Sign In

Communications of the ACM

Technology strategy and management

Reflecting on the Facebook IPO

Reflecting on the Facebook IPO, illustration

Credit: Anthony Freda

What should a company be worth? This simple question is difficult to answer and became a heated topic of debate with the public debut of Facebook in this past May. The social networking giant, now with close to one billion users, opened with an initial public offering (IPO) stock price of $38 and a market value of $104 billion. These numbers quickly fell 25% within a week.5,6 Then Facebook disappointed investors with its first earnings announcement in late July, showing "only" 32% growth over the prior year. The company had higher costs and took $1.3 billion in charges related to pre-IPO stock compensation. Analysts also worried that too many users were accessing Facebook on mobile devices, resulting in lower growth rates for ad revenues as well as reduced operating profits. Not surprisingly, the stock price dropped another 20% or so overnight.8 Where the company's market value will end up, no one knows. But how low or high should the IPO stock price have been?

Most new software and Internet services companies pay little or no dividends, do not place much emphasis on profits, and invest mainly in intellectual capital rather than physical assets. Therefore, traditional metrics used to value a company, such as dividend yields, price-to-earnings ratios, or return on assets, do not work well. Nonetheless, investment bankers will suggest a value, investors will consider whether or not to invest, and analysts will comment after the fact. One columnist for The Wall Street Journal, for example, argued that Facebook's initial price should have been no more than $13.80—if investors were to achieve average stock market returns (11% annually) and a Google-like market valuation-to-sales ratio five years from now (currently 6 or 7 to 1 compared to 15 times revenues at its 2003 IPO). The lower stock price implies Facebook should have been worth at IPO no more than $38 billion.2

Zynga and Groupon have also fallen steeply from their IPO prices, giving many investors reason to pause: Facebook and its entire generation of social media stocks may all have been overvalued. My purpose in this column is not to speculate on prices but to make two more-general observations: First, I can see no way to determine the short-term market value of a company, before or after IPO, with any type of "scientific" precision. And second, investors can still use some "objective" metrics to consider whether a company might become a good long-term investment. To make this judgment, we need to understand the underlying economics of the business—how a company makes and spends money, as well as how fast it is growing relative to its peers.

Facebook and its entire generation of social media stocks may all have been overvalued.

For example, Table 1 compares Facebook with several companies frequently in the news—Microsoft, Google, Oracle, Apple, Infosys, SAP, IBM, Salesforce, and LinkedIn. The data is from the prior fiscal year, with market values as of January 2012 except for Facebook, for which I use its IPO value from May 2012. The order is determined by the second column, which is operating profit percentage (profits from ongoing operations, before taxes, divided by sales—excluding extraordinary income like sales of real estate or a business division). Every year I compile such a list and Microsoft nearly always comes out on top—at least partially the result of selling not hardware but packaged software products, which comprise 90% of revenues (the last column in Table 1). On this measure of operating performance, Facebook does extraordinary well—beating Microsoft, 47% to 39%. (Of course, we now know these numbers excluded the $1.3 billion in stock-related charges that Facebook took in the first quarter after its IPO!)

At the bottom is the professional social networking firm LinkedIn, a prominent 2011 IPO and the smallest firm on the list in sales. It has only 4% operating profits. But a young firm should be spending heavily on salespeople, ads and marketing campaigns, and new products and services. Operating profits represent money left over after these expenses and the direct cost of sales (as well as some other expenses). It would not be surprising for LinkedIn to rank low in operating profits if it has been investing in growth.

The next column gets a lot of attention in business schools and the business press: gross margin. This represents sales minus the direct cost of making those sales. In conventional manufacturing, like automobiles or semiconductors, this is an important number because it is costly to bend metal, mold plastic, or transform silicon wafers into microprocessors. It is also revealing in professional service businesses like Infosys or the consulting divisions of SAP, Oracle, and IBM because it represents the direct costs (mostly personnel) of delivering those services. In software products or Internet services, this is not the case. The marginal cost of replicating any digital good is close to zero, though Internet services can require big investments in server farms. But, in general, companies that do not rely on physical products or labor-intensive services should do well on this metric, and this is what we see.

