Twitter: Out of the Nest, Into the Sky
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Artur Pylak
FOR IMPORTANT DISCLOSURES 203.901.1633 /.1634
sagawa@ / email@example.com
November 3, 2013
Twitter: Out of the Nest, Into the Sky
- TWTR sales growth will be driven by 5 factors: number of users, engagement by users, ad density, ad pricing and additional monetization vehicles. Monthly active users were up 39% YoY in 3Q13, with considerable further room for growth, given TWTR’s affinity for mobile. We expect a 25-30% CAGR through 2017. Timeline views per MAU were up 8% in 3Q, but had been growing at better than 16% previously. As TWTR’s media partnerships bear fruit, we expect engagement to reaccelerate somewhat in the near term and to average 8-10% growth over the next 4 years. Ad density is by far the biggest driver of TWTR’s 100%+ sales growth, with ad engagements per timeline view 2.5x vs. last year. FB’s results are clear evidence of users willingness to absorb more ads, and given our belief that TWTR is far more conducive for advertising, we project 60-70% annual growth in ad density going forward. In contrast, pricing has been declining 62% YoY as TWTR has increased the inventory offered in its ad auctions. We expect this to temper somewhat, with a mix shift to higher CPM video and increasing advertiser enthusiasm yielding a more comfortable 20-30% yearly deterioration. Finally, we believe that TWTR has substantial potential to offer new money making services on its platform, but have included no value for this optionality in our model. In aggregate, we are projecting ad sales growth of 60-70% and modeling $4.5B in 2017 revenues. Applying a FB multiple to that and discounting back at 30% yields a present value of $30B+.
- TWTR is far from saturation. TWTR has 232M MAUs, 75% with mobile access. That is over 13% penetration of smartphones, 9.5% of internet users, and 20% of FB’s user base. In developed markets, TWTR enjoys 24% penetration of internet users, vs. 22% in developing markets. 80% of potential developing market users are in countries where TWTR is well positioned, while 20% are in countries where TWTR faces competitive challenges, including 5% in countries, like China, where TWTR faces outright bans. Given its natural affinity for mobile service and the fact that many smartphone brands include it as a default app, we believe that TWTR can grow to at least 50% smartphone penetration in markets where it is well positioned, and 10% in markets where it is available but challenged. With more than 3.5B smartphone users expected by 2017, this suggests a target of 650-800M MAUs, and a 25-30% annual growth rate.
- Engagement can improve. TWTR users visit their timelines an average of 7.5x per day. FB users view an average of 44 pages on the service each day. While page views are not the same as site visits – TWTR users may view multiple “pages” on each visit – FB engagement is likely a 3-4x multiple of TWTR on an equivalent basis. TWTR visits/MAU slipped sequentially in 3Q, but have been on a rising trajectory over the past 7 quarters. With new media partnerships kicking in, we expect user engagement to return to growth in 4Q. Assuming an 8-10% CAGR, timeline visits per MAU per day ought to exceed 10 by 2017, and could reach 11.
- Ad density is increasing. The number of TWTR ad engagements, per timeline view, is 2.5 times higher today than it was a year ago. Still, sponsored tweets appear to be less than 1 in 50 posts for U.S. users. We believe that TWTR ads are relatively benign to the user experience, and thus, project that US ad density could triple again before becoming an annoyance to users. Outside of the US, where we believe ad densities are considerably lower, the increase could be considerably greater. With international timeline views outnumbering US by nearly 3 to 1, we believe ad density could increase at more than 60% annually through 2017.
- Declining ad pricing could reverse. TWTR’s average revenue per ad engagement declined 62% YoY in 3Q13, a sharp deterioration from the 40% decline noted vs. the previous year. Given TWTR’s auction process for selling its ad inventory, the change in trajectory is likely tied to the simultaneous sharp increase in ad inventory made available. TWTR has only been selling ads for 3 years, and we expect prices to stabilize as ad buyers grow more comfortable with the efficacy of the platform, a phenomenon that has been apparent in ad pricing for Facebook and Google before it. Furthermore, we expect a considerable mix shift toward video advertising, carrying much higher CPMs, as TWTR’s broad roster of new media partnerships begins to bear fruit. These positive effects will be somewhat offset by a continued volume shift toward international markets, which carry lower advertising pricing. Net net, we expect overall ad pricing declines to decelerate going forward, averaging 20-30% annually through 2017.
