Twitter: How Tweet it Is!
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Artur Pylak
FOR IMPORTANT DISCLOSURES 203.901.1633 /.1634
sagawa@ / email@example.com
October 8, 2013
Twitter: How Tweet it Is!
- TWTR’s S-1 affirms its considerable growth and profit potential. Sales have more than doubled YoY, driven by 44% growth in users, 16% growth in visits per user, and 23% growth in ad revenues per visit. We believe that all of these metrics will continue to improve on a steep trajectory, given TWTR’s unique and advantaged position as the primary Internet distributor of time sensitive information and its detailed insight into its users interests. Immediate opportunities from TV/media partnerships and a record of innovation with its product, such as Twitter Cards that allow users to embed rich media files, give us confidence that sales deceleration from its current triple digit pace will be gradual. Moreover, the successful launch of Vine hints at opportunities to expand the applications available to the TWTR user base and create future vehicles for monetization. We also believe that gross margins will expand from the current 64% to better than 80% as the business matures, and that positive EBITDA will grow to positive net earnings quickly as sales outpace investment in R&D and marketing. Reasonably assuming a 3 year annual growth rate of 60-65%, mature operating margins of ~25%, and applying a FB/LNKD trailing sales multiple suggests a year end 2016 valuation of $50B.
- TWTR is a unique info distribution service, NOT a social network. Twitter’s business is asymmetric, delivering news, commentary, rich media content and links from the community of newsmakers – e.g. media outlets, journalists, government agencies, artists, businesses, executives, celebrities, etc. – to a much broader community of followers. It is uniquely valuable because of its immediacy, its focus, its ease of use, the scope of information available, and the scale of audience that can be reached. In most of the developed world, there is no viable alternative and economy of scale and network effect barriers make market entry daunting to would-be rivals.
- User growth still on steep curve with plenty of runway. TWTR has 218M monthly average users (MAU), up 44% YoY. 49M of those users were in the US, up 22.5. For comparison, FB’s 1.15B MAUs were up 21% YoY in June, with 142M in the US and Canada, up about 9% YoY. 75% of TWTR’s MAUs access the service on mobile devices, suggesting penetration to just 26% of American smartphone owners, and 12% for the rest of the world. With the increasing public profile of the service as it is heavily promoted through TV tie-ins, advertising, sports and other venues, we expect adoption to continue apace for the foreseeable future.
- Engagement can also increase. TWTR measures engagement by timeline views, which rose 68,5% YoY worldwide and 45% in the US during the June quarter. The average views per user are more than 7.6 per day globally up 16% YoY, with Americans visiting more than 9 times a day. Several factors should continue to drive views higher going forward – i.e. growing penetration into smartphones, closer tie ins to TV and sports, increasing rich media content, improved apps, and growing public awareness.
- Advertising density can increase considerably. TWTR measures its ad density as advertising revenues per 1,000 timeline visits. On this metric, ad density rose 23% YoY to $0.80, or roughly $0.18 per MAU per month. TWTR restricts the inventory of placements available to advertisers in its auctions and charges its advertisers based on the level of user engagement with its ads. We believe that the amount of advertising served to TWTR users could increase significantly, noting FB’s success in raising its ad load to levels far above TWTR. Subjectively, TWTR ads integrate smoothly into the timeline, fit naturally into the context of an information distribution service, and are typically well-targeted to user interests – a product of TWTR’s “Interest Graph” data base of who each user follows and how they engage with the tweets that they receive. Moreover the Amplify program promises to add well-accepted pre-roll ads to desirable videos from TWTR’s media partners. As such, we believe that TWTR is less likely than FB to experience user push back on rising ad density.
- Ad pricing is likely slipping today, but should improve long term. The S1 points out that TWTR’s ad prices are somewhat lower for mobile placements, and that mobile is increasing as a percentage of its total. Mathematically, this will sap ad pricing, although we don’t expect management to provide the detail in the future. Still, we feel that this negative will reverse for several reasons. First, the likely influx of premium-priced video ads from the Amplify initiative will drive an immediate increase in CPMs. Second, 65% of TWTR’s ad sales are already on the mobile platform – the mix shift is likely to slow and any discount for mobile ads should vanish with time. Finally, advertisers are slow to appreciate the value of new online ad vehicles – TWTR has been selling ads for less than 3 years, and should see more robust demand as its track record of efficacy is better quantified and appreciated.