LinkedIn leads in gross margins (84%), followed by (80%), the pioneer in Software as a Service (SaaS), and then Microsoft (78%) and Facebook (77%). The laggard is Apple (41%), which also trailed Microsoft, Google, and Oracle (in addition to Facebook) in operating profit rate. Apple's innovative iPhone, iPad, iPod, and iMac products are at the core of its strategy, and it can still charge a premium. But the low gross margin reflects that Apple's direct costs of making products are much higher than at software or Internet service companies. The more revenue Apple generates from automated digital services, software sales, or transaction fees through iTunes, App Store, iCloud, iBooks, and iAds, then the more its gross margins are likely to rise.

What about market value times prior year sales? LinkedIn's IPO price in May 2011 was $45 per share, with an initial market value of $4.25 billion.7 This was 17 times 2010 revenues of $243 million. A year later, LinkedIn still leads with a valuation of 13 times 2011 revenues ($522 million). is next (9x), while at the bottom we see IBM (2x) and Microsoft (3x). Even Apple is only 4x—remarkably low for a company that grew sales 66%. Why is that?

LinkedIn's valuation makes more sense when we look at the next column—sales growth rate (versus the prior year). Here again, LinkedIn leads at 215%, though 2010 revenues were small. Facebook again impresses at 88%. By contrast, IBM lags at 4%. Low growth and low operating profits (19%) help explain why IBM's market value is so low.

Another metric is productivity—here defined as total sales divided by year-end employees. Our star is Apple ($1.7 million per employee), followed by Google and Facebook (each with $1.2 million per employee). But we need to look at other metrics to interpret this number. For example, comparing sales productivity with gross margins indicates if a firm has high sales costs and production expenses—which Apple does, making its high nominal sales productivity less impressive.

An IT services company like Infosys should struggle in sales productivity and it does ($46,000 per employee). It can compensate by paying low wages and using lots of people to generate revenues. In fact, Infosys does this and had impressive gross margins (42%, better than Apple) and a notably high operating profit rate (29%, merely 2% behind Apple). Investors apparently see this and high growth (26%) and value Infosys (5x sales) more than Apple, Oracle, and SAP (4x) or Microsoft (3x) and IBM (2x). But Infosys' dependence on labor-intensive services (95% of sales) suggests that future growth will be difficult. Historical data suggests Infosys grows revenues at approximately a 1:1 ratio to increases in employees. In 2011, Infosys had 131,000 employees and sales of $6 billion. It also reported employee attrition of 17%.3 To grow sales again by 26% in 2012, assuming the same sales productivity, Infosys must hire 34,000 employees plus another 22,000 to replace departures!

Other numbers help explain why gross margins can be high and operating profits low: general expenses that come after the direct cost of sales (but still before taxes). For software products and SaaS companies, and for Internet service companies like Google and Facebook, these expenses can be huge. spent a whopping 63% of 2011 revenues simply on sales, marketing, and general and administrative expenses. Add to this 11% of sales spent on research and development (R&D) and we get 74%. Combined with a 20% cost of sales (the inverse of the 80% gross margin), we can see why managed a profit rate of only 6% (see Table 2). The company was indeed investing heavily to achieve a 27% growth rate.

Similarly, LinkedIn spent 46% of revenues on sales, marketing, and general and administrative expenses, and 25% on research and development: 71%. This total, combined with 16% direct cost of sales (100 minus 84% gross margin), and another 9% of sales in other expenses,4 explains why LinkedIn shows just 4% of sales as operating profit. But investors seem to like LinkedIn nonetheless, possibly due to high growth and because it does not simply rely on advertising. Fifty percent of last year's revenues came from job placement fees paid by corporate clients and another 20% from premium subscriptions ("product sales").4

We can also understand Apple's operating profit rate of 31% more clearly, as well as its low valuation. Apple's growth has been so high (66%) that revenues have probably outpaced prior-year plans for expenses, leaving a surprisingly low percentages for sales, marketing, and administration costs (7% of sales) and research and development (2%). Only Infosys (2%) was in Apple's territory for R&D spending. But, as a service company, with merely 5% product sales, we expect Infosys to do little research and development. We expect Apple to do a lot, and its expense ratios will rise quickly if and when growth slows. Perhaps investors do not believe Apple can continue to grow so fast so cheaply.

