This weekend I talked to a friend of mine who analyzes website traffic. I scrapped most of the original draft of this article and replaced it with my friend’s input on Quantcast data, LinkedIn (LNKD) user habits and website advertising revenue. Also, I figured out the actual lock-up expiration day. Enjoy.
I am not shorting LinkedIn because the valuation is “crazy.” I am shorting LinkedIn because the company is misunderstood. Apparently a lot of people own LNKD because it is the closest thing to Facebook. This view is patently flawed. My report will not go into too much detail over the hiring solutions part of the business (42%+ of revenue), because I think that business is defensible over the long run. My concern centers on LinkedIn’s value to advertisers. Specifically, I do not believe LinkedIn’s “Marketing Solutions” business (one-third of revenue) provides its customers with much benefit; and therefore I do not believe that recent growth in the segment should be extrapolated. LinkedIn deserves a valuation closer to Monster Worldwide (MWW) than Facebook.
Where Are The Active Users?
A website’s value to advertisers is directly proportionate to its number of active users. Facebook is extremely valuable because people log in a few times a day to stalk attractive girls and to see what their friends are up to. Out of the 800 million registered users, about half log in every day!
My hunch was that LinkedIn has very few active users. I, for one, have logged in twice this year. Let’s investigate the Form 424B4. A search for “active user” and “active member” yields no results. A search for “visitor” brings us to this very interesting disclosure underneath some charts depicting growth in page views:
The number of registered members is higher than the number of actual members due to various factors. For more information, see ‘Risk Factors—The number of our registered members is higher than the number of actual members, and a substantial majority of our page views are generated by a minority of our members.’
OK, that first sentence is confusing. There is some elaboration in the Risk Factors section:
The number of registered members in our network is higher than the number of actual members because some members have multiple registrations, other members have died or become incapacitated, and others may have registered under fictitious names. Given the challenges inherent in identifying these accounts, we do not have a reliable system to accurately identify the number of actual members, and thus we rely on the number of registered members as our measure of the size of our network. Further, a substantial majority of our members do not visit our website on a monthly basis, and a substantial majority of our page views are generated by a minority of our members. If the number of our actual members does not meet our expectations or we are unable to increase the breadth and frequency of our visiting members, then our business may not grow as fast as we expect, which will harm our operating and financial results and may cause our stock price to decline.
My hunch was right! LinkedIn admits that it has very few active users. What I find interesting, though, is the vagueness of the admission. The Risk Factor disclosure has two separate points: first, that there are more accounts on LinkedIn than there are actual people behind them; and second, that a substantial majority of page views are generated by a minority of members. These are two totally unrelated pieces of information. And the second piece is far more important (relevant to advertisers) than the first. Why are they combined into one bullet point? A forensic statement analyst would say that the company is trying to mislead the reader. It is de-emphasizing a major shortcoming by preceding it with a minor one.
Since LinkedIn is intentionally vague about its active users, we have to go somewhere else for more detail. There are various services that compile data on website traffic and visitor demographics. I prefer to use Quantcast because it offers a paid service and needs to stand by its data. (Since Alexa doesn’t sell any services, it has no incentive to put out accurate information.)
Here is Quantcast’s report on LinkedIn.
The relevant section here is Traffic Frequency. Users are categorized as (1) Addicts–those who visit or interact with a site 30 or more times per month, (2) Regulars–those who visit more than once but less than 30 times per month and (3) Passersby–those who visit once per month. As you can see, only 1% of LinkedIn’s visitors are addicts. It is hard to convince advertisers that your site is valuable when 99% of visitors sign in just a couple times each month.
Compare this to Facebook. 76% of users account for 97% of visits! These are stats advertisers love, not even taking into account all the rich personal information on Facebook that advertisers can use to tailor their advertising for maximum effect.
Despite the paucity of active users, LinkedIn has shown impressive growth in ad revenue:
Net revenue from our marketing solutions increased $27.6 million, or 116%, as a result of an increase in customers’ overall online advertising budgets and improved productivity from our field sales organization.
I don’t understand why advertisers are stupid enough to pay for ads on a site with so few active users. I mean, look at Yahoo! (YHOO) and AOL (AOL). On virtually all metrics, they are far more attractive to advertisers than LinkedIn. They are having a terrible time growing their display ad revenue, and the market has re-rated their stock prices downward. Yet LinkedIn’s display ad revenue is exploding upward. What can I say? Luck can work especially well off of a small base, and LinkedIn has some skilled ad *cough* snake oil *cough* salespeople. I don’t think this side of the business has much of a long-term future, because I have faith that advertisers will realize that there are many superior alternatives to LinkedIn.
Which advertisers have taken the bait? Just go to LinkedIn.com and check the little buttons in the right sidebar. I see one ad for MIT’s Executive MBA program and another for some database of small businesses for sale. On the homepage, there’s a display ad for vodka. Who the hell clicks on this stuff?