- TWTR is well positioned to add new apps. The successful launch of the Vine video application illustrates TWTR’s opportunity to leverage its scale, interest graph, advertising platform and user base to expand its base of revenue generating services. We believe this represents a significant, but difficult to quantify, additional value to the TWTR platform.
- Business model offers significant cost leverage to scale. As confirmed in the S1, COGS consists primarily of data center costs, which have significant economies of scale and which rise to support the growing volume of tweets and timeline visits. While this is obviously tied to growth in penetration and engagement, it is not driven by ad density, and as improvements on this aspect of the business drive revenues, gross margins should improve accordingly. Operating expenses should grow much more slowly than sales, particularly once stock related compensation costs are recognized through the IPO. EBITDA has been positive for 5 quarters, and while GAAP earnings have been consistently negative, we believe that they could turn profitable before the end of 2014.
- TWTR deserves a significant premium vs. the obvious compares. FB and LNKN each trade at slightly better than 12.5x forward sales, on respective 2014 top line growth expectations of 32% and 41% respectively. TWTR should be able to deliver better than 60% growth for the reasons addressed above, with considerable more runway to sustain the growth trajectory than either of the social networking stalwarts. Furthermore, it seems likely that management has taken a conservative approach to advertising density pre-IPO, giving it dry powder to spark added growth out of the gate.
TWTR began in 2006 as a simple and cheap way to share 140 character “text messages” with a group. While early observers struggled to understand its usefulness, Twitter caught on with tech savvy celebrities, who used it to interact with fans, and with reporters, who used it to “time stamp” developing stories. On that basis, the model of prolific “Tweeters” and dedicated “followers” emerged, abetted by Rich Media Cards, which allowed link previews, photos, videos, etc. to be embedded in a Tweet. With 218M global monthly active users (MAU), visiting their timelines roughly 7.5 times a day, to view more than 500M daily tweets, Twitter has long passed critical mass that is both driving growth – MAUs up 44% YoY, daily timeline views per MAU at nearly 8 worldwide – and frustrating would be competitors. Sales have more than doubled YoY, still on a steep upward trajectory from the introduction of advertising just three years ago.
We believe that Twitter can sustain extra-normal growth for a very long time. First, the user base will likely grow much larger. Twitter is used monthly on just 20% of U.S. and 12% of international smartphones. Given that Twitter was conceived as a mobile app, and that its new TV initiatives, cultural presence, and publicity around breaking news are raising public awareness, it seems likely that double digit growth in MAUs can be sustained for years. Second, the growing use of Rich Media Cards, including new deals to distribute NFL video highlights and other TV content, makes it likely that user engagement will also increase. We also note that complementary applications, such as Vine or a 2nd screen TV commentary app, could also juice engagement and create new venues to serve advertising. Third, and perhaps, most importantly, Twitter can increase its advertising density as it likes, a strategy that Facebook has used to obvious effect in driving its own revenues forward. We believe Twitter has deliberately limited its ads, and that the nature of its service will allow it to expand advertising without significantly reducing the satisfaction of its users. Moreover, we believe as the advertising community grows more comfortable with the efficacy of Twitter advertising, its ad pricing will rise accordingly.
We also expect Twitter to begin making serious money soon. The business model has inherently high gross margins, with the cost of goods primarily stemming from data center investment that will scale with the number of users and tweets rather than the amount of ads sold. Without the need to archive billions of old photographs, Twitter’s eventual gross margins will likely prove closer to Linked In’s 87% than to Facebook’s 73%. Expenses may grow quickly, but likely NOT as quickly as revenues – Twitter can only hire so fast. Finally, while messaging platforms like Kakao and LINK are stealing some of Twitter’s thunder in specific markets in Asia and the emerging world, the network and scale effects of linking tweeters with followers are making it extremely unlikely that real competition will emerge in western markets and with users that care about international content. While Facebook may seem an obvious rival, the polyglot nature of its service makes much, much less effective for the time sensitive nature of Twitter’s primary use cases.
