Description

This was supposed to be a write-up on Meta. However, I figured it would be a waste once talking about social media to not speak on the other big US players (Twitter, Snap). 

 

There are always two aspects to discuss in social media: the ad market and, of course, the consumer time market.

Just a quick overview of the ad market.

The 2021 global ad market was about $780 billion in 2021, with digital making up about $490 billion. Search makes up less than 200 billion leaving over about 300 billion for general Brand/DR advertising. DR makes up a more significant amount and is growing faster than Brand. Display is taking share from Search and in the US accounts for about 57% of Digital against 40% for Search. Video is about 50% of Display and is taking share from others.

SMBs invest substantially more (all) of their budgets in digital and make up a more significant part than big businesses. For example, Google reportedly makes 60%+ of its revenue from SMBs.  According to the WSJ, meta, with its targeting abilities, has even more at 75% of its revenue.

 

The US makes up $190 billion of this digital ad spend and is about 55-60% of total US ad spend. 

 

Worldwide, Amazon, Google, and Facebook have 60-90% of ad markets.

The global US ad market is expected to grow about 10-15% for the next couple of years, according to Emarketer, or an additional $20+ billion a year with global to grow also double digits but perhaps at a slower pace. Search has been losing share to Display and is expected to continue. The primary beneficiary of all this continued spending will primarily accrue to the most significant players due to very few alternatives of big audiences.

 

TV is still a massive Ad platform around $150 billion worldwide and about $60+ billion in the US.

Although it will probably not fall off very soon, TV spending will likely not grow much with falling audiences. What does fall off will be replaced with AVOD services and digital TV spending. This, however, will not be over time enough to offset the declining users and will eventually accelerate its decline into Digital. It isn't happening faster because of the slow-moving nature of ad markets. This is especially so with big businesses and big ad agencies who have gotten used to the ad-measurement/creation/format. In addition, TV ads are much less targeted than Facebook and are dominated by big businesses. Moving to targeted inventory would force these big companies to be bidding against highly targeted ads over time.

 

Now just a bit more granular.

To better understand these dynamics, you must understand the different types of advertisers. The ad market is very different than other Tam calculations, where one must calculate the demand for a specific good. In general terms, Ads are just a cost of selling and will grow if the ROI gets higher; it essentially does not have a limit. For those economic-rule-based investors, any incremental ROI earned by better marketing will be competed away until the ad price reduces that return.

 

We classify two types of advertisers more granularly:  budget-based advertisers. These are generally big businesses, well-known brands, and not pure eCommerce plays. These companies typically do not act very nimbly, chasing every extra ROI possible and instead budget every year, giving a specific amount to different segments. They are return-based but are usually slower to move, especially the brand-based part of their budget (TV). They are the slowest in moving to Digital and still only give about 30-40% of their budget to digital advertising. These companies also need digital targeting less as they can be very broad-based in their advertising.

However, these companies may experiment more on new platforms even if they have less developed attribution and other services.

 

Then there are the ROI-based advertisers, mostly DR. These are generally SMBs and pure e-commerce plays. These companies are less budget-based, receptive to better returns, and will ramp spending with returns. These companies are often very niche-based and need much Targeting to survive online to create traction to their services affordably. This ad spend bucket is harder to give a growth rate to as they aren't as budget-based and rather react more to returns and margins.

 

There are levels in non-budget-based spending: the best examples of this are Marketplaces with built-in demand. Take 3rd party sellers on Amazon/eBay; these companies have recently ramped up ad tools. This ad spend is entirely incremental and is more a tax on sellers. It used to work that the biggest sellers on eBay/Amazon would be at the top and get the best returns on their accounts. Adding ad services allows smaller accounts with little to lose to bid for those ads and get to the top-selling status. This pool is coming from demand already on the site and is just changing around fulfillment to the highest bidder. 

Now social media companies compete for this ad spend in two ways: user time and targeting/ROI/measurement. 

Even without much targeting investment/measurement, companies with significant global/national audiences will attract ad dollars. TV is an example of this; it will generally be big broad-based advertisers who can hit bigger audiences economically and build their own/3rd party measurement tools and broadly target with their own research. 

