Twitter's board gave Dorsey a growth target and 20 months later sold to Musk instead

In late February 2020, one of the most feared activist funds on Wall Street revealed it had lined up its guns on Twitter — and it wanted the founder gone. Elliott Management had been circling since Jack Dorsey announced he would decamp to Africa while continuing to run both Twitter and Square at the same time.⁴ By March 9, the fight was over almost before it started. Dorsey kept his chair. Elliott got a seat. Everyone shook hands. It looked like the founder had won — and for twenty months, it worked. Then Dorsey resigned, and the company Elliott had wanted fixed was sold whole to Elon Musk instead.

The official story is that Twitter's board rallied around its founder and beat back an activist raid. That is not what happened. What happened is that the board handed Dorsey a stay of execution with an expiration date printed on it — a set of public, time-bound growth conditions — and then watched the clock run out.

A 4% stake nearly ended a founder's tenure: the leverage in an activist raid was never the size of the position

Start with the number that makes the whole episode strange. Elliott owned roughly 4% of Twitter's stock.² Not a controlling block, not even close — a minority sliver in a company with a diffuse shareholder base. On the strength of it, Elliott had reportedly prepared to nominate four directors and push the founder out of his own company.³ The lesson every board should take from this is that an activist's power is not proportional to its ownership. It is proportional to how many of the other shareholders quietly agree. Elliott's 4% was a lever, not a load. What made it dangerous was the possibility that the silent majority of investors — tired of a part-time CEO — would vote its way in a proxy fight. The settlement was the board's way of never finding out whether they would.

Note the misconception worth clearing here: some retrospective accounts later inflated Elliott's position to around 9%.⁸ The contemporaneous record — Twitter's own SEC filing and reporting from the day of the deal — says roughly 4%.²³ The gap matters, because the smaller the real stake, the more the story is about the board's fear rather than Elliott's firepower.

The reprieve came with a countdown attached: peace was purchased with a growth target, and targets have deadlines

This is the part the 'founder won' framing skips. The settlement was not an unconditional pardon. In its own announcement, Twitter committed to specific, measurable ambitions: grow monetizable daily active users by 20% or more in 2020 and beyond, and accelerate revenue growth past 2020.² It authorized a $2 billion buyback and took a $1 billion investment from Silver Lake, whose Egon Durban joined the board alongside Elliott's Jesse Cohn.¹² Most tellingly, the board formed a committee expressly to review the CEO and leadership structure.² Read plainly, that is a board saying to its founder: you may stay, but we have written down what you owe us, and we have appointed a committee to check the receipt.

Eight months in, the board reaffirmed him and the numbers didn't cooperate: confidence votes are loudest precisely when the evidence is thinnest

By November 3, 2020, the board publicly reaffirmed its confidence in Dorsey, pointing to a 14% year-over-year revenue rise to $936 million in the third quarter.⁵ But look at the metric the settlement had actually enshrined. That same quarter, monetizable daily active users came in at 187 million — short of the 195.2 million Wall Street expected.⁵ The board had staked its truce on 20%-or-more mDAU growth, and the one hard data point in the public record showed the company already missing expectations on precisely that measure, even as the directors stood up to say the arrangement was working. The settlement had been characterized as having bought Dorsey time to prove he could run two public companies at once.⁵ The reaffirmation was less a verdict on results than a decision to keep extending the loan.

The outcome arrived without Elliott firing a second shot: the raid failed, and everything the raid wanted still happened

On November 29, 2021 — a year and a half after surviving the ouster attempt — Dorsey resigned, announcing it, fittingly, via tweet. He was succeeded by CTO Parag Agrawal.⁶ Here is the crucial detail: no source ties this exit to a renewed Elliott campaign. The New York Times reported that Dorsey's leadership had been questioned by employees and investors who felt he was unfocused and spread across Square and other projects⁶ — the same complaint Elliott had made in early 2020, arriving on its own eighteen months later. The activist didn't need to come back. The concern it had voiced simply metastasized. Then, on April 25, 2022, Twitter's board unanimously approved the $44 billion sale to Elon Musk, and Dorsey left the board entirely, severing the last formal tie any founder had to the company.⁷

Didn't the settlement still work?: the honest case that Dorsey deserved the time he was given

The fair objection is that a founder who survives an activist raid and then chooses his own exit twenty months later has, by any reasonable standard, won — he left on his terms, not Elliott's, and revenue was climbing when the board backed him.⁵ There is real force to that. A proxy fight in March 2020 would have been bloody, and the settlement genuinely bought stability into a pandemic quarter. But notice what the concession quietly gives up. Elliott's original thesis was that Dorsey's divided attention was hurting the company. The settlement didn't refute that thesis; it deferred the test of it. And when the test finally came — through the board's own mDAU benchmark, then through the same investor discontent Elliott had first named — the verdict landed on Elliott's side anyway. The founder chose the timing of his departure. The activist chose everything else. Winning the timing of your own removal is not the same as winning.

The settlement worked exactly as a settlement is supposed to: it stopped the fight in the room. What it could not stop was the reason for the fight. A 4% stake never had the votes to fire a founder, but it didn't need them — it only needed to name the problem out loud and write the standard by which the problem would be measured.²³ Twitter agreed to be judged by that standard, missed it, reaffirmed the man anyway, and then handed the whole company to a buyer eighteen months after he was gone. Elliott lost the proxy fight it never had to hold, and got the outcome it actually wanted. The board bought time. It just never figured out what the time was for.