On May 7, 11 days before the IPO, Facebook management -- CEO Mark Zuckerberg, COO Sheryl Sandberg, and CFO David Ebersman -- stepped out of a black SUV in front of the Sheraton Hotel in midtown Manhattan for their first meeting with investors. A crowd of paparazzi greeted them, and a long line of onlookers wound around the hotel building. Inside the meeting, Facebook played a video introducing the business model to special clients of its underwriting banks.
Although an IPO roadshow is supposed to be an untarnished hype-machine for a company's prospects, back in California, those prospects were hurting. Facebook's new internal forecasts showed revenue growing slower than expected. The reason: Users were flocking to smartphones faster than the company could serve mobile ads.
On the first day of what may have been the most watched IPO roadshow in memory, Ebersman confessed to Morgan Stanley that Facebook had cut revenue projections -- a nearly unprecedented last-minute correction in an IPO of its scale. Even if the changes were small, statistically, in IPO showbiz statistics run second to momentum, and nothing kills momentum like a poorly timed downward revision.
Facebook and Morgan Stanley knew they had to make a public disclosure. But what to disclose? The law requires companies to share all information that would likely influence an investor's decision to buy stock. Plus, Morgan Stanley's research team was still advising clients based on figures that Facebook now considered wrong. With less than a week before the IPO, they came up with a solution that they thought would spare Facebook a modicum of embarrassment -- but would have fateful consequences for Morgan Stanley and investors:
• First, Facebook would file an amendment to its public birth certificate, the S-1, to include information about mobile usage cutting into revenue.
• Second, the company would call research analysts with much more specific information about the company's weakening projections.
IV: THE AMENDMENT
When a company makes an amendment to its S-1, the entire document can be filed again, without track-changes or highlights to specify the changes. When Facebook filed its "Amendment No. 6 to Form S-1" to disclose its ongoing challenge to serve ads to mobile users, changes were embedded on three pages of the 170-page document.
On page 14 and 17, the company said that its users were growing faster than ads due to the increased use of mobile phones and product decisions that allowed fewer ads per page. On page 57, Facebook said the mobile trend discussed elsewhere in the document had continued in the second quarter, due to users shifting from computers to phones. None of the changes suggested any major revision to Facebook's financial outlook.
Facebook's lowered revenue estimates did not appear in the S-1, nor was there any precedent requiring these numbers to be included. Even the most sophisticated retail investors, armed with a software bot that could comb the new S-1 for updates, could not have read what research analysts at Morgan Stanley, JP Morgan, Goldman Sachs, Citigroup, and many other investment banks would learn later that evening: That Facebook, already projected to trade at high multiples given its earnings figures, was slashing its annual projections.
V: THE CALL
Before an IPO of Facebook's size, research analysts and large investors play a massive multi-billion-dollar game of "Telephone," because analysts at underwriting firms are banned from publishing or emailing research about a new public company until 40 days after the IPO. The rule is designed to protect retail investors from taking analysts' notoriously bullish research too seriously. But it has an unintended side effect: Research analysts can pass on exclusive, last-minute information to institutional clients without a paper trail.
And Facebook's last-minute revelation was about to launch a historic game of "Telephone."
After the company's surprising eleventh-hour amendment, the unenviable job of explaining Facebook's revisions to the research analysts fell to Cipora Herman, Facebook's vice president of finance. On May 9, Herman and Michael Grimes huddled in a Philadelphia hotel for a rehearsal session. He played the part of an analyst, while she played herself. They practiced until the amended S-1 was officially made public on the SEC website, at exactly 5:03pm.
Before her first call to Scott Devitt at Morgan Stanley, Grimes left the room to avoid the appearance of breaking the law that bans investment bankers from influencing analysts. "I was far down the hall so I wouldn't hear anything," he testified in a Massachusetts suit brought against Morgan Stanley. "I took extra precaution to do that, and sat on the floor."
But Herman had a script in Grimes' handwriting detailing Facebook's new second-quarter revenue estimates:
"I wanted to make sure you saw the disclosure we made in our amended filing. The upshot of this is that we believe we are going to come in the lower end of our $1.1 to $1.2 bn range for Q2 based upon the trends we described in the disclosure."
The script also showed that Herman had added Facebook's year-end estimates:
"Trend/headwinds over the next six to nine months as this run through the rest of the year, this could be 3 to 3 and a half percent off the 2012 $5 billion target."
Herman's first three calls were to the research desks of Morgan Stanley, JPMorgan, and Goldman Sachs, and after their conversations, all three banks cut their estimates of Facebook's annual revenue by between 3.01% and 3.33% -- perfectly aligned with Herman's notes, as the following charts from court documents show:
The news spread -- from research analysts to their preferred clients -- that Facebook was slashing revenue estimates in the middle of the roadshow. Morgan Stanley, JP Morgan, and Goldman Sachs had effectively sounded alarm bells on Wall Street. Even analysts who hadn't yet made a single call about Facebook's new figures were being inundated with questions. "Our clients were asking us to confirm what they had heard, so we had to tell them what we knew," said a research analyst at one of the other underwriting firms.
Many large investors sensed a rare opportunity: With its revenue projections falling at such a peculiar time, the conditions could be perfect to place a massive bet against Facebook stock.
VI: THE BET
A week before the IPO, confusion reigned in the financial press.
On May 11, Bloomberg reported that demand for Facebook's stock among institutional investors was much lower than expected. But on the same day, Reuters had a conflicting report: Facebook was already oversubscribed, and one large unnamed institutional investor was calling around syndicate desks trying to get more shares.