A legal battle between
is expected to come down to the number of fake and spam accounts on the platform—a figure that data scientists say isn’t easy to objectively determine.
Twitter sued Mr. Musk Tuesday, seeking to enforce their merger agreement after his lawyers said in a securities filing last week that he wants out of the $44 billion deal. Mr. Musk alleged that the company misled him about the number of fake and spam accounts on its platform and didn’t provide him with the information he needs to audit its estimate. Mr. Musk’s team said the misrepresentation voids the contract, allowing him to ditch the acquisition.
Mr. Musk’s side is citing two different reasons tied to spam and fake accounts to leave the deal. One is that Twitter allegedly misstated facts about its user data in a way that could have a “material adverse effect” on its business. State law allows companies to nullify mergers if a material adverse effect has occurred, but courts have also tightly circumscribed the conditions for such an outcome.
The other is that Twitter allegedly violated its commitment in the merger agreement to provide all data and information that Mr. Musk needs to complete the deal.
The lawsuit lays the groundwork for a complex examination of how Twitter estimates fake and spam accounts, according to experts in Delaware law. The suit was filed Tuesday in Delaware Chancery Court because Twitter, like most large public companies, is incorporated there.
That path could involve dueling experts and reams of records that relate to Twitter’s estimation process. Those would be necessary to judge Mr. Musk’s claim that Twitter misled him about an important fact, said
executive director of the Institute for Law & Economics at the University of Pennsylvania.
“He is going to have to prove that in some demonstrable way what Twitter has disclosed and represented about that rate is materially false or misleading,” Mr. Hamermesh said.
Twitter said Tuesday that it had “bent over backwards” to address Mr. Musk’s “increasingly burdensome requests” for information. The company said that while Mr. Musk has continued to accuse Twitter of misrepresenting its spam or fake account estimate, he hasn’t offered “a single fact to support the accusation,” according to the complaint.
A spokesman for Twitter declined to comment. A lawyer for Mr. Musk didn’t respond to a request for comment.
Proving that Twitter’s estimate of fake or spam accounts is false could be exceedingly difficult, data scientists say. Coming up with a precise, objective and authoritative number of fake or spam accounts, where hundreds of millions of tweets are posted daily, is an impractical task.
There is no universal definition of fake or spam accounts, and highly sophisticated technology can make posts seem as if they were written by humans when they were made by bots for nefarious purposes.
Twitter has long estimated that less than 5% of its monetizable daily active users are fake or spam accounts. Mr. Musk has questioned that number, saying it is likely closer to 20%.
An expected point of dispute is whether Twitter did a good job in reaching its estimate, said
chief executive of Human Security Inc., which specializes in preventing bot attacks and online fraud.
“How hard did you try, and did you make all reasonable effort to get to an accurate approximation of the level of bot activity? There is an entire range of accuracy based on the answer to that question,” Mr. Hassan said.
Twitter has said its process for estimating fake and spam accounts on its platform can’t be replicated by outsiders in part because it involves multiple human reviews of thousands of accounts sampled at random over time. The company has said its process also relies on private user data it doesn’t disclose publicly and won’t share with Mr. Musk, such as IP addresses and phone numbers.
A more streamlined path for Mr. Musk to prevail in court would be trying to prove that Twitter breached the agreement by failing to give him information he needs to verify its estimates, according to the Delaware law experts. The agreement grants Mr. Musk the right to get information from Twitter that has a “reasonable business purpose” related to the transaction.
Delaware’s corporate-law courts are used to adjudicating claims of stonewalling in mergers in which one party argues the other isn’t providing information it sought.
“If he is going to have a viable claim of right to exit, I would think the informational shortcoming has to be material,” said Mr. Hamermesh of the University of Pennsylvania.
Twitter said in the lawsuit it had dedicated significant resources to providing Mr. Musk’s team with the data requested, “sweeping far beyond the bounds of what might conceivably be deemed reasonably necessary to consummate the transaction.”
That included access to its so-called fire hose data of near real-time tweets. The company also said it walked him through how it calculates the average number of monetizable daily active users on its platform—the figure it uses to reach an estimate of fake or spam accounts—which it most recently pegged at 229 million.
Twitter has hired Wachtell, Lipton, Rosen & Katz to represent it in its fight with Mr. Musk. Wachtell’s bench of Delaware law experts includes
the former chief justice of the Delaware Supreme Court who sat on the Court of Chancery earlier in his career.
As a vice chancellor of the Court of Chancery in 2001, Mr. Strine ruled against Tyson Foods Inc. when that company tried to abandon its acquisition of meatpacker IBP Inc. Mr. Strine forced the takeover. Mr. Strine didn’t immediately respond to requests for comment.
Mr. Musk waived his right to condition his offer on due diligence of Twitter’s business. But that probably isn’t by itself fatal to his claim, Mr. Hamermesh said. That is because Mr. Musk argues that he relied on Twitter’s public disclosures about the number of fake or spam accounts, and the merger agreement says its filings are complete and accurate. If Mr. Musk can show through the lawsuit that Twitter disclosed estimates that it knew were wrong, he could have a basis to walk away.
Twitter has gotten it wrong before: In April it said it had overstated its user base for nearly three years through the end of 2021 because of an error in how it accounted for people linked to multiple accounts. The revision reduced the number of monetizable daily active users by 0.9% for the fourth quarter of last year.
Arriving at a reasonable estimate of spam and fake accounts is possible with the right data and analysis, said
a former data-science lead at Twitter who is now chief executive of Surge AI, a data-labeling and content-moderation company. “It’s a mix of art and science.”
Twitter’s securities filings caution investors that the true number of fake and spam accounts “could be higher than we have estimated.” The company could argue Mr. Musk knew the risk he was taking, especially given he is the CEO of two large companies,
and Space Exploration Technologies Corp., said
Williams Chair in Law at the University of Richmond School of Law.
“It’s not like he’s an amateur,” Mr. Tobias said of Mr. Musk.
Twitter requested expedited treatment of the case, which it said would protect its shareholders from “the continuing market risk and operational harm resulting from Musk’s attempt to bully his way out of an airtight merger agreement.” Expedited cases in the Chancery Court typically play out over the course of a few months rather than several months or years.
It is possible that the two sides could negotiate a settlement that doesn’t require a trial. Mr. Tobias said it is in Twitter’s and Mr. Musk’s interests to reach a resolution quickly. The deal has already created turmoil for Twitter’s staff and tension for Tesla shareholders, and a court battle could be expensive and time consuming.
“That’s why people settle,” he said.
Copyright ©2022 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8