Facebook's $16 Billion IPO That Everyone Called a Flop

On May 18, 2012 Facebook sold roughly $16 billion of stock at $38 a share, valuing the company near $104 billion, the largest technology IPO ever at the time and the third-largest US IPO in history. The stock opened on a Nasdaq exchange that malfunctioned for half the morning, traded for hours while investors did not know whether their orders had filled, and closed at $38.23, twenty-three cents above the offer, a number the lead underwriter manufactured by buying stock all day to keep it from breaking the issue price. Within four months it traded at $17.55, less than half the IPO price. Years of shareholder litigation and regulatory fines followed. Then the company that had "botched" its IPO became one of the most valuable enterprises on earth.

That arc is why the deal is a case study rather than a cautionary tale. An IPO can be graded two opposite ways depending on who you think it is for. By the standard of the company raising money, Facebook's was close to flawless: it priced at the very top of a range it had already raised, upsized the deal, and surrendered almost none of the value that issuers normally hand to underwriters' favored clients through a first-day pop. By the standard of market integrity it was a debacle: a broken open, research estimates cut and disclosed selectively to institutions during the roadshow, a price held up by the bank that sold it, and retail buyers who lost half their money before the business thesis turned. The central question is therefore not "was it good or bad" but "what is an IPO supposed to do, and for whom," and whether a deal that did exactly what the issuer wanted can still be a failure. This account is built from the prospectus and the pricing release, the SEC's order against Nasdaq, the Massachusetts regulator's findings, contemporaneous reporting, an independent pre-IPO valuation, and the documented recovery, and it lets each side speak.

A profitable company priced for perfection

Facebook did not come public as a hope. In 2011 it generated $3.71 billion of revenue, up 88% year on year, and about $1.0 billion of net income, with 845 million monthly active users and 483 million daily users at year end. That is a genuinely rare object: a consumer platform at global scale that was already solidly profitable before its IPO. The bull case did not need imagination, only continuation.

The discomfort was the price against that base. Near $104 billion, the equity was valued at roughly one hundred times trailing earnings and more than twenty-five times trailing revenue, multiples that priced not the company that existed but a far larger one the market was asked to assume would arrive. The contrast that every analyst reached for was Google, which had come public in 2004 at about $23 billion, already solidly profitable and growing fast, rose roughly 18% on its first day, and then compounded many times over the following years. Facebook was asking for roughly four times Google's IPO valuation on a smaller revenue base, which is the same as saying it priced in the Google-style decade before any of that decade had been delivered.

The subtler warning sat in the quarterly trend: first-quarter 2012 revenue had come in below the fourth quarter of 2011, the first visible deceleration just as the company asked for a perfection multiple. Two further facts a diligence-minded buyer would have weighed: the revenue base was about 85% advertising and meaningfully concentrated, with the game developer Zynga alone responsible for a low-double-digit percentage of it, and just weeks before pricing Facebook had agreed to buy Instagram for about $1 billion, a then-unprofitable, barely-monetized photo app, which told you something about both the mobile threat and how the company intended to answer it. The fundamentals were strong and the valuation assumed they would become extraordinary, which is a different and more fragile claim.

The mobile hole in the story

The specific fragility had a name. More than 425 million people were already using Facebook on mobile devices, and Facebook earned essentially nothing from them, because it did not yet show advertising in the mobile app at all. This was a product choice as much as an engineering gap: the company had been reluctant to compromise the clean mobile experience with ads, and its advertising system had been built for the desktop news feed. Usage was migrating to phones faster than the company could monetize phones, which meant the single most important user-growth trend was, in the near term, a revenue *headwind* rather than a tailwind, the rare case where a company's flagship metric improving made its near-term economics worse.

The S-1 and the quiet period

The S-1 is the registration statement a US company files with the SEC before an IPO: the prospectus, risk factors, financials, and use of proceeds. From filing through roughly 25 days after the IPO the issuer is in a "quiet period," tightly limited in what it and its underwriters' analysts may say publicly, which is precisely why how and to whom information moves during that window is so legally sensitive.

This was not hidden. It was disclosed, and then it was disclosed more sharply, and the moment it was sharpened became the hinge of the entire controversy. Hold that thread: the same mobile gap that nearly broke the stock in 2012 was the exact thing the company fixed to rescue it in 2013, which is why neither the bears nor the bulls get to claim the story cleanly.

