Disclaimer: I shorted (bet against) Castlight Health on April 1 at $23.21 and held this position until Friday, April 11 when I exited at $16.99. I exited to lock in my winnings, but as this essay explains, Castlight is still overvalued and I may short them again. Any reader unfamiliar with “short selling” stocks should reference Investopedia’s introduction to short selling.
Additionally, it’s hard to complete these writeups in my evenings in a timely manner while the market continues to move. I began writing a week and a half ago when I shorted Castlight and had a number of friends, such as Andrew Virata, Reece Arthur, Ben Gilbert and others, aware of Castlight. This note is here to add credibility so it doesn’t look like I’m writing the analysis after the stock has already moved in my favor. Additionally, I have uploaded a zip folder where you can download Castlight’s income statement, balance sheet, and cash flow statements in Excel format and the PDF of their public FORM 424B4 filing from which most the research was derived.
“We expect to continue to incur operating losses for the foreseeable future and may never become profitable on a quarterly or annual basis, or if we do, we may not be able to sustain profitability in subsequent periods. As a result of these factors, we may need to raise additional capital through debt or equity financings in order to fund our operations, and such capital may not be available on reasonable terms, if at all. “ – Castlight Health, Page 13 of Form 424B4.
Castlight Health is an internet business which went public on March 17, selling 11.1 million shares at $16 per share, raising approximately $161.2 million for the company ($177.6 million from the public stock sale, but a $12.4 million fee for the bankers). On the first day of trading, the stock rocketed 149% to $40 per share, valuing Castlight at over $4 billion.
In 2013, this business lost $62 million on sales of $13 million. Since it went public a month ago, the stock has dropped over 50%. This is where I start my analysis of what Yahoo has deemed “the most overpriced IPO of the century.” I do not disagree.
What is Castlight Health? It is a web app for enabling employers and employees to easily view and manage their health care and insurance plans. It’s leading application is called Castlight Medical which they say, “simplifies health care decision making for employees and their families by providing highly relevant, personalized information for medical services that enable informed choices before, during and after receiving health care,” and “enables employees and their families to intuitively search for robust and comprehensive information about medical providers, including personalized out-of-pocket cost estimates, clinical quality, user experience and provider demographic information.”
How does one lose $62 million in one year building a website for health data? And how is it still in business?
Castlight’s Income Statement
“I would say there is no one here who can understand some new internet company. I said at the annual meeting this year that, if I were teaching a class in business school, on the final exam I would pass out the information on an internet company and ask each student to value it and anybody who gave me an answer I’d flunk.” – Warren Buffett’s Q&A at the University of Florida in 1998
When I first noticed Castlight and it passed my eyeball test as a terrible stock, I wanted to see if I could possibly value the company. I quickly relented.
Castlight filed its Form 424B4 paperwork to the SEC on March 14 which includes all the required information for a company selling public stock. This document contains the company’s historical performance going back to 2011. Typically a financial analyst will use these SEC filings to do a Discounted Cash Flow analysis (link to my introductory tutorial on the subject), but with a company burning through as much cash as Castlight (and many other internet businesses) the DCF analysis quickly breaks. The primary problem is that, unlike a more mature company, or even a barely profitable one, you can’t even begin to predict the company’s future with any degree of confidence or reliability. The two sub-problems when making those future predictions are projecting future revenue increases and cost decreases. In looking at Castlight, I realized I had few ideas on how either of those line items would look over the next five years.
Here is their slightly organized income statement (numbers are in millions):
The last row, Castlight’s earnings, paints a pretty grim picture of not just three consecutive years of losses, but increasingly large losses. While revenue (row three) has grown substantially, costs (rows five and nine) have grown faster.
Castlight is an enterprise software company, so initial development and sales expenses are likely to be higher than mainstream consumer software. Yet after three years, these numbers clearly show a company that’s been lighting dollar bills on fire with matches in front of large customers for a pittance.
Page 82 of Castlight’s filing lists its top customers: Wal-Mart, Microsoft, Eaton, Indiana University, Microsoft, Mondelez, Purdue, Safeway, Honeywell and others. While this is a seemingly impressive group, they only totaled $13 million in revenue. If you can’t become profitable with Wal-Mart and Microsoft as your customers (two firms with a large number of employees), how many customers and how large do your future orders need to be? By Castlight’s own admission, Wal-Mart accounted for about $2 million of the $13 million in revenue under a contract which expires in 2015. Are there potential customers bigger than Wal-Mart who are willing to pay more than $2 million? With $58 million in selling and administrative expenses, it’s unclear if Castlight has figured out how to sell services at a higher price than it costs to do the selling (I could have phrased this much less glowingly).