The Top Sites Puzzle
So I have more or less proved that LinkedIn is a bad deal for advertisers. But there was one more puzzle to solve. According to Quantcast, LinkedIn is the 22nd most visited site in the United States, based on the number of unique visits per month. I asked a web traffic analyst how this could possibly be. His response verbatim (since I can’t do any better):
As far as LinkedIn being the 22nd ranked site–it’s by uniques, which is a notoriously difficult number to get a read on. For one, it’s monthly uniques. So if you visit the site once in a month, you’re part of the uniques count for that month.
That said, if I were pressed to identify a reason why, I’d say it’s because of their enormous direct mailing list. Their website started to boom in Quantcast in January of this year, which is when they started to leverage their large but largely inactive user base (i.e., people like us who have accounts but don’t visit the site very much). I usually delete those emails, but it wouldn’t surprise me if they’ve based most of their “growth”, as it were, on these users who sporadically click the email links and visit the site.
Also, any time anyone searches a person on Google, that person’s LinkedIn profile (as well as Facebook) is going to (likely) be the most popular hit. So they’ll get lots of uniques that way.
“In the end, as far as advertising goes, this type of traffic is not worth much. What they want is the kind of traffic that Facebook gets.”
In the past month I have received more emails than usual from LinkedIn. Usually they alert me when a connection (someone I don’t know but for some reason have “friended” on the site) posts an update. I always delete the email. It’s junk to me. Apparently other people are more prone to clicking the links.
I hope advertisers are smart enough to realize that monthly uniques stats don’t matter nearly as much as the number of active users.
Speaking With One Of LinkedIn’s Largest Shareholders
I wanted to hear the bull case on LNKD. So I called an acquaintance of mine who is a managing partner of one of LinkedIn’s largest investors (the name of which will stay anonymous). All I needed answered was one question: How can LinkedIn monetize its site so that it can deserve a valuation of 25X revenues?
The partner answered with a sentence containing so many adjectives that it was barely coherent English. He must have used “compelling” twice in one sentence. There were also some common tech-related nouns you often hear VCs talk about: “platform,” “penetration” and “installed base.” In a few minutes, I figured that this partner had no more insight on LNKD than I did. I generally assume that other market participants understand situations better than I do, until I can make a smart argument to prove otherwise. Here, I was talking with someone who was discussing a lot of unknowables. This guy had just proved to be a very smart reseller, just as every venture capitalist is. They invest early in a concept, making the story sexier as it grows. Then they time the IPO during an overheated market and BAM! They’re rich! The hot potato passes from the VC to the naive public. Unfortunately the public has no one else to pass the potato to. If it burns, it really burns.
LinkedIn’s Core Business
I think LinkedIn will have more luck with “hiring solutions” than “marketing solutions.” Here’s the mumbo-jumbo:
Revenue from our hiring solutions is derived primarily from the sale of our LinkedIn Corporate Solutions and LinkedIn Jobs products. LinkedIn Corporate Solutions allow enterprises and professional organizations to identify job candidates based on industry, job function, geography, experience, education, and other specifications. Our customers can also purchase job slots to utilize job postings on our website throughout the contract term. We recognize the net revenue from sales of LinkedIn Corporate Solutions ratably over the subscription period, which is typically 12 months and billed annually, quarterly or monthly. We also sell LinkedIn Jobs on our website to enterprises and professional organizations of all sizes. These jobs are generally posted for 30 days, and revenue from individual job postings is recognized over the same period.
So LinkedIn is paid to be a headhunter. Except this headhunter uses the power of the Internet. It’s like a Craigslist of skilled people (not of the adult services variety). I have nothing negative to say here. It will probably work well over the long run. But there’s one concern. I hope bulls aren’t extrapolating growth rates from the S-1, because:
Net revenue from our hiring solutions and premium subscriptions increased primarily as a result of higher productivity from our expanded field sales organization as well as an increase in the number of subscribers of our premium subscriptions. To a lesser extent, net revenue from our hiring solutions increased as a result of the fact that we had a full year of sales of LinkedIn Corporate Solutions in 2009, as we had only recently introduced that solution during the last half of 2008. (Emphasis added.)
When Is The Actual Lock-up Expiration Day?
Most LinkedIn bears incorrectly believe that lock-up expiration is November 14, because this is 180 days after the IPO. However, page 129 of the S-1 reveals:
The 180 day restricted period described in the preceding paragraph will be extended if during the last 17 days of the 180 day restricted period we issue an earnings release or material news event relating to us occurs, in which case the restrictions described in the preceding paragraph will continue to apply until the expiration of the 18 day period beginning on the issuance of the earnings release or the occurrence of the material news or material event.
Since LinkedIn reports earnings on November 3, the actual date of lock-up expiration is November 21.