One founder, permanent control

Buyers were also purchasing a particular governance bargain. Facebook used a dual-class structure in which insiders' Class B shares carried ten votes each. Mark Zuckerberg held about a quarter of the economics but, with super-voting stock and voting agreements over others' shares, controlled roughly 57% of the votes. Public investors were buying the cash-flow rights of the business and almost none of the control of it.

Dual-class share structure

Two classes of common stock with identical economics but unequal votes, typically high-vote founder shares and low-vote public shares. It lets a founder raise outside capital while keeping voting control, insulating long-term bets from market pressure. The cost is accountability: ordinary shareholders cannot replace a board or force a sale, so the structure trades governance protection for founder conviction.

There were two honest readings, and the case keeps both. One: this protected a founder with a strong record from short-term market pressure, and Facebook's later mobile pivot is itself the argument for it, because a quarterly-driven board might not have funded a costly bet that depressed near-term economics. Two: it concentrated unaccountable power and made Facebook a landmark in the long governance argument over dual-class tech listings. The structure is not incidental color; it priced into the deal, and it shaped how investors should have weighed everything that followed.

The range that kept rising

Pricing an IPO is a negotiation conducted through a moving number, the last and most consequential step of the IPO process. Facebook's first public range was $28 to $35 a share. As the roadshow generated apparent demand, the company raised the range to $34 to $38, then priced at $38, the very top of the raised range. In ordinary IPO craft a price set at the top of a lifted range is a signal of strength; here it also meant the deal had been walked up twice and left no headroom for the aftermarket to do anything except hold or fall.

The mechanics of how a book is built and a range is moved are covered in how an IPO price range is set; the point specific to Facebook is that the process optimized for one variable, the price the issuer and selling holders received, and accepted that doing so removed the margin of safety the first day usually relies on.

Upsizing, and who was selling

Days before pricing, the deal also grew. The share count was increased by about 25%, to roughly 421 million shares, and the additional supply tilted toward existing holders selling, not just the company issuing. Of the total, Facebook itself sold on the order of 180 million primary shares, raising roughly $6.8 billion for the company; the balance was secondary stock, early investors and insiders monetizing into the most anticipated IPO in years.

Primary vs secondary shares

Primary shares are newly issued by the company; the proceeds go to the company's balance sheet to fund the business. Secondary shares are sold by existing holders; those proceeds go to the sellers, not the company. A deal heavy in secondary stock, especially one upsized late, tells you insiders are choosing this price and this moment to reduce their stake.

Read generously, the upsizing simply met heavy demand and gave long-holding investors and employees liquidity, which is part of what going public is for; the roadshow was drawing crowds and the book was reported as many times oversubscribed. Read critically, the deal added supply and insider selling, at a raised price, days before a deceleration the company itself was about to flag, and the people with the most information were increasing how much they sold into the people with the least. The decision sat with chief financial officer David Ebersman, who owned the size-and-price call and would later be cast by CNBC and others as "the man behind the IPO debacle," a personalization the case should treat carefully: the choices were defensible on theory, which is exactly why blaming one executive misses the point.

One more public signal landed in the window and is worth weighing because it cut against the euphoria. On May 15, three days before the offering, General Motors said it would stop paying for display ads on Facebook, about $10 million of spend, publicly questioning whether the ads worked. It was small in dollars and large in symbolism: a marquee advertiser doubting the core product days before the company sold itself to the public at a perfection multiple. The case does not need to assign motive to record the structural fact: more stock, more insider supply, a higher price, and a fresh public doubt about the business model all arrived in the same few days.

The strategy the buy side was not built to like

Underneath the mechanics was a deliberate philosophy, and it is the analytical core of the whole deal. Most US IPOs are intentionally priced below where they will trade, producing a first-day "pop" of roughly 10% to 15%. That pop is not free money; it is value transferred from the company and its selling holders to the (mostly institutional) investors who received allocations. Facebook, advised by Morgan Stanley's lead banker on the deal and driven by its own chief financial officer, set out to minimize that transfer: price to the top, capture the proceeds for the issuer and the sellers, and leave the underwriters' clients almost nothing.

