TWTR: Data is Gold in an AI World
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Artur Pylak
FOR IMPORTANT DISCLOSURES 203.901.1633 /.1634
psagawa@ / email@example.com
September 25, 2016
TWTR: Data is Gold in an AI World
CNBC’s report that the TWTR board was weighing M&A interest from several companies surprised some, as the company’s struggles to drive sub growth have frustrated investors and soured advertisers. Still, we believe TWTR retains significant unrealized value that could be unlocked by a synergistic acquirer. Key assets: 1. Its critical mass of tweeters and followers; 2. A strong start in live video streaming; 3. An underdeveloped but high potential customer service use case; 4. A well-designed ad platform with excellent native formats; and 5. A massive, up-to-date, and unique database of tweets and clicks that documents global interest trends. All of these may have more value for an acquirer – like GOOGL, which could easily generate synergies in combination with its own billion-plus user franchises – than TWTR can realize independently, but the biggest attraction is the data. Supercharged with deep learning, that database and live stream could yield extraordinary predictive insights to enterprises and a greatly enhanced experience for users. We believe GOOGL and FB could derive the greatest synergies, but see real value for many potential suitors, including AMZN, Snapchat, and even MSFT or CRM, at valuations up to $30B or more.
- TWTR is for sale. Unlike other founder-run tech companies, TWTR does not have a two-tier equity structure. While the board is loyal to founder and CEO Jack Dorsey, the languishing stock leaves it little choice but consider offers for the company. CNBC, likely sourcing bankers, has reported at least two such offers are on the table, from GOOGL and CRM. We believe that the interest is real, and that other suitors are likely to emerge, yielding upside to $40/share or more.
- Management has failed. In 15 months as CEO, Dorsey has shown little progress, with the sub count stagnant and the once prodigious ad growth slowing. Promises to ease onboarding and improve access to the most valuable content have gone unfulfilled. Potential users are poorly informed to TWTR’s primary use case as a real time content distribution service. A negative reinforcing cycle of weak user counts, bad press, investor distain, and resulting advertiser disinterest seems to have taken hold. As such, before the M&A speculation began, shares had dipped below $14, less than half of the IPO price.
- The user base is static but valuable. A committed base of hundreds of thousands of newsmakers posting their thoughts/links on a nearly comprehensive array of interests in real-time is at the heart of TWTR’s attractive, albeit static, user community. Would be rivals – FB, GOOGL, Snapchat, AAPL and others – have tried to supersede TWTR, but have failed to deliver its timeliness and completeness in the distribution of news and comment. We believe that a strategic buyer could more effectively resolve the obstacles to growth – e.g. a cumbersome and inflexible app, poorly framed messaging to potential users, etc. – while leveraging its own consumer franchises to extend TWTR’s reach.
- Live video and B2C marketing/service opportunities stand out. TWTR has made an excellent start in live video via the ad hoc Periscope platform and its streaming sports broadcasts, although it faces strong competition from FB and Snapchat. A synergistic buyer could add momentum to these efforts, perhaps combining them with its own video platform. In addition, TWTR has become an effective, but underdeveloped, channel for companies to reach users directly for customer service and proactive marketing. Monitoring TWTR allows companies to flag poor experiences as they happen and, perhaps, address them directly in dialog with the customer. Companies may also choose to reward customers who express enthusiasm for their products. This could be developed much more fully by an acquirer.
- TWTR’s data is gold to the right buyer. In the avalanche of 500M+ tweets per day are buried details about what is interesting to people RIGHT NOW. TWTR already analyzes these trending topics and expressed interests to recommend content and place ads, but the possibilities – e.g. for applying deep learning to predict future trends, to tease out unexpressed interests, to measure the spread of viral messages, etc. – go far beyond that. Those insights could be used to supercharge B2C marketing, to personalize TWTR services, and inspire new services yet to be conceived. Companies like GOOGL, FB, AMZN, or even MSFT and CRM, with strong AI capabilities could see this data as extremely valuable.
- Huge possible synergies for GOOGL or FB. Both GOOGL and FB have the engineering chops to fix the problems with the TWTR app, and the size to provide cover while they do it. Both have billion+ user bases that they could easily lever to increase the reach and engagement of TWTR, and both have the advertising sales engines to exploit that growth. Both have struggled with social news distribution and both are vitally interested in gaining advantage in live video. Both companies have ambitions to serve their advertising clients with more proactive services and could exploit the B2C customer service and outreach potential of the TWTR platform. Importantly, both companies have strong AI capabilities and would be able to apply deep learning to make TWTR’s data resources even more valuable. We think these synergies would justify a the rumored $30B asking price, IF the companies believe that anti-trust regulators in the US and EU would allow an acquisition. Given the perception of TWTR’s failing fortunes, we believe that a deal could be approved, particularly for GOOGL with its perceived weakness in “social” applications.