Given the steep slope and sustainability of Twitter’s growth and its likely near term transition to profitability, we believe TWTR deserves a considerable premium over its more mature comparables, Facebook and Linked In, both of which trade at just over 12.5x consensus forward sales. Moreover, the dry powder on ad density and the potential near term boost from TV related initiatives position it for a strong start out of the gate.
What Hath God Wrought?
At precisely 9:50 PST, March 21, 2006, 20 year old Jack Dorsey pressed send on the message “Just setting up my twttr”, and sparked a global phenomenon. Originally, it was conceived as a way to send an instant message simultaneously to a group of people, thus the arbitrary 140 character limit, borrowed from SMS text messaging. As Dorsey later explained to the LA Times, “…we came across the word ‘twitter’, and it was just perfect. The definition was ‘a short burst of inconsequential information,’ and ‘chirps from birds’. And that’s exactly what the product was.” Twitter was launched to the public in June of that year, and by March 2007, when the company installed public Twitter monitors in the hallways of the “South by Southwest Interactive” festival, which brought the tech glitterati together with trend setting celebrities, a cultural meme was born (Exhibit 1).
Exh 1: Twitter Major Event Timeline, Founding to IPO
At first, it was a simple tool that let groups of friends organize themselves, but it quickly morphed into a soapbox from which the so inclined could speak their minds in 140 character snippets. This proved to be an irresistible forum for extroverted celebrities, the savviest of whom began to cultivate long lists of followers. The Twitter-verse began to divide into prolific Tweeters – e.g. celebrities, news reporters, sports figures, business executives, government types, etc. – and their followers, who would read the assembled tweets from their list in their chronologically displayed timelines. Tweets often contained links to web-based content, and media organizations joined their reporters and web celebrities in using the service to draw traffic to their sites. Followers could “retweet” posts, to expand the distribution of content deemed particularly interesting to their own followers, or tweet back to append their comments to an original tweet. These features allowed followers to interact with popular tweeters, and to join conversations around trending topics.
In 2009, Twitter introduced “Trending Topics”, a real time index of the word combinations that were appearing in tweets, giving the site a new role as a measuring stick for social interest and giving users an enhanced ability to find content on topics of interest offered by tweeters that were not on their follow list. In the same year, Twitter began to play a real role in news dissemination, with citizens on the ground scooping the traditional media in the crash of a US Airways jet in the Hudson River, and in the demonstrations after a disputed election in Iran. In 2010, Twitter began to sell advertising, delivered to users as “sponsored tweets” appearing amidst their normal timelines. In the same year, it also revamped the presentation format of tweets – a click would expand an original tweet to include a range of rich media, including link summaries, photographs, and more recently added, videos, embedded directly within the application. This has made the service more valuable to tweeters, advertisers and followers.
Today, Twitter is often miscast as a social network or micro-blogging service. These labels fail to appreciate the asymmetric nature of tweeting and following, and the key role of promoting and linking to other media. Twitter is the de facto place to find up –to-the-second information on breaking events on an extraordinary range of topics. It is the de facto place to keep track of celebrities and newsmakers, and for those celebrities and newsmakers to find their most avid and engaged audience. It trumps the variety of the Sunday newspaper and the immediacy of TV headline news in a personally customizable and conveniently accessed platform. We believe that it will be the dominant information distribution mechanism of the next age.
Exh 2: Twitter Monthly Active Users by Region, Q1 2010 – Q2 2013
How Many Tweets Can a Tweeter Tweet?
Twitter is large and still growing quickly – the S1 gives plenty of evidence of that. 500M tweets a day, up 25% in just 6 months since the 400M mark was hit around the company’s seventh birthday. 218M monthly average users, up 44% YoY in the most recent quarter (Exhibit 2). More than 150B timeline views per quarter, up 68% YoY, with the average MAU visiting Twitter more than 7.5 times a day (Exhibit 3). Advertising revenue per 1,000 timeline views was up nearly 25% YoY. International growth is particularly brisk, with MAU up 48% vs. 22.5% for the US, timeline views up 95% (7.4 per day per MAU), vs. 47% US growth (9.1/day/MAU), and ad revenues per 1,000 views up 76% to $0.30, vs. up 26% to $2.17 in the US (Exhibit 4). Add it all up and Twitter’s overall sales are growing at an eye-popping 104% pace.
