"Twitter provides a compelling and efficient way for people to stay informed about their interests, discover what is happening in their world right now and interact directly with each other."
But as Twitter continues its march to a November 15th offering date, there's a storm brewing that just might end this love affair. The Company's recent filing raises a whole new series of questions about its strategy, business model, AND accounting that probably won't get answered before its initial public offering (IPO). There just isn't enough time left for capital market regulators to force answers (thanks to the JOBS Act), nor is management likely motivated toward transparency given the nature of the issues. As so aptly put by Dennis Berman in "The Big Hole in Twitter's IPO Filing:"
"In Wall Street parlance, Twitter remains inscrutable, a "black box." It's beautifully-lacquered, with a bow on top. But still a black box."
So What Is Twitter's Strategy?
Despite exhaustive and sometimes rambling narratives on value propositions for users, platform partners, advertisers and data partners, and discussions of the Company's growth strategy and related risks, Twitter never explicitly states its singular corporate strategy. This is a real problem! All I want to know is how the Company creates value for its customers, and differentiates itself from its competitors. After all, it is such a strategy that drives an entity's business model, related performance metrics, and ultimately, valuation.
To make sure I hadn't missed something, I searched the S-1 for the word "customer," and the results were revealing. Apparently, Twitter feels that "customer" is a dirty word when it comes to its own business. Yes, the term was used several times (S-1; 7, 104, and 105), but only when referring to how data partners and advertisers utilize Twitter data in developing and selling products to their customers. The word also appeared in the Risk Factor section (S-1,17) in the context of "customer service to users," yet users were not explicitly identified as "customers." So, are users "customers" or not?
Then, there are the standard boilerplate risk disclosures (S-1, 34) which use "customer" in a variety contexts (e.g. with users, affiliates, suppliers, and distributors). Finally, Twitter slips and acknowledges formally that its customers are its advertisers who purchase advertising services (S-1, 112). So, there we have it…Twitter creates value by providing advertising services! But as Berman points out, we really don't have much customer data:
"Not a peep on advertiser numbers, quality, or renewals. Not a peep on the sales force, except to state that there will be a "significant increase" to the sales budget that exceeded $77 million for the first six months of 2013. Not a peep showing the actual effectiveness of Twitter's advertising, save for stating that more effective ads will bring in more money."
Now, what about "differentiation?" How is Twitter different from its social media competitors involved in the advertising space? Well, word searches for differentiation yielded no "hits" at all, but a slight variation did yield the following:
"Twitter's public, real-time, conversational and widely distributed content and our differentiated ability to target users through their Interest Graphs enable advertisers to promote their brands, products and services, amplify their visibility and reach, and complement and extend the conversation around their advertising campaigns (page 106). "
Despite this reference, there is not a hint as to what the "differentiated ability" might be. This is troubling given that Twitter acknowledges "mobile" as the primary driver of its advertising business (S-1, 2), and that it competes:
"Against many companies to attract and engage users, including companies which have greater financial resources and substantially larger user bases, such as Facebook (including Instagram), Google, LinkedIn, Microsoft and Yahoo!, which offer a variety of Internet and mobile device-based products, services and content (S-1,18)."
So, there you have it…Twitter is in the business of essentially creating customer lists and selling contact access to advertisers. Not exactly the Andy Rooney business plan, but it can work. It's a fairly simple idea…but can the Company make money at this, particularly given its competitive and scale disadvantages? Maybe its business process will provide some clues.
What About Twitter's Business Model?
Again, the Company provides little insight into how it intends to execute its strategy. Neither "business model," "business process," nor "value chain" are mentioned even once in the S-1, and the only models explicitly discussed are the "asymmetric follow model" and the "Black-Scholes option pricing model." So, exactly how does Twitter carry out its market analysis; research and development; sales and marketing, procurement production, and distribution; and after-sale customer service functions? We are provided nary a clue, other than that:
"Our products and services incorporate complex software and we encourage employees to quickly develop and help us launch new and innovative features. Our software has contained, and may now or in the future contain, errors, bugs or vulnerabilities (S-1, 30)."
Twitter's admission of existing process problems, and its cavalier attitude toward "future vulnerabilities" are big concerns, particularly given the Company's inorganic "growth strategy." In the past 30 months, Twitter has made over 20 acquisitions, and added over 1800 employees since January 1, 2010. What's the likelihood that these deals have been successfully and seamlessly integrated, and that information systems and financial reporting controls are functioning efficiently and effectively? You know the answer.
To its credit, Twitter does acknowledge the risks associated with managing growth (S-1, 26). But in doing so, the Company also buries a warning in the fine print about future costs:
"Providing our products and services to our users is costly and we expect our expenses to continue to increase in the future as we broaden our user base and increase user engagement… our expenses may grow faster than our revenue, and our expenses may be greater than we anticipate. Managing our growth will require significant expenditures and allocation of valuable management resources."
