Control is not a defense against being wrong. It removes the thing that tells you
WeWork's supervoting structure was a bet that the founder's judgment would stay better than every check it disabled. Canadian Pacific rejected Hunter Harrison in a filing and appointed him six weeks later. Twenty-six control decisions on record.
What the structure actually buys
Twenty-six founder and control decisions are on record and twenty were never reversed. Control structures are durable, which is the point of them and also the risk.
The record's description of the WeWork decision is the sharpest line in this corpus. Granting Adam Neumann supermajority control through a multi-class share structure in the August 2019 IPO filing was a bet that the founder's judgment would stay better than every check it disabled.
That is what a control structure is in every case, stated without decoration. The supervoting structure was cut, the IPO withdrawn, Neumann's position unwound in a February 2021 Delaware settlement on terms far below the $1.7 billion legend, and WeWork filed for bankruptcy in November 2023 listing liabilities of roughly $10 to $50 billion.
Winning the vote is not keeping the strategy
DuPont chose to fight Trian's proxy slate in 2015 rather than concede board representation, seeking to keep its slate intact and vindicate management's strategy. DuPont won the vote. On 5 October 2015 it announced that Ellen Kullman would retire as chair and chief executive effective 16 October, and the same release cut full-year operating earnings guidance. The company moved toward the breakup Trian had demanded.
Canadian Pacific rejected Pershing Square's demand to install Hunter Harrison in a 2012 filing, on the stated grounds that Harrison was not the right leader and the nominees lacked Class I railroad operating experience. Fred Green resigned and Harrison was appointed president and chief executive on 28 June 2012, six weeks after the board's position collapsed. Pershing Square ended its association with CP in 2016 with an estimated $2.6 billion profit.
Both boards had the formal power to say no and used it. Neither outcome was decided by the vote.
The structures that held
Michael Dell took on debt rather than dilution in 2013 and the choice left him with concentrated voting power and roughly 46% ownership. LVMH locked ownership in 2022 and deliberately left leadership open, and the structure produced the board and ownership outcomes intended. Disney fought Nelson Peltz in 2024 and gave no seat, then installed the governance and succession structure Peltz had been asking for.
The pattern is not that these founders were better. It is that each separated the thing they were protecting from the thing they were refusing. LVMH protected ownership and conceded leadership. Dell protected voting power and paid for it in interest rather than equity. Disney refused the person and adopted the demand.
WeWork's structure protected everything at once, which is the same as protecting nothing in particular, and left nobody positioned to raise the question that mattered.
How to run this decision
Drawn from what separated the records that stuck from the ones that were undone. A step nobody in the record can be seen taking is not here.
- 01
Name the specific decision the control exists to protect
Long-term capital allocation. A refusal to sell. A product standard. Write it down as one decision, not as independence in general. A structure defending everything defends nothing specific, and it will be used to win arguments it was never justified for.
- 02
Write down what the checks would have caught
Before you disable a check, describe the kind of mistake it exists to catch, and say where else that mistake will be caught instead. If the honest answer is nowhere, that is the trade you are making, and it should be made deliberately.
- 03
Separate ownership from leadership, and say which you are keeping
LVMH locked one and opened the other. The two get bundled together in most founder-control arguments and they have completely different costs. Ownership survives you; leadership does not, and a structure that assumes otherwise has a fixed expiry.
- 04
Treat an activist's argument separately from the activist
Disney refused the seat and implemented the governance change. DuPont fought the slate, won, and lost the chief executive and the strategy within months. The two decisions — is this person coming onto my board, and is this person right — are independent, and conflating them is how boards lose both.
- 05
Price the control discount
Concentrated control costs something at every raise, every sale and every valuation, permanently. Put a number on it, compare it to the specific decision from step one, and decide whether that decision is worth that price. Dell paid in interest and kept 46%, which is at least an arithmetic anybody can check.
- 06
Build the reversal into the structure
A sunset, a conversion trigger, an independent chair, a defined event that returns the votes. The control decisions that ended badly here had no condition under which they would end, which is why the ending was somebody else's to write.
What it looks like when this is going wrong
- The structure protects the founder's judgment in general rather than one named decision.WeWork, 2019.
- No one can say where the disabled checks' work will now be done.WeWork, 2019.
- The board's public refusal and its private position have already diverged.Canadian Pacific Railway, 2012.
- The activist's argument is being assessed by whether you like the activist.DuPont, 2015.
- The structure has no sunset, conversion trigger or defined ending event.WeWork, 2019.