Exh 3: Daily Timeline Visits per MAU, Q1 2012 – Q2 2013
Exh 4: Ad Revenue per 1,000 Timeline Views, Q1 2012 – Q2 2013
Twitter was designed from the start to be mobile – it was initially an app for sending text messages simultaneously to groups of people, hence the still enforced 140 character limit on text posts. In the S1 management revealed that 65% of the company’s advertising revenues came from mobile and that 75% of its MAUs accessed the service, at least in part, from a mobile device (Exhibit 5). Given that those smartphone owning Twitter users likely mix in some desktop access, it seems reasonable to assume that the company is monetizing its mobile timeline visits at rates not too dissimilar to its desktop visits. This bodes well for the future, when mobile access is likely to significantly outpace desktop.
Exh 5: US Mobile Twitter Users and Penetration, 2011-2017
More is More
Twitter can’t escape comparisons to Facebook. Twitter’s S1 reveals, in comparison to Facebook at the time of its IPO 18 months ago, a relatively immature company. Facebook had a billion MAUs, Twitter has 218M (Exhibit 6). Facebook had $3.7B in sales for the 12 months before its IPO, Twitter has had just $390M. Facebook was already profitable, Twitter is still making losses. However, Facebook was launched in 2004 and began selling advertising shortly thereafter, while Twitter began public service two and a half years later, and didn’t begin selling ads until early 2010. Moreover, Facebook was easily at home on a desktop browser with a use case – sharing personal thoughts and photos with friends – that was instantly valuable to the millennials that were its first adoptors, while mobile-focused Twitter only began to hit its groove as smartphones began to proliferate and as the critical mass for its asymmetrical distribution use case was breached.
Exh 6: Adult use of social networking sites and Twitter over time
When Facebook had roughly the same revenues as Twitter does today, in 2009, it would go on to grow its revenues at a nearly 90% CAGR over the next three years. At the end of 2009, Facebook had 350M users, a number it would nearly triple by the end of 2012 (Exhibit 7). For this, and for recent signs that revenue growth need not slow nearly as much as many analysts (this one included) had assumed as the company pushes its ad density to new heights, Facebook has been rewarded with a market capitalization north of $100B. While a similarly spectacular three-year run may seem implausible for Twitter, we believe that it is a realistic upper bound, given the current trajectory and the monetization potential of the platform.
Exh 7: Global Facebook User Growth, 2004-2013
The Long March
Twitter’s Monthly Average Users (MAU) have grown strongly, driven, in part, by the global shift to smartphones. Twitter is particularly suited to mobile devices, and comes pre-installed on iPhones and on many popular Android smartphone models. The global smartphone market grew 46% in 2012, with total smartphone users topping 1 billion at the end of the year, and continued growth suggesting a population of better than 1.4 billion at the end of this year. Industry research firm eMarketer projects that the smartphone base could swell to more than 2.5 billion by the end of 2017 (Exhibit 8). With 65% of Twitter revenue coming from mobile timeline views, this trend is undoubtedly bullish for future growth.
Exh 8: Smartphone Users Worldwide, by Region and Country 2011-2017
Exploiting the growing population of smartphone users also dovetails with Twitter’s strategy for international expansion. While the app is still blocked from competing in China, with its 350M smartphone subs, Twitter is enjoying a rush of adoption overseas – non-US MAUs are up more than 48% YoY. This trend can be expected to continue, and the company calls out Argentina, France, Japan, Russia, Saudi Arabia and South Africa as markets where it is experiencing particularly robust user growth. Not surprisingly, 4 of the 6 are amongst the world’s 15 largest smartphone user bases, with Argentina a few steps back at #21. Japan is of particular interest, as it is the one country where Twitter users outnumber Facebook users, with 39% penetration to 36%. Twitter does face competition in some of the bigger smartphone markets, in particular Southeast Asian countries – e.g. Indonesia, Thailand and Malaysia – where 3rd party messaging apps like Kakao, WeChat, and LINE, have taken up the original group messaging use case of Twitter and have begun to establish capabilities for media distribution (Exhibit 9). Still, the closed nature of the chat communities weakens their value, particularly for content with broad cross- border or cross-cultural appeal – you can’t follow Justin Bieber on Kakao or read about news half a world away. Given that just 12% of non-U.S. smartphone users are currently Twitter MAUs, this is a massive opportunity for future growth.