And Twitter also provides another interesting nugget that warrants investigation (S-1, 69):
"Many of the elements of our cost of revenue are relatively fixed, and cannot be reduced in the near term to offset any decline in our revenue."
Well, as you might expect, this Grumpy Old Accountant just had to investigate what effect the Company's existing cost structure might have on profitability. After all, if you're considering investing in a company with net losses and negative operating cash flows, wouldn't it be nice to know how far Twitter is from "breaking even?"
Given the lack of cost data available for Twitter, I used the High-Low Method of Cost Estimation and the Company's 2011 and 2012 revenue and expense data to estimate fixed and variable costs, as well as breakeven point. The results are presented below:
The results suggest that last year's variable costs accounted for 76.1 percent of each revenue dollar. Armed with this data, I computed variable costs for 2012 to be $241,201 (76.10% times $316,933). I then deducted this amount from total costs and expenses of $394,016, which yielded a fixed cost estimate of $152,815. The Company's total revenue breakeven point is then calculated by dividing estimated fixed costs by the contribution margin percentage (1 minus 76.10%, the variable cost percentage). The result: Twitter needs revenues of almost $640 million to breakeven, assuming its cost structure remains relatively constant at 2011 and 2012 levels. That means the Company has to more than double 2012 revenues to turn a profit, while keeping costs under control.
Sounds possible, right? Wrong! Don't forget the Company's earlier statement that "our products and services incorporate complex software." Complexity means higher costs. And since Twitter prioritizes "innovation and the experience of users and advertisers on our platform over short-term operating results" (S-1, 28), it is more likely that the Company's inorganic growth strategy will contribute to increasing fixed costs as well, and even higher breakeven points. So much for the Company's business model.
Are There Any Accounting or Financial Reporting Mysteries?
Of course there are…after all this is a technology-based, social media company! Twitter avails itself of many of the usual internet company tools that muddy disclosure and transparency which the Grumpies discussed in "The Beauty of Internet Company Accounting." So, let's dig in. For the most part, I will be ignoring the unaudited results presented in the S-1, as they simply can't be relied upon…not that PwC's audited numbers are likely to be much better (see "Is FASB Killing the Auditing Profession?").
Balance Sheet Concerns
Much of the Company's asset growth over the past several years has been driven by non-cash transactions related to such activities as acquisitions and capital leases. The supplemental non-cash disclosures in the Company's consolidated statement of cash flows detail these transactions (S-1, F-9). Particularly interesting is the slight reduction in stockholders' deficit, which is largely the result of the equity issued by the Company in its acquisitions. Since the beginning of 2011, Twitter has financed over 87 percent of its acquisition purchases by issuing its own common and preferred stock. The result: the Company has been able to partially offset the negative effects of net losses on stockholders' equity by "doing deals." And equity issued for 2013's acquisition activity actually improved the stockholders' deficit.
What's the big deal you ask? To answer this question, I provide the following acquisition activity summary for Twitter:
The first thing that should get your attention is what Twitter actually received for its stock, mostly intangibles (developed technology) and goodwill. Now here's the problem. Where did the valuations assigned to the developed technologies come from? Equally troubling is where did Twitter get the valuations for its stock? Although the Company does provide a narrative on how it values its common stock (S-1, 86), at the end of the day, we all know management relied on paid consultants and financial advisers, and investment bankers for these valuations. I am having flashbacks to the S&L crisis of the 1980's when thrift managers inflated their balance sheets using questionable valuations. Those of us with gray hair remember a joke back then…what's the first thing you need to do if you open an S&L? Buy yourself an appraiser! Things haven't changed all that much, have they?
This all means that Twitter may have significantly overstated both assets and equity, all under the auspices of today's generally accepted accounting principles (GAAP). In these acquisition transactions, both the seller and acquirer have incentives to get the purchase price as high as possible. The seller's incentives are obvious, but Twitter's may be a bit more devious. First, there is the positive effect on stockholders' equity that I previously discussed. But Twitter also has incentives to keep its equity prices as high as possible and rising, as it nears its IPO. The Company's statement of stockholders' equity (S-1, F-7 and F-8) reveals acquisition common stock issuance prices per share of $8.11 in 2011, $18.39 in 2012, and $17.58 for 2013. It's probably not a coincidence that almost all of the recent acquisitions have been for stock at very high purchase premiums that created large amounts of goodwill. So, there is a very good possibility that the balance sheet is overstated by inflated intangible and goodwill amounts. And of course, there are no impairment charges despite the absence of positive operating cash flows or earnings. Is that because the asset impairment tests are based on the equity valuations prepared by the Company's paid consultants? Yes, that's a real grumpy thought…
And the acquisition accounting problems don't end there. There's also the issue of potentially unrecorded liabilities for contingent consideration associated with Twitter's recent acquisitions. The Company has committed to paying contingent consideration on several of its acquisitions (S-1, F-19 to F-21), but it's unclear as to how much of this liability (if any) has actually been reported in its financial statements. Finally, Twitter's subsequent events note (S-1, F-47) appears to have omitted discussion of two other acquisitions which occurred in August of 2013, Trendrr and Marakana, even though it did discuss September's MoPub transaction.