The same posture showed up in the fee. A syndicate of more than thirty banks worked the deal, led by Morgan Stanley with J.P. Morgan and Goldman Sachs, yet Facebook paid an underwriting discount of only about 1.1%, on the order of $170 million to $180 million in total, a strikingly thin gross spread for an IPO when smaller deals routinely paid the conventional 7%. A company with a hotly contested book has the leverage to compress both the pop and the fee, and Facebook used it on both. That is the through-line of the entire pricing story: every lever was pulled to move value from the intermediaries and their clients to the issuer and its selling holders.

By the logic of corporate finance this is defensible and even admirable: underpricing is a real cost borne by the issuer, and a company that eliminates it has, on that metric, executed well. The full economics of that trade-off are laid out in money left on the table at IPO. The natural counterfactual is Google in 2004, which disliked the same wealth transfer and tried to solve it differently, using a modified Dutch auction to let the market, rather than the banks' allocation list, set the clearing price; Google still rose about 18% on day one. Facebook's answer was blunter: keep the conventional bookbuild but push the price to the top and beyond the cushion. The objection is not that the theory is wrong but that it ignored a second job an IPO does, seeding a stable, confident shareholder base, and that pricing for zero cushion on a perfection multiple guaranteed that any disappointment would look like collapse rather than a dip. Both propositions can be true at once, and in this deal both were.

Get the technical reps behind a deal like this. Practice 1,000+ interview questions on IPO pricing, the bookbuild, and equity valuation in our iOS app: download it here.

The thirty-minute exchange failure

The first thing that went wrong was not Facebook's doing. Nasdaq runs an "IPO cross," a single auction that gathers all the opening orders and computes one opening print. On the morning of May 18, the volume of orders and last-second cancellations overwhelmed a poorly designed piece of that software, and the cross fell roughly nineteen minutes behind the orders actually arriving. The opening was delayed about half an hour, and tens of thousands of orders entered before the open were left in limbo, neither filled nor cancelled, for more than two hours.

The consequence was not abstract. For roughly two hours after trading began, investors who had placed orders did not know whether they had been filled, so they could not tell whether they owned the stock, had sold it, or were exposed to a position moving against them; confirmations did not flow normally until around 1:50 p.m. Several major market-making firms, including UBS, Knight Capital, Citadel, and Citi, absorbed the resulting chaos and later reported aggregate losses on the order of half a billion dollars, with UBS alone claiming losses in the hundreds of millions. Nasdaq's own remediation fund for member firms, on the order of tens of millions, did not come close to covering the claimed damage, which is itself part of the story: the venue's failure was an order of magnitude larger than the venue's willingness or ability to make participants whole.

The SEC ultimately charged Nasdaq and the exchange paid a $10 million penalty, the largest ever levied against an exchange at that point. One reading is a freak technical accident on unprecedented volume; another is that Nasdaq reached for a marquee listing it had not engineered its systems to handle and made poor real-time decisions when the software failed. The record supports the second more than the exchange would like, because the SEC's findings were specifically about systems design and decision-making under stress, not about bad luck.

A price defended, not discovered

The second thing that went wrong was quieter and more important. Facebook opened around $42.05, then sank toward the $38 offer price within minutes. It did not break $38 on day one, and the reason it did not is mechanical. The syndicate had sold about 15% more stock than the base deal, roughly 484 million shares against a base near 421 million, leaving the underwriters short the extra shares through the standard overallotment.

Greenshoe, overallotment and stabilization

Underwriters routinely oversell an IPO by up to 15%, creating a short position. A "greenshoe" option lets them buy that many extra shares from the company at the offer price to cover the short if the stock rises. If instead the stock sags, they cover by buying in the open market at or near the offer price, which both closes the short profitably and supports, or "stabilizes," the price. The stabilizing bid is legal and disclosed; it also means the first-day price is partly engineered, not purely discovered.

Because the stock would not hold, Morgan Stanley, the lead underwriter and stabilization agent, covered that short by buying heavily in the market at around $38 rather than exercising the greenshoe, which simultaneously defended the issue price and let the bank profit on the short it had sold higher. The legitimate description is textbook stabilization, exactly the mechanism described in first-day trading, the greenshoe and aftermarket stabilization. The uncomfortable description is equally true: the $38.23 close was not the market clearing the stock, it was the seller of the stock buying it back to keep the headline from saying "broke issue." The flat first day did not show resilient demand. It concealed that real demand was already below the offer price.

