How to value Twitter?

Twitter Announces IPO: The Valuation (Originally Posted: 10/08/2013)

A little more than a week ago, I posted my first take on Twitter and argued that even in the absence of financial information from the company (since the prospectus had not been filed yet), you could price the company. Based on prior transactions in the company (VC infusions and acquisitions) and the multiples of revenues/users for other companies in the space (the social media medley, as I called it), I argued that Twitter would be priced at about $12 billion by the bankers. 

I also argued that as a long term  investor, focused on value, you could not buy the stock, at any price, unless you gauged its value first, and promised to return to the company and value it, once the filings were made. Twitter did file its S-1 (the pre offering filing) late in the day on October 3 and I am going to give it my best shot. Since I will reference this filing through my valuation, you should download the filing and use it to not only follow my estimates but to change those that you don't agree with. As with all of my valuations, I would hasten to add that this is my valuation and while it informs my decision on whether to buy or sell the stock, you should make your own best judgments about the company. (I know that this resembles the boilerplate disclosure that you see in every email that you get from your broker but I really mean this and I am not saying it to avoid getting sued.)

The state of the company

Before I embark on the ambitious task of forecasting the future, I will begin by looking at the company as it stands now. The financial filings paint the picture of a young company with little in terms of conventional earnings or cash flows to back it up, but plenty of promise (a dangerous word). Let’s start with the income statement. In the table below, I list the company’s key income statement numbers from 2010 to 2012, with the 2013 data in the last three columns.
The numbers paint a picture that should not be surprising. The company’s revenues have grown rapidly, with the revenues in the first 6 months of 2013 jumping more than 100% from the revenues in the first 6 months of 2012. Notwithstanding the jump, though, the revenues over the twelve months ending June 30, 2013, were only $448 million (which is well below the rumored $583 million that I used in my pricing post). The company has reported operating losses through its entire life, though to be fair, R&D expenses account for a chunk of the operating expenses. The company does report an adjusted EBITDA, and while the trend is positive, I cannot read much significance into a number that is based on the fiction that employee expenses can be added back because they are non-cash. Allowing for the fact that R&D is really misclassified capital expenditures (by accountants) and assuming a 3-year amortizable life for R&D does improve the operating income somewhat, as does the capitalization of operating lease expenses (moved to debt). Even with this improved operating income, Twitter’s pre-tax operating margin is barely positive (0.96%). 

Looking at the balance sheet just adds to the narrative of Twitter as a young, early-in-the-lifecycle company, as can be seen in the following table:
These numbers illustrate how completely useless accounting balance sheets  are at young firms, especially in technology. In fact, the only number in this balance sheet that has any relevance is the cash balance, with the intangible asset item measuring none of the true intangibles and goodwill playing its usual (and useless) role as plug variable. I capitalized their past R&D expenses and am showing it as an asset. The book value of equity is negative but as an investor, that tells me nothing, and the shift to a positive book value of equity in the pro forma statement matters even less. The company has little debt (a capital lease) and a series of convertible, preferred share offerings, reflecting its multiple VC capital infusions, which will get converted to equity on the public offering (removing a major headache in the valuation). Comparing Twitter’s financials to Facebook’s at the time of its public offering (see my valuation of Facebook on the day before its offering) should make it clear that Twitter is much earlier in its growth path than Facebook. 

The IPO set up

Valuing a company, just ahead of a public offering, is tricky for multiple reasons. The first is that there is a feedback effect from the offering itself, since the cash retained from the offering augments the value from the firm (if the founders don’t cash out). The second is that the share count is a key and shifting number, with conversions of other types of securities (preferred in this case) and shares used in employee compensation (restricted stock units (RSUs) and options) overhanging per share values.

1. Magnitude of the initial offering: While this may be putting the cart before the horse, the first number that you need to estimate is the dollar value that the company hopes to raise in the initial offerings. In most IPOs, only a small fraction of the company is offered and Twitter seems to be no exception. This news story suggests (and I tend to believe it) that Twitter plans to raise about $ 1 billion in the offering, at a stock price of roughly $20/share. Note that I am not suggesting that this is the right value for the share and it will really not affect my valuation.
2. Use of the offering: On page 16 of the S-1, Twitter specifies that it plans to retain the proceeds from the offering in the company and use it to cover investment needs (acquisitions, capital expenditures and working capital). That effectively means that the day of the offering Twitter’s cash balance will increase by roughly $1 billion, if that is the offering amount. (In some IPOs, the founders of the company cash out a portion of their ownership and take the offering cash out of the company. In that case, it would not augment the cash balance). 
3. Post-offering shares outstanding: As I noted in the last section, Twitter has a whole series of convertible preferred offerings. On page 17, the company notes that all of the convertible preferred shares will be converted into common shares, removing one potential entanglement. On the same page, the company specifies that it expects to have 472.613 million shares outstanding after the offering, but then proceeds to say that this number excludes 44.157 million employee options (potentially convertible to shares), 86.915 million shares of restricted stock units (also granted to employees), 0.117 million shares issuable on a warrant and 14.791 million shares to be issued to MoPub stockholders as payment for the acquisition. There is absolutely no valuation basis for excluding these shares and the total number of shares that I will use in my per share value is 574.44 million (472.613+86.915+0.117+14.791). The options also represent a claim on equity value, but I will deal with them separately. 