In Castlight’s defense, they dedicate page 83 to “Customer Case Studies”, examples of their application saving companies or their employees money on health insurance. For Honeywell, “employees who used Castlight to shop for laboratory services paid 14% less than those who did not search (February 2012 through September 2013).” This would be more impressive if it weren’t preceded by: “Castlight developed a health care management platform that connected Honeywell health care vendors in a single integrated solution and offered employees a consistent user experience and message.“
Although Castlight Medical is a web application, their business model is not entirely Software-as-a-Service. A large component of their revenues is from “professional services”.
Professional Services is their consulting business of customizing its software suite for each client, which they say typically takes three to twelve months. While there are highly successful consulting businesses, the profit margins on such an employee-heavy business model are typically worse than a pure software business.
In its own documentation, Castlight almost creates a paradox. On page 50 under the section “Key Factors Affecting Our Performance”, the company says about its Professional Services business:
“We believe our professional services capabilities support the adoption of our subscription offerings. As a result, our sales efforts have been focused primarily on our subscription offering, rather than the profitability of our professional services business. Our professional services are generally priced on a fixed-fee basis and the costs incurred to complete these services, which consist mainly of personnel-related costs, have been greater than the amount charged to the customer…. These factors contributed to our gross loss percentage from professional services of (249)%, (596)% and (739)% in 2011, 2012 and 2013, respectively. The increase in gross loss percentage in 2012 was due to non-recurring professional services fees. The increase in gross loss percentage in 2013 was a result of an increase in the number of customers and complexity of our customer implementations. We expect to continue to generate gross losses on professional services for the foreseeable future as we focus on adoption of our subscription offerings.”
That can be summarized as: Professional Services loses money. We have no intention of it directly making money, but using it to up-sell our customers into profitable subscriptions.
That’s not so bad a proposal, until page 51’s “Costs and Operating Expenses”:
“Cost of professional services consists primarily of employee-related expenses associated with these services, the cost of subcontractors and travel costs. The time and costs of our customer implementations vary based on the source and condition of the data we receive from third parties, the configurations that we agree to provide and the size of the customer. Our cost associated with providing implementation services has been significantly higher as a percentage of revenue than our cost of providing subscriptions due to the labor associated with providing implementation services…. We expect to continue to generate negative gross margin on our professional services for the foreseeable future. As our implementation processes and technologies mature and our use of automation increases, we expect our gross margin on our professional services to improve.”
And back on page 50 under the “Revenues” section: “We expect professional services revenue to constitute a significant portion of our total revenue in the future.“
This reads to me as, “Professional services is going to generate a lot of our revenue, but because it’s so personally customized at big upfront costs for each client, it’ll create even greater losses, even if our margins improve.”
Speaking of margins, they are usually easier to improve when you’re not in a competitive industry.
Page 22 highlights the industry’s competitive landscape:
“While the enterprise health care cloud market is in an early stage of development, the market is competitive and we expect it to attract increased competition, which could make it hard for us to succeed. We currently face competition for sub-components of our offering from a range of companies….These competitors include Truven Health Analytics Inc., ClearCost Health, Change Healthcare Corporation, Healthcare Blue Book and HealthSparq, Inc. In addition, large, well-financed health plans, with whom we cooperate and on whom we depend in order to obtain the pricing and claims data we need to deliver our offering to customers, have in some cases developed their own cost and quality estimation tools and provide these solutions to their customers at discounted prices or often for free. These health plans include Aetna Inc., Cigna Corporation, UnitedHealth Group, Inc. and WellPoint, Inc. Competition from specialized software and solution providers, health plans and other parties will result in continued pricing pressures, which is likely to lead to price decline in certain product segments, which could negatively impact our sales, profitability and market share.”
The fundamental premise behind Castlight is valid; web applications that allow companies and their employees to more easily manage their health insurance and doctor relationships will be the future. This is not an original concept to the industry’s existing players or other startups, which they clearly admit here.