- TWTR could pull other monetization levers. Today, TWTR generates nearly 10% of its revenues from sales of its data to licensees. This stream has and will grow more slowly than advertising revenues, but should still account for 2-3% of sales in 2017. The MoPub acquisition could make TWTR a leading network for mobile ads on other sites, a potential source of significant new revenue. TWTR could also pursue other opportunities to monetize its primary service – e.g. premium content subscriptions, click-to-buy commerce, etc. – or to launch additional revenue generating services that leverage its user base and interest graph – e.g. a unified messaging platform, an Instagram-like photo sharing service, etc.. We have not included revenue from new products in our model, but the opportunity is there.
- Revenue projected to grow at 60-70% CAGR through 2017. Assuming a 25% CAGR in MAU, 8% in daily user timeline views, 60% in ad density, and 20% decline in ad rates, TWTR’s advertising revenue growth rate would be just over 60%/yr and overall 2017 sales would be just under $4.3B. Taking the upper end of our user growth, engagement and ad density estimates, but applying a 30% annual decline in ad pricing, suggests 70% advertising sales growth and over $5.2B in 2017 sales. Our model, toward the lower end of this range, assumes 65% annual ad sales growth and projects 2017 sales of $4.5B.
- TWTR is likely worth more than $30B. FB and LNKD both trade at trailing sales multiples of more than 20 times. If TWTR is able to grow at the rates that we believe that it can, applying a similar multiple to our projected $4.5B in 2017 sales would yield a future valuation of $90B or more. Of course, relying on future projections is extremely risky, but applying a 30% discount rate to account for that risk suggests a present value of more than $30B for the company. Even a 40% discount rate would value TWTR at $23B, about double the low end of the preliminary pricing range.
Twitter – How Much Bigger Does That Thing Get?
TWTR is growing VERY quickly. 232 million MAUs worldwide, up 39% YoY. 7.5 timeline views/ user/day, up 7.9 % YoY. $0.97 in ad revenues per 1,000 timeline views, up 49% YoY, with 153% growth in ad engagements per timeline view partly offset by a 62% drop in price. Sales have more than doubled YoY in every quarter of the past 3 years, with ad sales accelerating from 116% growth in 2Q13 to 130% growth in 3Q13. We think that the trajectory going forward will be more of the same.
Management points out that TWTR’s 232 million MAUs are just 9.5% of the world’s internet users and just 13% of the smartphone users. Facebook has 5 times as many. Arguably, this is an indication that Twitter has considerable room to grow, particularly if it gets serious about communicating its compelling use case to consumers, many of whom seem confused about how to get started. Of course, Twitter is challenged in certain countries – for example, it is largely prohibited in China. Accounting for that, we see 20-25% penetration of smartphone users as a reasonable target, or 650-800 million MAUs by 2017. Engagement reversed course in 3Q13, down <1% QoQ, possibly an indication that strong user growth was sharply backend loaded. Timeline views per user had been growing at a double digit YoY pace previously. With new co-promotional partnerships with major media players kicking in this fall, we believe that engagement will reaccelerate in coming quarters. We have modeled 8-10% annual improvement in timeline visits per MAU, and will monitor the metric closely for confirmation
With users and engagement driving ~50% growth in total timeline views, TWTR is successfully monetizing those visits. TWTR only charges its advertisers if users engage with their ads, and these engagements are growing at a stunning 280% rate, the product of increasingly dense placement into timelines and user interest in those ads. The increase in density was partly offset by a 62% YoY drop in average ad pricing, a consequence of putting so much more inventory into an auction process. Still, as advertisers get more comfortable with TWTR and as the media partnerships drive a mix shift toward much higher CPM video ads, pricing should stabilize and could even begin to grow a la FB. We believe ad density can continue to grow, albeit at more controlled 60-70% rates, with price declines moderating to 25-30%, yielding annual “ad sales per 1,000 timeline view” growth of 30-45%. In this, we see FB’s clear success in driving advertising growth as a model, with TWTR’s use case, perhaps, even better suited to higher degrees of ad density.
Our valuation model does not give TWTR any credit for opportunities beyond timeline advertising, although these obviously exist. 10% of sales already come from data licensing, although we expect this business to grow much more slowly than advertising. The MoPub acquisition also has promise, setting up TWTR to be a leading network for placing ads on mobile sites. Just as FB is launching advertising on the recently acquired Instagram, TWTR could pursue a similar strategy to lever its user base and interest graph for affiliated ad supported services. However, even without valuing MoPub or this optionality, we believe TWTR, with ad revenues projected to grow at 60-70% through 2017, ought to be worth more than double the high-end valuation suggested by the preliminary IPO price range. Taking estimated 2017 revenues of about $4.5B and applying a FB or LNKD like 20x multiple, suggests a future valuation of ~$90B, and discounting that back at 30% discount rate that acknowledges the risk inherent in our assumptions yields a better than $30B value today.