Exh 9: Major Messaging Applications by Users
While the lurid smartphone growth stats are in the emerging markets, there is also substantial opportunity at home, and in other developed markets. The US smartphone penetration has jumped over 1000bp from the start of the year to 64%, a high figure but on a trajectory that suggests that the peak will prove considerably higher (Exhibit 10). Moreover, just 20% of the US smartphone population uses Twitter on a monthly basis – we believe this percentage can increase considerably.
In the past year, Twitter has moved up in the public consciousness, with now ubiquitous Twitter handles, follow requests and hashtags appended to advertising, TV broadcasts, print publications and even Wall Street research. News reporting routinely sources Tweets from the scene, which may now include photos and video. Artists of all ilk openly encourage their fans to follow them. Twitter’s 6 month old Amplify program has signed deals with CBS, the NFL, Fox, ESPN, Major League Baseball, the NBA and a dozen other media partners to put exclusive video clips on Twitter, to promote the Twitter connection on broadcast programming and co-ordinate advertising campaigns across platforms (Exhibit 11). As media partners make more and better use of the rich media capabilities of the platform, its utility to a broader audience of potential MAUs will only increase and become more apparent. We believe that a concerted effort to explain the usefulness of the service and to stimulate adoption and use could dramatically increase the number of MAUs and the frequency of their visits.
Exh 10: US Smartphone Penetration, Q1 2009 – Q3 2013
Exh 11: Twitter Amplify Partners
What is the Frequency Kenneth?
The average Twitter user visited the site 7.6 times a day in the June quarter, up from 7 daily views in the previous spring. Breaking this measure of engagement out between US and international MAUs, the figures are even more impressive, with American daily views up almost 11% to over 9 and international views up 22% to 7.2 per day (Exhibit 12). We believe this impressive growth in engagement can continue, and although disproportionate growth in the less engaged global market may mute aggregated engagement figures, the overall trend is unambiguously positive.
Exh 12: Daily Timeline Visits per MAU, Q1 2012 – Q2 2013
In the U.S., views per user per day have plateaued over the past two quarters, but there is reason to believe that the Amplify co-marketing deals with CBS, the NFL, and other media partners signed in the last few months will not only attract new users to join, but will also induce existing users to visit more often. Furthermore, the growing penetration of smartphones, the relative bandwidth efficiency of Twitter, and the increasing role of Twitter in the culture of the day should also spur rising engagement going forward. The International story is also straightforward. As local newsmakers and celebrities in each market establish ongoing presence on Twitter and as smartphone penetration increases, a self-reinforcing cycle of engagement will drive daily visits ever higher. We expect international daily views per monthly average user to slowly close the gap vs. the US, even as American engagement edges higher. We expect total daily views per MAU to top 8 before year end, and to approach 9 by the end of 2014, continuing the strong trajectory of increasing engagement that has already been well established.
Show Me the Money
Continued growth in MAUs and engagement per user have increased the opportunity for Twitter to place advertising. Total timeline views are up 68% YoY, with US visits up more than 45% and international visits up nearly 80%, with each timeline view a chance to place advertising. In the S1, Twitter provides a simple metric for its advertising density – ad revenues per 1,000 timeline views. This metric, which combines the effects of changing ad placement and ad pricing, has been steadily growing, up 25% YoY to $0.80 in the June quarter. The split between US and non-US ad markets reveals a severe dichotomy, with the U.S. generating $2.17 per 1,000 views, up 26% YoY, and international ad sales at just $0.30 per thousand, but up more than 114% vs. the same quarter in 2012 (Exhibit 13). Twitter cautions that the disparity between the advertising monetization in the US and abroad, combined with the higher growth of timeline views outside of the US will likely distort this measure of advertising density, and arguably, it already has.