The obvious unknown is how much and how fast Facebook will continue to grow, and how much that growth will cost.

The last column in Table 1 points to percentage of revenues coming from new product sales. The balance of revenues comes mainly from services or maintenance, including software license renewals. New product sales or subscriptions in software and some hardware businesses (like the iPhone) have high gross margins and high growth potential without adding costly headcount or expensive factories. It is a useful metric for Microsoft, Oracle, SAP, and, and even for Apple and IBM, but has little meaning for Google and Facebook. Their "products" are free automated services like searching the Internet or connecting with friends. What they sell is primarily advertising. Facebook logged some virtual goods and technology sales (mostly to Zynga), but these amounted to only 15% of revenues.1 SaaS companies are hybrids with a product delivered and priced like a service, with some technical support and maintenance bundled into a monthly fee. But they still can have very high gross margins and scale economies. This is why and LinkedIn should have high valuations as long as they are growing fast. But their high expense ratios are worrisome for the future.

Back to Top


The Facebook IPO price seems very expensive in retrospect and assumed a continuation of extremely high growth (and profit) rates. But there really is no way to determine what the price should have been in advance. On "objective" measures like sales growth as well as operating profits, gross margin, sales productivity, and expense ratios, the company looked very strong prior to the IPO and solidly in the class of Microsoft, Google, Apple, and other elite performers. The obvious unknown is how much and how fast Facebook will continue to grow, and how much that growth will cost. User behavior with mobile ads is apparently different and less profitable than with PCs, and Facebook will have to adjust. Investors must also adjust their expectations as they speculate about the future. For example, if Facebook can average 32% annual growth for the next five years, and if the stock price and market cap recover to the IPO level, then the multiple will be a Google-like 7x in 2016. Facebook will then seem much less expensive. But these are big "if's." The bottom line? There may be no science to IPO pricing, but even a quick look at the economics of the business, with some reasonable extrapolations and comparisons to peer firms, does shed some light on what a company is likely to be worth in the future.

Back to Top


1. Facebook, Inc. Form S-1. United States Securities and Exchange Commission, Washington, D.C., Filed February 1, 2012.

2. Hulbert, M. Facebook's stock should trade for $13.80. The Wall Street Journal (May 25, 2012);

3. Infosys Technologies Ltd. Form 20-F. United States Securities and Exchange Commission, Washington, D.C., Fiscal Year Ending March 31, 2011.

4. LinkedIn Corporation. Form 10-K. United States Securities and Exchange Commission, Washington, D.C., Fiscal Year Ending December 31, 2011.

5. Raice, S. and Letzing, J. Parsing the Facebook fallout. The Wall Street Journal (May 30, 2012), B5.

6. Saitto, S. et al. Playing the Facebook blame game. BusinessWeek (May 28-June 3, 2012), 45–46.

7. Selyuk, A. and Baldwin, C. LinkedIn IPO prices at $45, but risks real. Reuters (May 18, 2011);

8. Sengupta, S. Facebook shares plummet in an earnings letdown. The New York Times (July 27, 2012), B1.

Back to Top


Michael A. Cusumano ( is a professor at the MIT Sloan School of Management and School of Engineering and author of Staying Power: Six Enduring Principles for Managing Strategy and Innovation in an Unpredictable World (Oxford University Press, 2010).

Back to Top


T1Table 1. Comparing Facebook with other companies.

T2Table 2. Examining the profit rate.

Back to top

Copyright held by author.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2012 ACM, Inc.