What’s Black and White and Read All Over?
For a company that only started selling advertising three years ago, Twitter is making up for lost time. Overall revenues are growing at a 105% pace, with ad sales – now 90% of total revenues – expanding at a 120% annual clip (Exhibit 1). While it is typical for fledgling companies to experience hyper-growth before an inevitable deceleration, Twitter has pulled out of the stall to maintain a steady doubling over the past three quarters, with the prospect that new revenues tied to its Amplify media partnerships can keep the party going.
Exh 1: Twitter Revenue by Segment, Q1 2011 – Q3 2013
Looking at advertising revenues more granularly, there are four inter-related elements to the equation: 1. How many people are using the service? 2. How often do those users view their timelines? 3. With how many ads do they engage, each time they visit? 4. What price does Twitter earn for these engagements? Finally, there is a fifth element driving revenues outside of timeline advertising -5. What other monetization vehicles can Twitter introduce to lever its user base and the detailed data graph of their interests? We believe breaking down each of these factors gives insight into the sustainability of Twitter’s sales growth and the plausibility of future revenue targets (Exhibit 2).
Exh 2: Twitter Revenue Drivers and Model Assumptions
Take Me to Your Tweeter
In the beginning, Twitter was intended for an entirely different purpose than its use case today. Jack Dorsey, Evan Williams, Biz Stone, and Noah Glass were casting about for a new direction for their start up Odeo, since that company’s primary podcasting business was looking increasingly like future road kill in front of the iTunes juggernaut. What they came up with was a service intended to make it easy to manage SMS text communications amongst a group of people – essentially a precursor to the instant messaging apps that are now slugging it out in so many emerging markets. Early on, Twitter was fortunate to gain the notice of the glitterati of the tech and new media world, and curious newbies flocked to the public service to see what these notables were saying. A breakout experience at the 2007 South-by-Southwest festival brought tech savvy Hollywood types to the suddenly hip service, and the snowball started rolling (Exhibit 3).
Users themselves established the community norms of referring to themselves by their @twitterhandles and of tracking topics by appending a descriptive #hashtag to their posts. To their credit, Twitter management recognized the shifting zeitgeist around their service and rushed to incorporate these and other conventions into the basic service structure, allowing users to easily find Tweeters by their name or handle and to follow “trending topics” by their hashtags. Twitter was morphing from a peer to peer network of acquaintances communicating with one another to a distribution service for “Tweeters” to distribute content in the form of 140 character posts with attached web links, out to multitudes of “Followers” who used the platform as a news service of sorts with a self-curated list of sources.
As this became the primary use case, the platform attracted anyone and everyone who had valuable content to share (or at least, tease). Actor Ashton Kutcher became the first to hit a million followers, only to be erased by the wave of pop stars, sports figures, political figures, journalists, business executives, and entertainers that flooded the medium. Breaking news, like the Arab Spring and Sully Sullenberger’s emergency landing in the Hudson, started showing up first on Twitter, which has now become THE go to source for up to the minute information, even amongst the reporters who are on the scene. In this way, Twitter has become the logical successor to the morning newspaper and CNN for the next generation of information consumers.
Exh 3: Twitter Major Event Timeline, Founding to IPO
As the use case developed organically, Twitter let word of mouth be its marketing. Newscasters were quoting Tweets on the air, popular TV shows were riffing on trending Twitter hashtags, and advertisers were tagging their spots with corporate Twitter handles. Meanwhile, Apple and Microsoft negotiated deals to integrate Twitter as a default app on their mobile platforms, while most of the major Android OEMs – e.g. Samsung, HTC, LG, Motorola, et al. – were shipping it pre-loaded on their popular smartphone models. It was built and a lot of people came.
However, by letting the users define the use case and spread the word, Twitter has been slower to spread into demographics that are less plugged in, with smartphone users who don’t entirely understand why they would use Twitter and are not sure how to get started. An education campaign explaining the panoply of information available on the service – NFL and MLB replays, up to the second news stories from trusted sources, exclusive previews of popular TV shows, juicy celebrity gossip and opportunities to interact directly with the stars, ongoing debates amongst experts on almost any topic, blah, blah, blah – would be a good start, along with reassurances that no one expects users to Tweet if they don’t want to. Rethinking the “getting started” process for new users would also be a good idea, including a more aggressive marketing program to move past word of mouth.
