How to Score Whether Your Company Survives Losing Its Key Person Tomorrow

On January 1, 2026, Greg Abel became chief executive of Berkshire Hathaway, and the most closely watched succession in American business stopped being a hypothetical.¹ For most of the six decades before that morning, Berkshire and Warren Buffett were close to interchangeable - a company built so completely around one person's judgment that asking whether it could survive without him was really asking whether it had a soul independent of its founder. That question isn't unique to Berkshire. Every company built around one hard-to-replace person eventually faces its own version of it, usually with far less warning and far less preparation. What makes Berkshire useful isn't the drama of the handover - it's that the readiness behind it can be measured, on the same four dimensions any company can grade itself on before its own version of the moment arrives: bench strength, board alignment, knowledge transfer, and whether the person currently in charge is actually willing to let go.

That kind of edge is exactly what makes a succession dangerous: the more of a company's result traces back to one person's judgment, the more it's quietly carrying a risk it has probably never priced. A scorecard forces the pricing. Instead of a memo that says 'we have a plan,' score four things that can each be checked against evidence: bench strength, board alignment, knowledge transfer, and the incumbent's actual willingness to let go. Do that honestly and you get something worth having - a number you can argue with, and a clear answer to which single weakness would break the handover first, if it happened tomorrow. Berkshire is the extreme case, because no company has ever depended this visibly on one person's read of a balance sheet. But the same four questions apply whether the irreplaceable person compounds capital for a conglomerate or is simply the only one who knows why the biggest client still calls back.

Bench strength: a name on a memo, or a person who's already been tested: the difference between a designated successor and a proven one

Bench strength asks a blunt question: if the top person left tomorrow, is there someone who could actually run the place, or just a name someone floated at an off-site once? The honest test isn't whether a successor has been identified - it's how long they've already been tested in a role big enough to fail in. Abel joined Berkshire in 2000 through the energy and utilities business, was given oversight of most of the non-insurance empire as vice chairman in 2018, and was publicly named Buffett's successor in 2021.² Each step was a real audition in a job big enough to matter, run in view of the board years before the board actually had to decide. That sequence is the whole point of scoring bench strength rather than just asking about it: a plan can be announced in an afternoon. A person capable of running a company can only be demonstrated over years, in public, where being wrong would have shown.

Board alignment: united behind a plan, or scrambling when it's forced: a fast vote is evidence; a fought one is a warning

Board alignment measures something different from bench strength: not whether a good candidate exists, but whether the people with the actual power to install them agree on the plan, the timeline, and the pick. A board that's divided, or has never discussed succession out loud, improvises under pressure - and improvising under pressure is how bad handovers happen. Berkshire's board didn't improvise. Buffett recommended Abel take over as CEO effective January 1, 2026, and the board voted to confirm him unanimously the next day.¹ That kind of speed is what real alignment looks like from the outside - not because the vote itself was the hard part, but because a board only moves that fast on a decision it has, in every way that matters, already made.

Knowledge transfer: written down and distributed, or trapped in one head: the difference between a system and a memory

Knowledge transfer is the dimension most companies fail quietly, because it's slower and less visible than naming a successor or holding a vote. It asks how much of what the departing person knows - the relationships, the judgment calls, the reasons things are done a certain way - has been written down, distributed, or built into the system, versus how much simply leaves with them. Berkshire's answer was structural rather than documentary: long before the handover, it had already decentralized dozens of operating subsidiaries to run themselves day to day, funded from their own cash flow, with only the allocation of Berkshire's excess capital staying centralized.³ That meant daily survival was never actually contingent on Buffett's personal knowledge in the first place - the knowledge that ran the businesses had already been transferred to the managers running them. What stayed genuinely concentrated, and what no org chart can fully redistribute, was Buffett's own feel for capital allocation - a harder problem this scorecard comes back to.

Willingness to let go: watch what the incumbent does, not what they say: the only dimension you can't verify from a policy document

Willingness to let go is the dimension every other one depends on, and the hardest to score in advance, because it can't be verified from an org chart or a board resolution - only from what the incumbent actually does once someone else is nominally in charge. A founder can name a successor, season them for years, and win a unanimous board vote, and still quietly keep making the real decisions from the chairman's seat - because letting go is the one dimension that costs the incumbent something personally, in status and relevance and the feeling of being needed, while costing nothing to fake in a press release. The test was never the announcement. It's what happens the first time the new person has to act without checking in first.

Buffett was talking about Berkshire's agreement to acquire the homebuilder Taylor Morrison for $72.50 a share in cash, a deal Berkshire valued at roughly $8.5 billion in enterprise value and announced in May 2026 - the first acquisition of real size struck entirely on Abel's watch.⁴ The deal terms aren't really the evidence; plenty of successors get to sign a press release while the old guard still reviews every term beforehand. What's diagnostic is Buffett's own account of his involvement: none.⁵ That is what a passing score on willingness to let go actually looks like in the wild - not a retirement speech, but a former chief executive who can point to a multibillion-dollar decision he had no hand in, and say so, on the record, without flinching.

The whole build, compressed into one timeline: from a junior hire to the corner office, none of it improvised

What the scorecard can't fully capture: the real test was never the plan - it's what happens under pressure

Here the honest caveat matters, because a high score on all four dimensions isn't the same as eliminating the risk - it's the most any company can actually do about it. Two things resisted every dimension of Berkshire's engineering. Buffett's specific talent for capital allocation was never fully written down as a repeatable system; it lived in judgment built over six decades, and judgment is the one asset a scorecard can flag as concentrated but can't force anyone to transfer. And Berkshire's patience - its willingness to sit on cash for years and pass on fashionable deals - was protected for decades by a reputation only Buffett had earned, the kind that lets a company say no to Wall Street without being punished for it. A successor inherits the balance sheet. He doesn't automatically inherit the credibility that let his predecessor be boring on purpose.

None of this makes succession safe, and it isn't supposed to. What a scorecard does is convert 'we'll figure it out when it happens' into something you can actually see coming: four scores, one of them almost always lower than the other three, and a clear answer to which weakness would break first under real pressure. Buffett spent decades pushing his own four scores up before anyone forced the question. Most companies don't get decades. The ones that survive losing their key person aren't the ones who got lucky with the timing - they're the ones who ran this scorecard honestly while there was still time to act on what it found.