Split Is Winning. So Why Do Boards Keep Failing?” - covers the structural side. Written from an Australian board and governance perspective. Which archetype have you seen up close?
References & further reading
- Reporting on Travis Kalanick’s removal as Uber CEO (June 2017) and the investor letter “Moving Uber Forward” (Reuters; Fortune; Bloomberg; Harvard Business Review, “When Founders Go Too Far”).
- Reporting on WiseTech board resignations, February 2025 (Sydney Morning Herald; Fortune; Startup Daily).
- Wharton, Darden and Yale SOM case analyses of Theranos governance and board oversight.
- Delaware Court of Chancery rulings on Elon Musk’s 2018 Tesla pay package; CNBC (Nov 2025) on the ~US$1tn replacement package; Tesla SEC filings (Chair Robyn Denholm).
- Nike SEC filings and reporting (2004–06) — William Perez’s appointment and departure after ~13 months “citing differences regarding leadership” with chairman Phil Knight (Nike Form 8-K; Associated Press).
- Reporting and shareholder-association commentary on Harvey Norman governance, board independence and the 2020 remuneration protest vote (SmartCompany; Michael West Media; Tyndall Asset Management ESG note).
- Reporting on the Volkswagen Piëch–Winterkorn power struggle, April 2015 (The Conversation; Fortune; Reuters).
- CNN and contemporaneous reporting on the 2004 Disney shareholder revolt (“Save Disney”) and Michael Eisner losing the chairmanship.