More MAU, Please
232 million users is a good start, up 65M from a year ago. The initial prospectus spooked investors a little bit, as the June quarter numbers showed US monthly active users up just 2% sequentially, but the 8% QoQ jump in the September quarter that was revealed in the subsequent amended filing appears to have put those fears to bed. As of 3Q13, both US and International MAUs were growing at better than 32% YoY, for an overall 39% rate. That’s good, but as suggested in the previous section, we think they could do better, or at least, keep it going for a long time (Exhibit 4).
Exh 4: Twitter MAUs by region, Q1 2010 – Q3 2013
Exh 5: Potential Twitter User Growth
Twitter’s target market is internet users, of which there are 2.4B in the world, or more specifically, smartphone users, of which there are 1.7B, according to Twitter’s roadshow presentation (Exhibit 5). As a point of further comparison, Facebook, which has a roughly 4 year head start on Twitter and a somewhat less confusing use case, currently claims 1.19B MAUs (Exhibit 6). The smartphone population is growing at a better than 45% annual clip and is expected to number more than 3.5B by 2017, more good news for Twitter, which was designed from the start to be a mobile application. Many smartphones, at least those at the higher end of the spectrum, come pre-loaded with the Twitter App, making it all the easier to convert that smartphone buyer into a Twitter user. So, how many of those smartphone users will turn out to be Twitter users?
Exh 6: Global Facebook MAU User Growth, 2004-2013
Right off the top, China (along with Iran and North Korea) has banned Twitter. Although it has been reported that many Chinese have found ways around the ban, it is probably prudent to assume that the censors will win, and that these markets, and the 1 billion 2017 smartphone users that they represent, will be out of reach for Twitter (Exhibit 7). In parts of Southeast Asia, Twitter faces a competitive challenge from wildly popular instant messaging platforms, like Kakao, LINE and What’sApp (Exhibit 8). While the service that these applications provide is now quite different from Twitter, in countries where Twitter is not well established, such as Thailand, Malaysia or Vietnam, the messaging providers could subvert its plans to penetrate the market by offering localized versions of the content distribution functionality. We have assumed that the potential penetration of these markets is considerably lower than in the rest of the world (Exhibit 9).
Exh 7: Twitter Penetration as % of Internet Users, Q2 2013
Exh 8: Major Messaging Applications by Users
Exh 9: Countries making up the 80th percentile of Twitter users ranked by penetration
Breaking the smartphone population into three buckets – banned, challenged and well positioned – and assuming a 0%, 10% and 50% respective penetration ceiling for each of the buckets, we believe that the upside for the 2017 user base is about 800 MAUs, which would suggest a 30% annual growth rate over the next 4 years. Tempering that growth to 25% would pull back the 2017 user base to 650 MAUs. Given the current 39% growth rate in MAUs, the prospect that the higher profile created by Twitter’s media partnerships can generate even further interest, and the wild success of Facebook before it, we are comfortable that this range is conservative.
Making a list, Checking it Twice
The average user checks his/her timeline 7 and a half times each day, with Americans checking in more than 9 times a day (Exhibit 10). Twitter, which is very efficient and does not tax smartphone users data plan allowances, has a lot of room for improvement on this statistic. In contrast, Facebook users view approximately 44 “pages” on the site every day, and while a single site “visit” may entail multiple page views by Facebook’s metrics, it is likely that the social network leader enjoys 3-4 times the engagement of Twitter (Exhibit 11).
Before 3Q13, Twitter engagement had been cruising along at a double digit pace of growth, with 2Q timeline visits per MAU up 17% YoY. September saw the metric slam on the breaks, up only 8% YoY and, shockingly, down 1% QoQ. Thus far, management has vaguely suggested that the sequential decline in engagement was related to changes that the company had made in the product in order to improve the user experience. Twitter DID discontinue support for an older version of its APIs in mid-June, a move that invalidated many 3rd party apps that allowed users to access the service – if this is the culprit behind the sequential engagement drop, it is likely a temporary effect.