Exh 13: Ad Revenue per 1,000 Timeline Views, Q1 2012 – Q2 2013
We believe that Twitter can and will serve more ads to its users. Twitter only began selling advertising in 2010, a full six years after Facebook, and on an unscientific basis, counting ads on both services on mobile and desktop platforms from a handful of users, appears to be offering half as many ads to its users. We believe that this disparity is largely artificial, and that just as Facebook has been able to accelerate its top-line growth in recent months by push ad density, so shall Twitter.
In some ways, Twitter may be better able to increase its advertising load than it rival. The nature of its timeline, a stream of 140 character posts with appended rich media files that appear as a user clicks on an item and expands it, is well suited to advertising. Ads are easily identifiable, quickly addressed, and do not disrupt the user flow of sorting through the stream of posts, fitting well into the context of gathering information from many disparate sources. In contrast, Facebook’s less focused use cases may make advertising a greater burden on users as they navigate the site to meet their immediate needs, be they photo sharing, messaging, social gaming, searching for friends, or simply killing time. Similarly, Facebook’s extraordinary and broad “social graph” may be less effective in targeting advertising to interested parties than Twitter’s “interest graph”. All of these advantages are relevant factors in the success of Amplify, which positions Twitter as on-line marketing partner of choice for traditional media outlets, a status obviously coveted by Facebook.
From the demand side, the Amplify program also portends significant additional advertising opportunities, much of it high priced video pre-roll attached to video clips provided by Twitter’s media partners. While revenue from this program will be shared, the typical price is likely to be a multiple of the text-photo-link ads that currently make up the bulk of Twitter advertising. Moreover, as the efficacy of hybrid TV/Twitter campaigns becomes better appreciated, demand for rich media ads from outside the Amplify partnerships is likely to accelerate. Twitter’s sales pitch asserts that its own data shows that consumers that watch a TV spot and then interact with the advertiser via Twitter report a 58% higher likelihood of buying than viewers that just watched the TV ad alone. If this effect can be verified in practice, advertisers will beat a path to Twitter’s door. Ultimately, this co-operative approach will also be available as the company continues to expand into international markets as well.
Exh 14: Internet ad CPM Ranges by Media, 2013
The Price is Right
Twitter’s ad pricing is set by auction, and as such, is dependent on market demand and on the inventory that the company makes available for sale. A study from 1Q2012 of 45 billion ad impressions on Twitter and Facebook by online ad placement agency TPG Digital suggests that demand has been far outstripping ad inventory, with Twitter ad CPMs for that quarter running at more than $3.50, while Facebook ads, which at the time were disproportionately skewed to low priced banner ads, were priced almost 90% lower. Facebook CPMs are believed to be considerably higher now on the success of ~$4.00 CPM “page post” ads interpolated into a user’s news feed and “sponsored stories” which often tie to product “likes” and location “check-ins” to generate even higher pricing (Exhibit 14). Given that Twitter’s advertising is exclusively placed directly into its users’ timelines and that it appears to be tightly controlling inventory, it is likely that Twitter still enjoys a modest pricing advantage over Facebook.
That pricing advantage could grow considerably, if Twitter gets a better jump on video advertising. CPMs for pre-roll video ads can run as high as $30-40 CPM for high quality placements, and, obviously, a mix shift toward rich media would have a corresponding effect on Twitter’s ad pricing. Even if Twitter is sharing revenue with its Amplify partners, Twitter’s share would be a hefty boost to average CPM. Moreover, as Twitter only charges for ads that evoke a user response – e.g. expand the tweet, click through a link, retweet, etc. – its realized pricing should also improve as more engaging advertising formats induce increased responses.
Of course there is a downside to ad pricing. Ad rates outside of the US are considerably lower than they are in country, and as international growth continues to outstrip American expansion, average rates across all markets are likely to slide. Twitter is only issuing a single metric for its advertising, revenues per 1,000 timeline views, and cautiously notes in its S1 that “to the extent the number of international users and engagement by international users grow faster than U.S. users and engagement by U.S. users, total advertising revenue per timeline view may be adversely impacted even if total advertising revenue continues to increase.” This is true, and thus, overall deterioration in this metric should not be viewed as deterioration in the demand for Twitter advertising.