Also dangerous is that there is little stopping health insurance providers from cutting off their data supply to Castlight if they deem it a competitor to their business or cutting into their revenues from companies/end users. Castlight already acknowledges this: “If health plans perceive continued cooperation with us as a threat to their business interests, they may take steps that impair our access to pricing and claims data, or that otherwise make it more difficult or costly for us to deliver our offering to customers. “ It’s possible for the insurance companies to offer this service themselves. An agnostic third-party application is a good idea, but the clear competition and clear stranglehold the data providers have on them makes it a tough business that could have its valuation multiples suppressed.
There is one potential positive to support Castlight’s valuation: Its “backlog” of customer contract agreements to be fulfilled in the future. My reading of their filings points to the backlog information as one of the few indicators of potential revenue.
Castlight explains what constitutes its backlog on page 60: “At any point in the contract term, there can be amounts that we have not yet been contractually able to invoice. Until such time as these amounts are invoiced, they are not recorded in revenue, deferred revenue or elsewhere in our consolidated financial statements and are considered by us to be backlog.”
Potential future deals that haven’t been recognized in their numbers seems promising. The actual backlog size is , as of December 30, 2013, $108.7 million for the “total backlog”, but only half of it, $50.9 million is non-cancellable. Perhaps they give the $108.7 million number to give people a reason to be optimistic, but it seems near-meaningless to me. The $50.9 number is interesting as its a lot larger than last year’s $13 million in revenue and could indicate huge growth for Castlight.
Except Castlight says they do not “expect to fulfill our non-cancellable backlog as of December 31, 2013 over a period of approximately four years, with the substantial majority expected to be fulfilled after 2014.” $50.9 million spread over four years, with most of it in the last three, does not indicate much, if any, growth. Castlight admits as much: “Accordingly, we believe that fluctuations in our backlog may not be a reliable indicator of our future revenue.”
I was hoping the backlog might help them out, but it seems pretty useless except for Castlight to say that some potential customers are kinda sorta considering paying them between $50 million and $100 million over the next four years. And who knows how much it’ll actually cost Castlight in Selling and General Administration expenses to service those contracts.
After having highlighted many of Castlight’s numerous negatives, one may be tempted to try and bet against them (as I am). Despite Buffett’s statement, it could still be worth a back-of-the-napkin valuation so as to have an idea of when the stock may stop its decline.
Since Castlight lacks any significant cash flows to discount, our best guess would be to compare it to similar companies and see if Castlight is overvalued relative to them. The two I will use are BenefitFocus and athenaHealth. Warning that for the purposes of this essay, these valuations are very rough and guaranteed to be inaccurate. I don’t worry too much about the inaccuracy of these valuations considering the company doesn’t make any money.
BenefitFocus is a web application “that enables its employer and insurance carrier customers to more efficiently shop, enroll, manage, and exchange benefits information.” The company went public last September. They are valued at about $1 billion with 2013 revenue of $104 million and $-30.36 million in net income. Still not profitable, but large revenues and valued at about ten-times those revenues. Castlight has $13 million in revenue with -$62 million in net income, so about a tenth of BenefitFocus’s revenue with twice as many losses at three times the market valuation for the company. By using the same 10x revenue multiple as BenefitFocus, Castlight would be valued at about $130 million. In fairness to Castlight, their revenues have been growing at a faster rate the past few years than BenefitFocus, so we’ll double that multiple to 20x revenue and the valuation to $260 million. The stock price would be about $1.60 at a market capitalization of $260 million. The price is currently $18.09.
Now we’ll use athenaHealth, a health care and health insurance software company which offers, “cloud-based services are packaged as four integrated offerings: athenaCollector for revenue cycle management, athenaClinicals for clinical cycle management, athenaCommunicator for patient cycle management, and athenaCoordinator for referral cycle management.” athenaHealth went public in 2007 and is profitable with a mere $2.59 million in net income last year but $18.73 million in 2012. athenaHealth deserves its own research, as it’s currently trading at a price-to-earnings multiple of 8,000-plus. Using the same revenue multiple approach used with BenefitFocus, athenaHealth trades at about 9.2-times its annual revenue ($595 million in revenue and current market value at $5.47 billion). This 9.2x is pretty close to BenefitFocus’s and should arguably be higher considering athenaHealth is profitable and has grown revenues by about $100 million every year for the past five years.
Castlight is currently valued at over 225 times its annual revenue. From a Yahoo! Finance interview: “Jay Ritter, a professor at the University of Florida and my go-to source on IPOs for the past few decades, tells me that Castlight’s insane level of valuation – 107 times revenue (not profits, as they had huge losses last year) – of the original IPO pricing hasn’t been seen for a tech deal since the year 2000, the twilight of the 20th century. Of the prior 13 deals priced at 100 times revenue or more and sales of at least $10 million, the average 3-year return was -92%.”