These platforms will have lower spending when all else is even due to inadequate targeting/measurement tools that don't allow more bidding/efficiency on demand fulfillment. From there, spending will be more distributed based on the targeting/measurement abilities of those with the most data and the most money to invest in those tools. 

In addition, it will not always be efficient for another specific reason. There is a learning curve to learn to spend on each Platform efficiently, and one needs broad demographics, large size, considerable sales force, and experience by users to drive spending on their Platform.

 This is especially so for smaller SMBs who do not always use big DSPs and instead focus on one or two specific platforms DSP like Meta or Google.

 

Meta will be the biggest beneficiary (besides Youtube perhaps)of this dynamic. They have the most enormous cost structure in RD, the most in-house data from multiple massive platforms. More particularly, Facebook itself is more info intensive than other social sites. They also by now have the most businesses advertising on their site by far outstripping competitors, have a more significant sales force, and have the most extensive reach to different demographics and areas. 

 

Another topic that we must discuss regarding the ad market is ATT. This is why some Social stocks sold off and therefore must be explored.

 

Apple recently announced their ATT in that they are stopping access to the IDFA needed to track users across other apps. In addition to that, they are making rules (enforceability is debatable) regarding sharing PII for targeting.

 

The ad market over the past ten years was built on two things the Mobile device identifiers (IDFA and AAID) and cookies which were accessed predominately by third-party pixels (web) and SDKs (Apps). In addition, URLs got used for tracking App to web conversions. Marketers used these tools for two critical points in the digital ad market: Targeting and Attribution. 

 

Apple released their Skadnetwork to replace Mobile Attribution and will use their Private Click measurement to replace App to web and web conversions.

There are many consequences, of which some are still unclear; we will give an overview of our outlook.

 

The first is on Targeting; there are three fundamental ways Ad networks target ads to users. The first is contextual ads based on the App's content (Dicks will advertise on a sporting blog); this will not be affected by the switch to Skad but is least effective. Then, there are intent-driven and personalized ads, which will get affected majorly. 

Intent-driven is straightforward; a user goes onto Ralph Lauren and looks up a scarf with the user being identified by a cookie placed on the browser or their IDFA, after which the user does not buy it. They then go onto Facebook, where FB realizes who the user is judging from the cookie or IDFA and shows them an ad again for that scarf and can be shown again on Instagram and other platforms. These ads have a much higher chance of driving a conversion because there is a high intent to buy the product. This plays out in both retargeting and other places too. For example, if Meta realizes a consumer is redoing their house, they can again show very targeted intent-driven ads against that. 

 

This style of advertising will mostly get killed for now. Retargeting will be impossible on IOS users, and intent readings from 3rd party sites will disappear. However, the effect on the ad market will not be that much. Targeted intent-driven advertising is accretive to advertisers, especially when the intent is from their site/product but not the Network. For example, retargeting a user can be done by an advertiser very cheaply because no one else will pay the amount they can for that ad. Ads are accretive more to the Network when there is broad appeal and knowledge of that user and their Ltv, not for a specific product.

 

Another dynamic is that this type of advertising is relatively easy for ad networks to build and use and is not very knowledge/compute/model intensive to produce. By taking this away, smaller publishers and small ad networks will lose common ground with extensive self-owned networks in this niche.

 

First-party data will become king for this type of advertising as intent signals will have to come 100% from within an owned app. The more intent signals any social app has, the more value advertisers will have for them on the platform. Apps like Pinterest will be the biggest gainers in this new ecosystem, and apps like Snap (less Discovery) will have the least in this regard. All companies will try to build the most intent signals for their apps, where shopping comes in (discussed later).

 

Then is personalization, which is building up a database of who users are and what they may be interested in. Again, this comes from first-party data and third-party data. 

This gets killed by losing all granular specific user data from third-party apps. Instead, all user data will have to come from on-site, which will cause the info-intensive networks, especially those with multiple types of Apps, to be the biggest gainers (Meta). It should be noted here that Apple allows an IDFV number for cross-app tracking for apps owned by the same Network, so Meta will collectively benefit from all their apps.

In addition, the signals will become less clear than before and will have to be understood by algorithms from posts and many views. In the old days, you saw a user on Nike a lot and understood a big part of his intents. Now that isn't so simple. This again will benefit Networks that have the most to invest in compute power AI/ML models and those with the most data to take apart (Meta).