Twitter describes its COGS as follows:
“Cost of revenue consists primarily of data center costs related to our co-located facilities, which include lease and hosting costs, related support and maintenance costs and energy and bandwidth costs, as well as depreciation of our servers and networking equipment, and personnel-related costs, including salaries, benefits and stock-based compensation, for our operations teams. Cost of revenue also includes allocated facilities and other supporting overhead costs, amortization of acquired intangible assets and capitalized labor costs. 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.”
Given the relatively immature status of the company’s revenue stream, it is not surprising that gross margins, at 64%, are running considerably below its more mature comps, Facebook’s 75% and LinkedIn’s 87%, and management points out later in the S1 that unrealized stock compensation will weigh on COGS through the next few quarters (Exhibit 15). However, as Twitter grows and gains economies of scale, its gross margins will undoubtedly expand, but by how much?
Exh 15: Gross Margins – Facebook, LinkedIn, and Twitter
We believe Twitter’s mature margins will be closer to Linked In than Facebook for several reasons. First, like both compares, Twitter’s COGS are tied to the volume of data generated by users and by the number of times that users visit the site to view that data, NOT to the number and price of ads that are included within that data. Because of this, rising ad density and prices will typically cause sales to grow faster than infrastructure requirements.
Second, the nature of the data generated by Twitter users is compact – largely text and URL links, with a growing but small flow of more storage and processing intensive embedded rich media. In contrast, Facebook has become the world’s largest repository of massively storage intensive digital photographs. In this, Twitter is far more like Linked In than Facebook.
Finally, the “up to the second” nature of the Twitter use case means that Twitter users do not typically perform long deep archival searches of historical material and many embedded rich media files can be discarded after a modest holding period, and replaced with a link to an original file stored by a media partner. Again, Facebook bears a considerable relative burden in maintaining a deep and searchable archive of image and video filed. Linked In is better off than Facebook in this regard, but still has a greater responsibility for archiving and searching historical material then Twitter. All things being equal, this is a cost advantage for Twitter, and adding this to the two earlier points, we believe mature gross margins for the company should easily top 80% and could approach 90%.
Twitter spends a LOT of money on R&D, currently at more than 44% of total revenues. In comparison, Facebook spends 27.5% of sales and Linked In spends 26.5%. In the near term, Twitter’s R&D spend will be bloated by stock compensation expenses, but long term, it wouldn’t seem likely that the company’s R&D needs will be any more pressing than Facebook’s. The same is true of SG&A, which is running at 44.5% of sales for Twitter, vs. 25.3% for Facebook, and 48.0% for Linked In, which carries the added burden of an enterprise sales force needed to market its Human Resources services (Exhibit 16). Ultimately, we believe the mature expense structure of Twitter should be no more than 62.6% of sales spent by Facebook in 2012, and likely (and also likely true for Facebook itself) lower than that.
On this basis, the long term operating margins for Twitter ought to be higher than those generated by either Facebook or LinkedIn, should top 20% and could reach 30%.
Exh 16: SG&A- Facebook, LinkedIn, and Twitter
The acquisition and successful launch of Vine, a separate app that allows users to create short 6 second videos that can be subsequently posted to Twitter, or to other sites, is an early indication of Twitter’s ability to leverage its position with Tweeters, followers, media partners, and advertisers into related synergistic businesses. Facebook has proven the same with its purchase of Instagram, as did Google with its 2006 acquisition of YouTube. For all of these companies, scale gives them the ability to add value to acquisitions that if executed properly can become significant additional conduits for monetization (Exhibit 17). While it is difficult to place a specific value on this optionality, it is an obvious asset to Twitter as it moves forward to compete.
Exh 17: Acquisitions by Twitter
We believe Twitter is a well-managed company with a well-conceived strategy and a strongly advantaged position to be the dominant distributor of time valuable information and content. We believe that the business is attractive enough to drive better than 35% annual growth in users, 10% growth in timeline views per user, and 10% growth in ad sales per timeline view, over the next three years, which would push 2016 revenues to nearly $2.6B. With costs approaching mature levels, it is possible that operating profit could top $650M in that year, with the company remaining on a strong growth trajectory. Applying a Facebook or Linked In trailing sales multiple of 21x suggests a year end 2016 valuation of well over $50 billion. Discounting that back with any reasonable rate suggests a current value of $25B or more.