I’ll reiterate that you can refine these multiples all you’d like or use other metrics, but you can’t refine Castlight’s revenue. Professor Ritter’s stat is the most damning evidence against Castlight’s stock: The average three-year return of companies as overvalued as Castlight going back a decade is -92%.
Venture Capitalists Cover Castlight’s Losses
If Castlight hasn’t made any profits and lost over $100 million over the past three years, how has it stayed in business? With the help of $177 million in investor funding from 2009 to 2012, of which there was only $67 million left going into last year. Crunchbase includes a quick list of its investors and the size of each investment round. Prior to IPO, but at the end of 2013, Castlight had $67.17 million in cash and short term investments that could be liquidated into cash. Their losses for 2013 were $62 million, so they had about one year’s worth of cash left. The IPO raised an additional $161.2 million for the company, so they have about $166 million in cash now, minus whatever they’ve spent in 2014. This is also probably off by plus or minus $20 million, but it should be in the ballpark.
The investors have primarily been Venrock (the largest shareholder and a large Silicon Valley venture capital group), Oak Investment Partners, Maverick Capital, Fidelity Investments, The Wellcome Trust (a British charity endowment specializing in healthcare funding), T Rowe Price, and Morgan Stanley (also one of the investment banks who helped take Castlight public).
Given their history of increasing expenses more than they increase revenue, we can reasonably expect Castlight to spend somewhere in the range of $60 to $100 million this year. Their current cash reserves should last them two years before they will have to raise even more money, either through a secondary stock offering, sell bonds, or a credit lines with a bank. I guess they could accomplish that through increased sales too, but I’ll reiterate that for Castlight increased sales means increased costs.
Castlight’s Two-Class Ownership Structure and Management Compensation
With all these big investors, who owns Castlight?
Pages 113 and 114 list the major owners of the company and their total voting power for company decisions following the IPO:
- Venrock and affiliates own 20.6% of the class A shares and 18% of the voting power.
- Oak Investment Partners owns 15.8% of the class A shares and 13.8% of the voting rights.
- Maverick Capital owns 10.2% of the class A shares and 8.9% of the votes.
- Fidelity Investments owns 9.8% of the A shares and 8.6% of the votes.
- The Wellcome Trust owns 8.7% of the A shares and 7.6% of the voting rights.
- CEO Giovanni Colella owns 8.2% of the A shares and 7.2% of the votes.
- Chief Operating Officer Randall Womack owns 1.7% of the A shares and 1.4% of the votes.
Combined, these majors shareholders control 75% of the A shares and 65.5% of the management votes.
What do I mean by “Class A” shares? Well, the public offering is of “Class B” shares. Castlight has two types of shareholders: “Class A” a.k.a “Those with lots of power” and “Class B” a.k.a “Those with little power”. Public investors are Class B.
What’s the difference? Right on the very first page of Castlight’s SEC filing, they explain: “We have two classes of common stock, Class A common stock and Class B common stock. The rights of the holders of our Class A common stock and Class B common stock are identical, except with respect to voting and conversion rights…. Immediately following the completion of this offering, outstanding shares of our Class A common stock will represent approximately 98.6% of the voting power of our outstanding capital stock.”
I’m pretty sure voting rights are an important right and the legalese here, in context of everything I’ve already shown, comes across as flippant.
Pages 9 and 116 elaborates on the rights of these different classes. Class A shareholders have the majority vote on the sale or merger of the company, the sale or lease of property and assets, the dissolution of the company, changes to the certificate of incorporation, or “every matter” if an outside individual or investment group has or announces intent to buy 30% of the Class A and B stock combined. Even outside of these issues, one Class A share counts for 10 times as many votes as a Class B share. There are only 11.1 million B shares compared to 75,469,707 Class A shares.
I have to note that this setup is actually relatively common with public companies (where there are two or more types of stock with certain investors having more control than others). It just looks a lot worse for those Class A people when the company doesn’t look healthy.
The list of investors and executives above control the decision making of the company. But if they’re making good ones for the business, the class B public investors could be okay with the setup.
Or the Class A investors could just let management, some of them former employees at the investment firms, pay themselves large sums of money, as shown in this chart from page 100:
That’s right. In the last year, the top five executives were paid $965,156 in salary and total compensation of $6,183,012. This is during a year when the company lost $62 million and only had sales of $13 million.