Exh 10: Daily Timeline Visits per MAU, Q1 2012 – Q2 2013
Exh 11: Engagement rate among social networks
With Twitter’s Amplify program and its media partnerships – partners include NBCUniversal/Comcast, CBS, ESPN/Disney, Fox, The NFL, Major League Baseball, the NBA, Conde Nast, and others – gaining traction in 4Q13, we are confident that engagement will be up again in the quarter (Exhibit 12). Longer term, we are modeling 8-10% annual increase in Timeline Views per MAU.
Exh 12: Twitter Amplify Partners
You Are Who You Follow
In just three years of ad sales, Twitter has gone from zero to a projected $600M for 2013. Ad revenues per user are running at about $0.66 per quarter – that’s $2.11 in the US and just $0.23 overseas (Exhibit 13). While Facebook is generating $1.72 of revenues per user, that number does include non-advertising revenue and is spread over a much larger number of daily visits (Exhibit 14). Adjusting for the 3-4x difference in engagement, Facebook’s revenue per user visit is a bit to a lot lower than Twitter’s, a fact that is all the more surprising in light of Facebook’s comments that it may have reached the point of saturation in interpolating advertising into its users’ newsfeeds. Part of this is Facebook’s more international user base – US only ARPUs are not available, but are likely a magnitude higher than the International numbers. Still, the comparison is a testament to the success that Twitter is having in selling its platform to advertisers.
In its amended S-1, Twitter reveals torrid QoQ growth rates in the number of ad engagements that it has tallied over the past 7 quarters. Compounded over a full year, engagements were up an eye-popping 280% YoY in 3Q13, including a 124% sequential jump between 1Q and 2Q. Even before that big jump, ad engagements had been up 170% YoY in 1Q. Adjusting for the ongoing growth in users and visits, the growth in engagements per timeline view is still up 150% YoY. Note that Twitter only charges its advertisers once users engage with the ad by expanding it, following its link, or retweeting it, so the growth in engagements is the product of an unknown mix of more ad placements and the changing propensity of users to engage with the ads.
Exh 13: Ad Revenue per MAU, Q1 2012 – Q2 2013
Exh 14: Worldwide Facebook ARPU, Q3 2012 – Q3 2013
We think that Twitter is still far away from saturating its timelines with advertising. An unscientific tally of sponsored tweets in 1,000 tweets viewed at different times of the day puts the ad density at about 1 in 65, a level that seems far from obtrusive. Moreover, the nature of the Twitter timeline – a long list of posts with varying levels of interest – and the purpose for which users view their timelines – to absorb information – are conducive to conveying advertising. Add in that Twitter’s knowledge of its users REAL interests – there is no posturing or subterfuge by users that are simply consuming tweets – appears to allow very accurate targeting of ads.
Casually comparing the service to Facebook, it seems likely that Twitter could increase the density of ads on US timelines by a factor of 3 or 4 before any serious challenge to the user experience (Exhibit 15). Based on the very low levels of monetization for International markets, it appears very likely that the ad density of timelines is considerably below US levels, and thus even further from ad saturation. Recognizing that the 150% growth in ad density for the most recent quarter is unsustainably high, we are modeling a 60-70% annual increase in ad engagements per timeline view.
Exh 15: Average advertising density of major social websites and apps
Come On Down! It’s Your Turn to Play the Price is Right!
Twitter’s ad pricing has been dropping like a stone, down 62% in 3Q – a likely side effect of the massive increase in inventory implied by the torrid growth in total engagements (Exhibit 16). Twitter sells its ads in a real time auction process that is obviously subject to the laws of supply and demand, and the biggest quarterly price decline coincided with 124% increase in total engagements notched in 2Q13. As management tempers the growth in available placements, pricing should stabilize as a matter of course.
Beyond simple supply and demand, there are other reasons to be optimistic about Twitter ad pricing. First, the advertising community has only had three years to assess the value of Twitter as a platform – most Internet services have found ad pricing rising over time as ad buyers grow more comfortable with the efficacy of the service as an advertising vehicle. This was certainly the experience with Facebook. It will also help that Nielsen has launched a metric for evaluating the attention paid to media on Twitter in conjunction with its industry standard television rating service. Second, Twitter has enhanced its platform to support rich media embedded into its expanded Tweet cards and to show images directly within the timeline itself. This allows advertisers to present users with compelling image and video content, and to encourage more users to engage with sponsored tweets. Third, Twitter’s media partnerships bring with them exclusive video content – sports replays, show previews, etc. – that can carry high priced video advertising. CPMs for pre-roll video ads run 10-15 times higher than ordinary display ads, and while the media deals may entail sharing some of this revenue, a mix shift in this direction will be hugely beneficial to Twitter’s overall ad pricing (Exhibit 17).