- Attractive to other buyers too. The potential of TWTR as a B2C channel and a source of consumer insights has CRM on the list of potential buyers, and MSFT should be on the list for the same reasons. Snapchat has obvious synergies – TWTR could supercharge its efforts to beat FB in news and live video – the question is how it would finance a $25B+ acquisition. AMZN would certainly know what to do with the data, although a megadeal would be out of character. AAPL could buy TWTR with its loose change, and the company would fill some major holes – cloud operations, successful advertising platform, livestream video, presence across platforms, etc. Certainly AAPL could improve the kludgey app and integrate the TWTR feed into the iOS platform. Still, spending $25B+ is not a very AAPL thing to do. Media companies – like FOXA or DIS – might want TWTR for its news distribution dominance, but have relatively little real synergy. The same is true for cable and telecom players, but that hasn’t stopped VZ, T and CMCSA from making big deals with dubious synergy in the past.
On Friday, CNBC’s David Faber reported that at least two companies, Alphabet and Salesforce, had approached Twitter with interest in acquiring it. Given the company’s foundering share price, and a narrative of weakness that has spread from investors to advertisers, we expect the board, with pressure from shareholders, to be very open to negotiations. While founder and CEO Jack Dorsey has been opposed to a deal, his 15-month tenure has not produced a promised turnaround, and without controlling equity, he likely does not have enough leverage to repel an attractive offer.
Despite the narrative of Twitter as a failed service and its struggles to sustain growth, it remains a unique platform with substantial potential value to would be acquirers. We see several key assets, starting with its critical mass of committed newsworthy tweeters and their millions of followers. This community, supported by an infrastructure designed for simultaneous global distribution, has made Twitter resilient to periodic attempts, by Facebook, Google, Snapchat, and others, to dethrone it as the primary means of real time dissemination and discussion of breaking news and commentary. While its growth has slowed to a near standstill, it is still uniquely valuable to opinion leaders, interest followers and advertisers.
Twitter is also holding its own in the emerging field of live streaming video. Its Periscope platform pioneered the ad hoc video broadcast, and recent contracts with the NFL and other sports organizations elevates the platform against aggressive competition from Facebook and Snapchat. Twitter is also a leader as a proactive customer service vehicle – companies can find their customers amidst the 500m daily tweets and reach out to resolve complaints and acknowledge fans as their experiences unfold. Again, Facebook is taking aim, but lacks the real time urgency of Twitter. While the company has lost some luster with advertisers in the midst of its swoon, the ad sales platform has been largely successful, with the help of a range of excellent native ad formats.
An acquirer could work with all of these assets, re-energizing user growth and engagement through interplay with existing consumer franchises, expanding on Twitter’s live video initiatives and combining them with its own, giving the B2C service and marketing thrust the attention and resources it deserves and re-engaging advertisers by positioning Twitter as a unique plank in a broader digital portfolio. However, we believe the most valuable asset for Twitter’s suitors is its data.
It’s in the Data
Twitter has built a second-by-second record of consumer interest sliced by personalities, brands, events, geography, demographics, and myriad other criteria. The record – built on tweets and reactions to tweets (clicks, retweets, favorites, etc.) – is unique. For example, Google data does not capture immediate reactions and does not facilitate timely discovery in the same way that Twitter does. Information also takes time to be disseminated on Facebook, and reactions are cluttered by the personal relationships that users have with each other. This firehose of cultural information has, thus far, been under exploited.
Combined with deep learning, Twitter might be used to predict events and reactions to events. It might give deeper insights to marketers, provide leads to sales and service representatives, and greater personalization to Twitter users. It could be used to create new information services that have yet to have been articulated. This is the true pot of gold for would-be Twitter acquirers. Data was seen as the key asset in Microsoft’s $26B takeout of LinkedIn – Twitter has much, much more data with potentially even more value.