The May 9 amendment and the analyst cuts

On May 9, nine days before the IPO, Facebook amended its S-1 to state more pointedly that user growth was outpacing the growth in ads delivered, particularly because of the shift to mobile, and that this could weigh on revenue. It was a public filing, available to anyone. What happened next was not equally public. Analysts at the lead underwriters, including Morgan Stanley's internet analyst, cut their 2012 revenue estimates for Facebook, in aggregate from roughly $5.1 billion toward roughly $4.8 billion, in the middle of the roadshow.

Cutting an estimate after a company files a more cautious outlook is, by itself, ordinary analytical hygiene. The problem was distribution. The reductions were communicated by the banks' institutional sales desks to large clients by phone, in at least one instance the lead bank's internet analyst walking major accounts through the revised numbers directly, while the retail investors being marketed the same deal at the same price were left with the rosier picture in their heads and a dense risk factor they were unlikely to have re-modelled. Two different information sets walked into the same offering, sorted by client size, and the more cautious one went to the clients least likely to need the protection.

Selective disclosure and the bill for it

Regulators did not treat that as a grey area. The Massachusetts Securities Division fined Morgan Stanley $5 million for improperly influencing research analysts in connection with the offering, conduct it found violated the 2003 Global Research Analyst Settlement, the post-dot-com agreement under which the major banks had promised to wall their bankers off from their research to stop exactly this kind of selective, deal-serving analysis. Citigroup was separately fined over a junior analyst's handling of information around the deal. The investors' consolidated securities class action, in turn, did not allege the mobile risk was hidden; it alleged the opposite was the problem, that the offering was marketed while a materially more cautious revenue view was being circulated selectively, so the disclosure that existed was not the disclosure everyone received. That framing is why the case was legally durable rather than a complaint about a stock going down.

Morgan Stanley's defense, made publicly and not without force, was that all of this followed standard practice: the amendment was public, analysts may revise, and sales coverage is differentiated by client. That is precisely why the case matters. If the conduct that drew fines and litigation was, as the lead underwriter argued, normal, then the indictment is not of one bank but of the structure, and the Facebook IPO is the cleanest public example of it on record.

Go deeper before an interview: our 160-page PDF covers IPO mechanics, deal frameworks, and the technicals end to end, access the IB Interview Guide and pair it with these case studies.

The slide and the lock-up overhang

Once the underwriter stepped back, the engineered floor gave way. Facebook fell below $38 on its second day and kept going. The decline was then amplified by a predictable, mechanical force: lock-up expirations. Insiders and early investors had agreed not to sell for set periods after the IPO, and as each window opened, a wave of new supply hit a stock the market was already re-rating downward. Around the first major unlock in mid-August 2012, roughly 271 million shares became saleable and the stock fell under $20. By September 4, 2012 it reached an all-time low of $17.55, about 54% below the offer price.

The bears looked vindicated. The mechanics of why post-IPO supply lands when it lands are detailed in lock-up expirations and the post-IPO overhang; the case point is interpretive. Almost none of the 2012 collapse was driven by the business getting worse. It was sentiment plus a pricing that left no cushion plus a calendar of supply, which matters enormously for the verdict, because a stock that halves on mechanics and mood is telling you something very different from a stock that halves on deteriorating fundamentals.

Lock-up expiration

A lock-up is a contractual agreement, typically 90 to 180 days, barring insiders and pre-IPO investors from selling after the offering. When it lapses, a large block of previously restricted stock can hit the market at once. A staggered set of expirations, as Facebook had, delivers that supply in repeated waves, which can depress a stock for months independent of how the business is performing.

The pivot that rewrote the verdict

The thing the bears were most right about was also the thing that saved the company. Facebook had almost no mobile advertising revenue at IPO. Zuckerberg put the company on what insiders described as a mobile "lockdown," reorienting the product and the ad system around the phone. The first mobile ads appeared in August 2012. Mobile then went from essentially zero to a minority, then a near-majority, of advertising revenue: by the second quarter of 2013 mobile was about 41% of ad revenue, and rising. After the company reported that quarter in late July 2013, the stock reopened at $38.22 on July 31, 2013, back to its issue price for the first time in roughly fourteen months, and from there it compounded for a decade.

The magnitude is what makes the verdict bite. An investor who bought at the "disastrous" $38 and simply held did not merely recover; over the following decade the position multiplied many times over as Facebook, later Meta, grew into one of the largest companies in the world, a return that dwarfs the roughly 18% first-day pop the deal was criticized for not delivering. (That long-run figure is, by its nature, a moving and dated number; the point is the order of magnitude, not a specific quote.)