Also very interesting is the Company's "mixing" of its tax valuation allowance accruals. The Company admits that its net deferred tax assets (DTA) of $12.247 million require a full allowance at 2012 (S-1, F-42). However, instead of reporting the appropriate income tax expense to increase the valuation allowance for these net DTA, Twitter argues that its reported liability for uncertain tax positions of $12.156 million covers this net DTA overstatement. So, essentially it is using its uncertain tax position liability as a pseudo valuation allowance for its net DTA. Not only is this very interesting, it is not at all common. The problem with this practice (in addition to the obvious balance sheet misclassification) is that this treatment's net effect is correct only IF the net DTA are related to the Company's uncertain tax provisions…and Twitter has not made that link. Warning…there may be a $12 million DTA write-down in the Company's near future. These financial statements were audited, right?
And what would any Grumpy Old Accountant rant be without a complaint about lease capitalization. Yes, Twitter reports significant operating lease commitments, just NOT on its balance sheet (S-1, F-44). So, using the Company's commitment data, I made a couple of assumptions to estimate Twitter's off-balance sheet lease liability. I used a discount rate based on NYU Professor Damordaran's cost of capital estimate for Twitter (i.e., 11.22 percent). Also, "thereafter" lease obligations were assumed to occur equally over the remaining five years, and tax effects were ignored. Even with these "crude" assumptions, the Company's estimated unrecorded lease liability approximates $111 million, which represents 13.35 percent of total assets at the end of 2012. And the liability to assets ratio jumps from 24.92 to 33.76 percent when the unrecorded lease obligations are considered. But GAAP is what it is…by the way, the liability is likely even greater as operating lease commitments have swelled to over $221 million according to the unaudited June 30, 2013 numbers.
Income Statement Questions
One big surprise is Twitter's use of multiple element accounting (S-1, 83). Why the Company finds it necessary to "enter into non-standard sales agreements" (those are Twitter's own words) in the first place is a mystery. Nevertheless, in these arrangements, management has significant discretion as to when it records revenue:
"When neither vendor-specific objective evidence nor third-party evidence of selling price exists, we use our best estimate of selling price (BESP) to allocate the arrangement consideration on a relative selling price basis to each deliverable."
It's this "best estimate" that's the problem.
Also, the Company is aggressively recording costs as assets to defer expense recognition, all perfectly legal under today's flexi-GAAP. Examples include software and website development costs (S-1, F-17) and acquired developed technologies (F-18 through F-21). It is particularly interesting that Twitter has dramatically extended the amortization period for acquired developed technologies during the past three years from 12 months in 2011 to over 40 months for 2013 acquisitions. This practice decreases reported period costs for these questionable "assets," and increases operating earnings.
Finally, there is Twitter's use of non-GAAP metrics (S-1, 13-15). As you may recall, recently I voiced my objection to such measures in "Non-GAAP Metrics: Is it Time to Toss Out the SEC's Reg G?" But here we go again. In discussing its incentive compensation plan (S-1, 137), the Company suggests that it does have real performance measures available. These include sales bookings, product defect measures, product release timelines, productivity, return on assets, return on capital, return on equity, return on investment, return on sales, time to market, total stockholder return, and working capital. Why aren't these metrics disclosed? Instead, the Company employs the increasingly common "adjusted EBITDA" disclosure to transform reported GAAP net loss into a profit, and now introduces "Non-GAAP net loss." Twitter says that it uses these measures to help investors see its "operating results through the eyes of management." Now, that's a scary thought. Remember, Facebook didn't feel compelled to use these crazy metrics in its IPO.
So there you have it. Just when you thought you couldn't get any more drama from a social media company, along comes Twitter. Undifferentiated strategy, unclear business model, and uncontrolled accounting and reporting.
Yet, I still love Twitter…but only as a user. Maybe Comcast (CMCSA) or some other big entertainment giant will come along and buy Twitter at post-IPO lows, and keep my Twitter ad-free…one can only hope.
Disclaimer: This essay reflects the opinion of the author and not necessarily that of The American College, or Villanova University.