We see Alphabet as the company with the most to gain, and while the antitrust review would be painstaking, we believe Twitter’s struggles would ease objections. Facebook is another high synergy option, although the antitrust objections might be more muscular given most analysts’ conception of Twitter as a social network rather than a distribution and discovery service. Microsoft or Salesforce could use the business to consumer marketing and customer service angle to justify a bid, but lack other elements of synergy available to Alphabet and Facebook. Amazon could be a dark horse, as could Apple. Both companies would be able to capture many synergies, but in each case, a $25B+ acquisition would be quite out of character. Snapchat would be an intriguing buyer … if it could raise the money to do the deal. China’s Tencent is another company with obvious synergies, but one that could be blocked by regulators.
Other parties that have been mentioned as possibly interested do not seem to have enough synergy to stay in the bidding. Media companies, like Fox or Disney, would love to keep a major distribution channel out of the hands of the likes of Google or Facebook, but offer little benefit to the core Twitter app and have limited expertise in exploiting its data. The same is true for telecom or cable carriers, and Twitter is unlikely to be a bargain pickup like AOL or Yahoo. We also think the price will be too rich for private capital.
Why Would Alphabet (Or Facebook) Want Twitter?
There are five major areas of synergy for Alphabet to exploit in combining Twitter with Google:
Fixing the App. The Twitter app fails on many criteria – it’s hard to navigate a timeline, to separate disparate interests, to filter unwanted tweets, to search for historical context, etc. It is also famously difficult for new users to get started. Jack Dorsey’s 15-month reign as CEO has not brought changes fast enough. It is likely that Google could do better.
Reigniting Growth. Alphabet can leverage 7 products with more than a billion active users. Twitter could be tightly integrated with search, increasing the prominence of tweets in search results for trending topics and including the service as tab alongside images, video, news, etc. in the main app. It is also a natural fit for Google Now (and the just released Google Assistant) which look to anticipate interests, allowing more timely alerts on breaking news in specifically tailored categories of interest. All of this would be particularly present for Android users, who would see the main Twitter app added as a primary platform service. Casual Twitter users would be automatically registered through Google, and their activity and interests used to enrich ad targeting not just on the Twitter app, but across all Alphabet services. Twitter would also benefit from explicit inclusion in Google’s ad auction platforms.
Accelerating New Businesses. The fledgling live video service could be integrated into YouTube, greatly increasing its reach and its attraction as a partner to content providers. B2C customer service and marketing outreach could gain improved reach via integrated notifications in Android and Now, ties to the Allo and Hangouts messaging platforms, and a more powerful marketing pitch to enterprise partners.
Reducing Costs. Twitter margins are relatively poor owing primarily to its subscale datacenter infrastructure and high employee compensation costs (necessary for retention, given its stock performance). Alphabet should be able to easily resolve these issues (Exhibit 1).
Using the Data. While Twitter has a small but excellent AI team, Alphabet is, by far, the most advanced deep learning company in the world (Exhibit 2). In its hands, Twitter’s data could enable services of extraordinary value to enterprises and consumers, enhancing not just the Twitter service but Google’s core products, and enabling new capabilities that have yet to be defined. We believe that this is the main attraction for Alphabet management as it contemplates an offer.
Facebook would have a fairly similar profile of synergies, albeit with a larger potential regulatory hurdle. We note that the company has tried and failed to displace Twitter in its core use case, and that it faces a new challenge from Snapchat, with which Twitter could be a very useful weapon. We also note that CEO and founder Mark Zuckerberg has been willing to make extremely bold moves – spending 10% of the company’s value on the unmonetized WhatsApp messaging platform, for example. Given the potential competitive risks of an Alphabet or Snapchat deal, we would expect Facebook to be in the mix.
Exh 1: TWTR Valuation Metrics
Exh 2: AI Scientists and Citations by Company, September 2016
What Price is Reasonable?
Before its recent M&A speculation fueled bump, Twitter traded at just over 5 times its trailing sales, a significant discount to Alphabet’s 6.6x multiple. Our base case, which projects less than 1% annual growth in users and about 10%/yr sales growth through 2021, suggests a fair value of about $12B assuming that the multiple holds as margins improve with scale (Exhibit 3-4). However, we believe that synergies – more subs, more engagement driving higher ARPUs, new products, cost improvements, and positive impact on the acquirer’s own business – could more than triple the value for the right buyer. Given the current share price, there is considerable room for a deal that could satisfy the Twitter board and provide significant further upside to shareholders.
Exh 3: TWTR Revenue Trajectory – Base Case versus Takeover, 2016-2021
Exh 4: TWTR Valuation Sensitivity – Base Case versus Takeover