This is the detail that denies either side a clean win. The mobile risk the May 9 amendment sharpened, the risk the selective disclosure was about, was real: it genuinely depressed the stock for over a year. It was also early rather than fatal: the same risk, once addressed, became the growth engine that revalued the company far above the IPO. The IPO did not misjudge the business so much as the market misjudged how fast the single most important question would be answered, which is precisely why the people who called it a disaster and the people who called it perfectly executed were each half right.

For the issuer, it worked as designed

Judged by what an IPO is supposed to do for the company raising the money, Facebook's succeeded almost completely. It raised about $16 billion, priced at the top of a range it had twice lifted, took roughly $6.8 billion onto its own balance sheet, let long-term holders monetize at a rich price, paid a fee of only about 1.1% when the convention was far higher, surrendered almost none of the value that a conventional first-day pop hands to underwriters' clients, and kept founder control intact. On the corporate-finance scorecard, compressing both the pop and the fee is execution, not failure.

For everyone else, the process failed

By every other standard the deal was mishandled, and the costs were paid in cash and in rules. The exchange failed mechanically and the SEC fined Nasdaq $10 million; Nasdaq later settled a class action over the open for $26.5 million. The selective treatment of the estimate cuts drew the $5 million Massachusetts fine against Morgan Stanley and broader regulatory scrutiny of research conflicts. The securities class action, consolidated before Judge Robert Sweet in the Southern District of New York, was litigated for years and settled in 2018 for $35 million, an outcome later affirmed on appeal. Retail buyers who took the marketed story at face value were down roughly half their money within four months. None of that is excused by the company's later success; it happened, it was adjudicated, and it was expensive.

The verdict the record supports

So the honest verdict is split along the exact line the deal itself drew. The independent valuation bears, Damodaran among them, were right about the price for fourteen months and decisively wrong about it forever after. The process critics were right, and were vindicated by regulators and a federal court, that the open was broken and the information flow was unfair. The issuer-execution defenders were right that, as a capital raise, it did its job better than almost any large IPO of its era.

What the record settles, it settles firmly: the process failed and the failure was paid for, and the issuer's objective of maximizing proceeds with minimal underpricing was met. What it does not settle, and should not be forced to, are the two questions the case actually turns on. First, whether deliberately pricing for zero cushion is good practice or merely good arithmetic that ignores the cost of a shattered shareholder base, a debate the eventual recovery flatters but does not resolve. Second, whether a founder-controlled company that was mispriced for over a year and then vindicated proves the dual-class, capture-the-proceeds model works, or simply got rescued by a pivot that could as easily have failed. For a reader preparing to argue this both ways, the framing in how to discuss a landmark IPO in an interview is the relevant tool. The disciplined answer is not that the Facebook IPO was a disaster or a triumph. It is that it was both, on different axes, and that calling it only one is the mistake the case is designed to expose.

1. 1Facebook, "Facebook Announces Pricing of Initial Public Offering" (May 17, 2012).
2. 2US Securities and Exchange Commission, "SEC Charges NASDAQ for Failures During Facebook IPO" (May 29, 2013).
3. 3Aswath Damodaran, "The IPO of the decade? My valuation of Facebook" (February 2012).
4. 4The Washington Post, "Facebook IPO: Questions about disclosure swirl as stock stabilizes" (May 23, 2012).
5. 5CNN Money, "Facebook IPO: What went wrong?" (May 23, 2012).
6. 6TechCrunch, "Facebook Amends IPO S-1 To Admit Advertising Biz Hurt By Increasing Shift To Mobile" (May 9, 2012).
7. 7The Network, Berkeley Law, "Morgan Stanley Settles for $5 Million with Regulator for its involvement in Facebook IPO" (January 2013).
8. 8CNBC, "How Banks Cash In On Flailing Facebook Shares, Accidentally" (May 23, 2012).
9. 9TechCrunch, "Facebook Shares Open At $38.22, Finally Return To IPO Price" (July 31, 2013).
10. 10CNN Money, "Facebook's biggest lockup expiration" (November 2012).
11. 11Bernstein Litowitz Berger & Grossmann LLP, "Facebook, Inc. (IPO)" (settlement, 2018).