Valuing Twitter
The value of Twitter lies in what it can do with its 215 million users (the estimate in the S-1) rather than what it has done in the past. This is the section where I am sure that we will have to agree to disagree but the following sections summarize my assumptions.

A. Cash Flows/Earnings
1. Revenue Growth: The first leg of value creation for Twitter is for it to be able to grow its revenues out, from the $448 million in the most recent twelve months. To get some perspective on what the potential for revenue growth is in this sector, I started by looking at the latest assessments of the size of the online advertising market. While the estimates vary across sources, this one (from eMarketer.com) looks like it is close to reality (with the percent market shares and dollar revenues in millions):

There are two factors to keep in mind. The first is that the online portion of the advertising market is continuing to capture a larger share of overall advertising revenues (as attested to by the woes of print and traditional media companies); applying a 5% growth rate to 2013 online ad revenues yields a value of $190 billion for the overall market in 2023. The second is that there is a large segment of the market currently that is splintered among thousands of other companies, some conventional press media and many very small. Thus, the good news for Twitter is that there is a large potential market, but the bad news is that there will be plenty of competition from both the existing players and new entrants. In fact, one interesting and disquieting aspect of the inflation of market values of many of the companies on this list is that the market does not seem to be factoring in the finite size of the overall market. Thus, Google's current market cap implies that market expects its revenues will increase to $75 billion by 2023 and Facebook's market cap implies a revenue of $60 billion for that company; if the market is right, those two companies alone would account for 60-70% of the overall market in 2023.  I do think that Twitter starts with some advantages. While it does not have Facebook's user base (or expansive interface) and Google's easy reach, it does have a much-used and unique platform and an active user base. I will assume that Twitter's revenues will reach $11.5 billion in 2023. That will be more than a twenty fold increase in revenues and translate into a revenue growth rate of 55% for the next 5 years, scaling down to stable growth (of 2.7%) in year 11.
2. Target Operating Margin: Twitter's losses may be getting smaller and capitalizing R&D and lease expenses does make their operating margin for the last twelve months slightly positive (0.96%), but it is clear that Twitter's value as a company will eventually be determined by how much profits they can generate in the future. To make an estimate of the pre-tax operating margin that Twitter will be able to generate, once it gets through its growth pains, I took a look at a mix of firms that I would classify as the  social media medley: It is difficult to compare margins across these companies, since some (like Netflix) derive all their revenues from subscription revenues, some (like Pandora) have a mix of advertising and subscription revenues and some (like Google & Yahoo) are search engines. The company that is closest to Twitter in its advertising revenue model is Facebook and the company delivers an impressive 30% pre-tax operating margin, but Facebook's margin has shrunk as its revenues have grown. Will Twitter be as profitable as Facebook? There are news stories that suggest that Twitter gets less revenues from advertising per user than Facebook, but those may be reflective of where Twitter is in its growth phase. Twitter, with its 140 character limit, has a more constrained format for ad delivery but may work better in mobile advertising (which is the cutting edge of online advertising) than Facebook. Overall, though, I would anticipate Twitter to have a slightly lower operating margin (25%) than Facebook does now (30%), especially since Facebook's margins will also compress over time.
3. Reinvestment to deliver growth: Growth is never easy, nor is it ever free. With high growth companies, the tool I use to estimate reinvestment is the ratio of sales to invested capital (with higher numbers translating to more productive growth). To get a sense of what this number will look like for Twitter in the future, I took a look at Twitter's limited past and at Facebook's numbers: Note that for Twitter, I have computed the ratio of incremental sales to reinvestment each year from 2010 to 2012, and that my reinvestment number includes acquisitions, change in working capital capitalized R&D and is net of depreciation.  I have also computed the total sales to invested capital  for Twitter, Facebook and the sector in 2013. While Twitter's incremental sales to reinvested capital ratio has risen over time, it is still below the industry average. Put in intuitive terms, Twitter is spending large amounts (on R&D and acquisitions) to deliver its revenue growth and you have to hope for improvement as the company gets larger; Twitter specifically forecasts about $225-275 million in acquisitions for 2013 (S-1, Page 51). Based on these data, I assume that for every $1.50 increment in future revenues, Twitter will have to invest a dollar in capital; this allows me to estimate the reinvestment (including net cap ex, change in working capital, R&D and acquisitions) each year.
B. Risk/Cost of capital
 Is Twitter a risky company? Of course, but to estimate the rate of return that I would demand to cover its risk, I looked at three components:

1. Business mix: While the bulk of Twitter's revenues come from and will continue to come from advertising, Twitter does have a data trove of past tweets that may be mined for commercial or research reasons. In the first six months of 2013, Twitter generated 12.6% of its revenues from its data services and this proportion will probably decline in the future. Using a business mix of 90% advertising and 10% from data services yields a beta of 1.40 for the company. (Beta for Twitter = (0.90) (Beta for advertising) + 0.10 (Beta for data services) = 0.9(1.44)+0.1(1.05) = 1.40)
2. Geographic mix: While Twitter generated very little of its revenues from outside the US in 2011 and 2012, about 25% of its revenues came from the the rest of the world in 2013. Using an equity risk premium for the US of 5.75% and a GDP-weighted average equity risk premium of 7.23% for the rest of the world, with the current weights of 75% and 25% for each, yields a value of 6.15% for Twitter. Given that Twitter now has far more followers outside the US than in the US (S-1, Page 67), the proportions and equity risk premium may shift in the future. 
3. Financing mix: Twitter has capital leases of $71 million and the capitalized value of operating leases is $127 million. Collectively, debt accounts for 1.69% of capital and Twitter's cost of capital, given these assumptions, is 11.22%.

C. Loose Ends
Getting from the value of the operating assets to the value of equity per share in Twitter requires us to get over a series of speed bumps:

1. NOL & Taxes: The company's operating losses have resulted in a net operating loss of $468 million (S-1, page 217) which I use to shelter the company's income, when it does start making money, from taxes. As a result, I do not expect the company to pay taxes until year 5. Once it starts paying taxes, I assume that it will face an effective tax rate of 30%, which over time will move to the marginal tax rate of 35.50% (S-1, page 211) after year 10.
2. Cash & IPO Proceeds: I add up the cash and short term investments of the company (see page 173) to arrive at a cash balance, which is added to the value of the operating assets. Since I have assumed that the IPO proceeds will be $1 billion and that they will be retained by the firm, I add that value to the cash.
3. Capital & Operating leases: As mentioned in the risk section, I did convert the operating lease commitments of the company (page 214) to debt and added it to the capital leases to arrive at a total value of debt of $299 million.
4. Survival Risk: While young companies with operating losses are susceptible to failure, I will assume that Twitter's deep pocketed equity investors will bring in capital, if the company gets into trouble, rather than leave value on the table. I have assumed that there is a 100% chance of the company surviving.
5. Option overhang: The company has 44.16 million options outstanding, with a strike price of $1.82 (S1- page 207). Halving the remaining life on these options, to reflect the empirical reality that employee options get exercised about halfway through their lives, gives me a life of 3.47 years and in conjunction with an estimated standard deviation from the company of 53.6% (see page 207) yields a value of $805 million for these options (net of tax benefits in the future).
6. Classes of shares: The best news I see in this filing is that there is no mention of two classes of shares (with different voting rights) or special corporate status (which I did see in the Facebook filing). While that may change in future revisions, that does make my job easier in terms of estimating value of equity per share.

The net effect of these adjustments is to get a value of equity of $9.97 billion for the equity in common stock and a value per share of $17.36. The picture below captures the various assumptions in the valuation and you can download the spreadsheet with the valuation.

If you do not like my assumptions, please change them and come up with your own estimate of value. If you are so inclined, please do enter your numbers in the Google shared spreadsheet that I have created.
Decision Time
Having learned from the Facebook fiasco, I  expect the bankers and the company to make the Twitter IPO a smoother offering. That process will of course start with the road show, where they will package the company like a shiny new present, and unwrap their “offering” price. I am sure that Goldman’s bankers, working on this deal, are a capable lot and will price the stock well, with just enough bounce to make those who receive a share of the initial offering feel special.  As I watch the frenzy, I have to remind myself of two realities. The first is that there will be lots of distractions (like this one) during the IPO, most designed to take my eye of the ball. The second is that the bankers have their own agenda, and I cannot make the mistake of assuming that it matches mine. Watching out for my interests, here is how I see Twitter: at a $6 billion market cap ($10/share), I think it is a very good deal, at $10 billion ($17.5/share), I am indifferent to it, and at $20 billion ($35/share), it is a moon shot. Could I be wrong? Of course, but I would rather be transparently wrong (hence the long blog post detailing every assumption that I made) than opaquely right. I welcome disagreement (though I would much appreciate your phrasing it agreeably).