Oversupply Of Shares
LNKD represents the Class A shares, of which there are 7.8 million outstanding. There are also 86.7 million Class B shares, out of which 42.9 million (45% of the shares outstanding) are eligible to be registered, converted, and sold as Class A shares after the 180-day lock-up expires. For more detail, here is a paragraph from page 29 of the Form 424B:
After this offering, the holders of 42,941,315 shares of Class B common stock, or 45.4% of our total outstanding common stock, based on shares outstanding as of March 31, 2011 and giving effect to the sale of shares by the selling stockholders, will be entitled to rights with respect to registration of these shares under the Securities Act pursuant to an investors’ rights agreement. Shares of our Class B common stock automatically will convert into shares of our Class A common stock upon any sale or transfer, whether or not for value, except for certain transfers described in our amended and restated certificate of incorporation to become effective upon completion of this offering. If these holders of our Class B common stock, by exercising their registration rights, sell a large number of shares, they could adversely affect the market price for our Class A common stock. If we file a registration statement for the purposes of selling additional shares to raise capital and are required to include shares held by these holders pursuant to the exercise of their registration rights, our ability to raise capital may be impaired. We intend to file a registration statement on Form S-8 under the Securities Act to register up to approximately 35.0 million shares of our common stock for issuance under our Amended and Restated 2003 Stock Incentive Plan, 2011 Equity Incentive Plan and 2011 Employee Stock Purchase Plan. Once we register these shares, they can be freely sold in the public market upon issuance and once vested, subject to a 180-day lock-up period and other restrictions provided under the terms of the applicable plan and/or the option agreements entered into with option holders.
The 42.9 million-share overhang not only dwarfs the current float, but the shares short (about 2.3 million) as well. If even a small portion of insiders decide to take the money and run, there will be a severe oversupply of shares. It is unlikely that enthusiastic buyers will come in to absorb this. Since the market turmoil began in late July, Groupon (GRPN) has decided to cut its implied IPO valuation in half. Facebook shares on the secondary market have declined from $38 to $31 (and this is an artificial market because anyone who bought shares in the secondary market cannot turn around and sell). In fact, the past two Facebook auctions have failed to clear any shares, because the average bid ($28.47) is too low. LinkedIn insiders should be a lot less sanguine about their investment than they were just three months ago. Besides, they have the opportunity to realize one of the biggest investment returns in recent memory.
Series E, Oct 2008
Bessemer Venture Partners
Series D, Jun 2008
Bain Capital Ventures
Bessemer Venture Partners
Series C, Jan 2007
Bessemer Venture Partners
European Founders Fund
Series B, Oct 2004
Series A, Nov 2003
I’m not aware of any other funding rounds. If this is all the money insiders put into the business, they are sitting on some of the largest gains ever scored by venture capitalists. Sequoia owns 17% of the shares outstanding equal to $1.5 billion. It put up less than $30 million! How’s that for an amazing IRR? The CEO Reid Hoffman still owns 21% of the stock. That makes him a paper billionaire. If insiders are able to get out at just the IPO price of $45, there will be dozens of multimillionaires in Mountain View. I am betting that their greed will kick in.
Risks To Being Short
LinkedIn announces earnings after the bell on November 3. The conspiracy theorist in me fully expects that it will pull some shenanigans. Perhaps it has deferred hiring this quarter, or otherwise cut expenses. The consensus estimate is that it will report a loss of $0.04 cents. I am expecting a huge headline beat. The stock may spike to $120.
To see what the “other side” is thinking, here are the comments of Dan Zanger, a swing trader, regarding LNKD:
Earnings due out any day now will propel the stock one way or the other. Buy area is a move above $92 after earnings.
This gives you a sense of how much this might run up in case of a short squeeze. My observation over the years has been that Dan Zanger’s picks outperform in the very short run, but they underperform over the long run.
Whose Bread I Eat His Song I Sing
Another interesting factor is that Morgan Stanley (MS) (owning a 23% stake in LNKD) is about to take Groupon public. The investment bankers will use the public market valuations of companies like LinkedIn and Pandora (P) to justify Groupon’s offering price. One might argue that Morgan Stanley is motivated to wait to sell its LinkedIn stake until Groupon has gone public (the underwriting fee should amount to $35 million). Morgan Stanley is in an interesting position because it has the right to sell LNKD shares at any time, as it is not subject to the lockup period. To my knowledge Morgan Stanley has yet to sell a single share, and most of its stake was acquired on the open market (not pre-IPO). They may be engendering the second Internet Bubble. (Remember Mary Meeker?)
LNKD went public when the market was a voting machine, but now the market is a weighing machine. Over this time period, many other expensive stocks (Netflix (NFLX), OpenTable (OPEN) and Green Mountain (GMCR)) have collapsed. LNKD has miraculously held up. I believe it is a good short candidate because the advertising part of the business is not an attractive value proposition for advertisers compared to other websites. This is not anything like Facebook! With an impending share overhang, I believe we will finally see LNKD weighed more appropriately in the market.
Disclosure: I have a short position in LNKD call options.