The reasonable argument is that there is a job market for executives and, given their resumes, the Castlight team could get similar or better offers elsewhere. But these individuals are surely already rich given their previous jobs as venture capitalists for Greylock, having sold previous health care companies, and worked in executive capacity for other billion dollar public companies (page 92 lists the biographies of these five executives). I do make an exception for Head of Product Dena Bravata, who has been a practicing doctor and I assume has made good, but not outlandish sums of money prior to Castlight.
But view the situation from a different perspective: What kind of system pays people hundreds of thousands of dollars for losing tens of millions? Only a venture capital funded business could operate this way. They would presumably argue that sometimes you will need to take losses for a while in the early stages to grow the business. I retort that once you have a multi-billion dollar valuation, you are not in the early stages, even if the industry itself (healthcare IT) is not yet fully mature.
Amazon is the common exception people cite for the burn-money-for-years-to-dominate-an-industry, but Amazon’s stock was abysmal for years, going from above $100 in 1999 to $5 in 2001, and 2001 is when you should have bought their stock. Don’t buy Castlight while it’s expensive. If you believe in them, buy them after their stock collapses and they show some signs of success.
Having hopefully made the case that Castlight is a horrifically overvalued company that has offered no prospects for investors to earn a return on their investment (other than for the investment banks and their hand-picked clients who received the initial stock), who is responsible for the inevitable shareholder losses?
I don’t blame Castlight’s management. I am giving them the benefit of the doubt that they believe they have the solution to health care management for employers and employees, and the public investment markets are willing to invest in them, they are incentivized, if not obligated, to raise more money while they can.
The reality is much more grey with blame to be partially owned by everyone. I’m at best an armchair philosopher, but surely there is a moral/ethical argument to be made against Castlight’s management taking public money at multi-billion dollar valuations to pay executive salaries and stock options which put them in the top 1% of earners while losing tens of millions annually just because they can. Someday I will have better philosophical expertise to make a more foolproof assertion.
More blame falls next to the IPO underwriters and brokers: Goldman Sachs, Morgan Stanley, Allan and Company, Stifel, Canaccord Genuity, and Raymond James. In our post-SOX, IPO starved world, investment banks are starting to sell whatever they can now that investors have cash they’re willing to put to work, stock market returns looking more promising than any other asset class, and distance from the original dot-com bubble. Companies come to them with the willingness to be taken public. Bankers compete for the opportunity to support the company’s public offering and get their cut of the deal. Taking companies public is their job. Yet it seems short-sighted of them to sell obviously overpriced companies to public investors. The public has no chance of earning a return on its investment, but the banks and their closest clients get the public’s money. Over time, these kinds of offerings degrade the public’s trust in markets, as we saw in the dot-com boom and bust. The Goldmans of the world claim to have standards that separate them from the scams such as the one depicted in The Wolf of Wall Street. But when the public markets want to invest in stocks and new businesses, the big banks lower their standards to create supply to meet the demand. All this does is put public money into businesses that fail, losing even more money for everyone involved except bankers and executives.
Lastly, the public and the institutions which represent them. Why do we let companies, venture capitalists, and banks get away with making money from selling us stock in companies that lose money? I suspect it’s for two primary reasons: People do not take the time to learn how the financial industry and markets work and they put too much trust in the institutions managing their money without knowing what they are doing either. There is probably academic research measuring those two points. For the institutions like mutual funds, pension funds, and insurance companies, they are looking to make money from investing their cash from the public. New stocks may hold the potential to grow into the next Microsoft, Exxon, or Walmart. Both the institutions and the individuals they represent want to believe this is still possible, and invest accordingly in new stocks with their cash and hope.
That’s the real root of Castlight, the current market boom, and most investment bubbles. We all want to believe the future is brighter, regardless of the present conditions. And it’s usually true that technology will continue to, paraphrasing Gordon Gekko, “mark the upward surge of mankind.” What is lost is that the future is brighter because of today’s work. Though they may help, no amount of high finance or VC funding replaces the work that goes into sustainable solutions to hard problems. That’s what a healthy economy is: sustainable. Currently, Castlight and many internet companies of its kind are not.
If you have any questions, feel free to email me at firstname.lastname@example.org.
UPDATE: Monday, April 14, 2013:
Jim Cramer directly addresses technology IPO oversupply in 2014 and lists Castlight Health as an example of a disappointing IPO harmful to the market. Maybe he reads my blog.