At the same time the trajectory has been sharply downward and with the ongoing shift toward International, pricing growth is likely too much to ask for. We have modeled 25-30% annual price drops through 2017, with the expectation that this will turn out to be very conservative.
Exh 16: Twitter Cost per Ad Engagement, Q1 2012 – Q3 2013
Exh 17: Digital Video Versus Facebook CPMs, 2013
It’s a Floor Wax AND a Dessert Topping!
Timeline advertising is not Twitter’s only source of revenue – nearly 10% of sales still come from licensing contracts for the company’s user data base. While data licensing is still growing at an attractive 36%, it is not a priority business for management, which has many ideas on how it might use the data itself. As such, we expect this business to fall off of the pace fairly quickly, and have forecast just 13% annual growth over the next four years (Exhibit 18).
Exh 18: Twitter Revenue, Q1 2011 – Q3 2013
While licensing is almost certain to decelerate, Twitter has many other, intriguing possibilities to leverage its data, infrastructure and user base. The recently acquired MoPub is one such possibility, offering a platform to expend Twitter’s already valuable interest graph to include broader activity on Twitter affiliated sites and use it to target ads outside of Twitter. As Antonia Garcia, the architect of Facebook Exchange, wrote on Medium.com,
“The MoPub acquisition allows Twitter to fundamentally change how mobile ads are purchased and places them at the forefront of how mobile, Web, and social ads interact. This makes Twitter the most interesting company in advertising right now.”
He goes on to make the point that the combination of Twitter’s interest graph data and identity tracking across platforms with MoPub’s real time bidding exchange for mobile advertising inventory, is much more than the sum of the parts in providing value to advertisers. Garcia believes that this gives Twitter a major leg up on the obvious exchange competitors, Facebook and Google, neither of which, he believes, offers as complete a solution for mobile advertisers. If he is correct, this could be a VERY valuable business for Twitter, and one for which we have allocated zero revenue in our valuation model.
Facebook, now on the brink of monetizing its $1 billion acquisition of Instagram with the services first ads, provides another monetization blueprint. Vine, acquired by Twitter, which presided over the launch of the popular video clip app, is probably not yet substantial enough to support an advertising component, but could become the basis for a more complete social network platform. A future deal for a social network such as Snapchat, a messaging platform such as What’sApp, an on-line content play like Buzzfeed, or an e-commerce platform such as Etsy or Pinterest, could have interesting synergies with Twitter’s existing service and infrastructure. Again, we have recognized no value at all for this optionality in our model.
Exh 19: The SSR Twitter Model
What’s it Worth to Ya?
Combining 25-30% growth in monthly active user with 8-10% growth in average timeline views per user, yields an expected annual growth in total timeline views of 39%. Combining 60-70% growth in ad engagements per timeline view, and 25-30% declines in the price of ad engagements, yields an expected growth in advertising sales per timeline view of 19%. Considering both usage growth and increasing advertising yields points to a projected growth rate of about 65% give or take several percentage points in either direction. We believe that the assumptions that go into this calculation are conservative, in particular, the continuing sharp decline in ad pricing.
Using 2012 seasonality as a guide, we project 4Q13 revenues of roughly $243 million, bringing our full year 2013 sales estimate to $665 million, $601 million of which is advertising. Applying the 65% sales estimate to the ad sales, and applying a normal deceleration curve over the next four years, we project 2017 advertising revenues at $4.436 billion, with 13% average growth in data licensing sales yielding an additional $106 million, for a total of $4.54 billion. Note that this estimate does NOT include incremental revenues for MoPub, Vine or any other subsequent acquisition by Twitter.
Both Linked In and Facebook trade at roughly 20 times trailing sales. Arguably, the performance that we expect from Twitter would merit a similar future valuation. Applying it to our 2017 sales estimate implies a better than $90 billion future valuation. The current capital environment implies historically low discount rates, but even adjusting sharply upward to account for the risk inherent in projecting sales four years forward, Twitter looks attractive. For example, a 30% discount rate yields a better than $30 billion present value, while a 20% rate would push fair value to nearly $40 billion. Given that the estimated pricing range for the IPO implies a fully diluted market cap of just $11.8 to $13.9 billion, we believe that there is significant upside for investors, even if the deal prices materially higher than the preliminary range.