It will also affect the ability of Networks to understand the value of specific users and limit the ability of Advertisers to target their most valuable users. This is because it limits specific user data and instead aggregates it. In addition, it limits long-term conversion data, which again limits LTV calculations.

What will happen is that Meta and others will try to build databases (perhaps not so specific) of users LTV from many variations of a user being in an Apple campaign ID. This is probably part of why Meta is limiting advertisers to only about 45 Skadnetwork campaign IDs and using the other 55 for itself. This will help them with reporting, and more importantly, get data on users for their ML models. In addition, Android and opted-in users will benefit them in building models to speculate on opted-out users.

Again, this will give the most powerful platforms with the most amount of varying ads across different platforms an advantage in understanding a user's LTV and will also require a lot of ML building and computing power, benefiting Meta and Google.

 

All in all, the biggest losers here are first small businesses, second the small publishers with no data, then the small social platforms with fewer users, less data, and fewer platforms with different use cases. In addition, the biggest winners are those with economies of scale in spending on computing power and AI/ML engineering. 

The ad market is not getting killed because a lot of the ROI loss was in specific users for specific businesses where there isn't a lot of bidding. Even more importantly, short-term advertisers are more the least-worst option minded in delegating ad dollars, especially the big players, which causes ad spending not to go down.

Another big issue that ATT caused is attribution problems. It used to be that a user would click on an ad that captured their IDFA and was then used for conversion purposes when the user downloaded the App and then used it and perhaps engaged in different actions all measured by Ad networks/MMPs.

Now that signal is gone, we have mostly Apple's Skad-Network Private Click measurement and Meta's Aggregated Event measurement. A key drive of Apple is not to let any shared data be used to infer who the user is based on their data, causing issues that will be hard to solve on multiple fronts.

Skadnetwork works by aggregating data to specific campaigns limited by Apple to 100, including any particular data point. They also can get more detailed conversion in a value attributed to particular campaigns. The Ad network or MMP will receive conversion numbers within 48 hours after it happens. Then if users complete more valuable conversions, Apple will have a timer after each conversion of 24 hours to wait for another. Finally, they will share the highest conversion value up to 48 hours after the last was finished.

 For example, a user receives an ad for candy crush and then gets downloaded after clicking to the app store, the advertiser will get that data within 48 hours, and that is it. If the user completes a buy for stars within 24 hours, a clock starts again with that data point shared if no new conversion happens. The advertiser will get the best conversion result (outside of basic conversion) in a metric after it is finished in an aggregated manner.

 

There are multiple issues with this new system, with the main ones below.

 The delay in getting the data, especially the granular conversion data, causes advertisers to be blind and not optimize during the beginning and middle of a campaign. 

Another issue is that conversions will get underreported due to the limit on conversion reporting after the time frame set by Apple is up; this has been something cited by Meta. For example, View-through attribution is limited to a 24-hour window causing view-through conversion reportings to be limited.

In addition, advertisers do not have as much granular data because Skad limits campaign IDs to 100 and Meta limits to 45. It used to be that advertisers would get granular information on what creatives are working and in what region, and to what audience. Now that each aspect of a campaign is limited to 100, that data is minimal. In addition, the fact that conversions are given in very un-granular terms does not let one know the amount of time it takes to convert, how many converts on one user there were, and what their LTV is. This again makes it hard to optimize ad spending.

 

 These problems can get solved in a couple of ways by using extensive modeling and predictive AI. The issue is that the data feed needed to model this is not very widespread. 

The cohorts that are opted in and perhaps Android will become very valuable now for ML models. The issue is that this will be limited. First, users will have to opt into both Facebook and the advertiser app for tracking to work. This causes advertisers with matching opt-ins to be limited in this base. In addition, Google will probably have to follow suit soon, so Android users cannot be relied upon long term.

A very controversial tactic is something called probabilistic modeling. This is essentially fingerprinting in a new name only. They use it by identifying device data points to attribute ads in real-time. The critical difference that the industry is touting is that they do not share specified user data and do not keep specific user databases. In addition, they perhaps do not use so many data signals to get positive reads on particular users. Instead, they are just a probability model on an aggregated group that can improve over time.

 There is a lot of unclarity if Apple allows this, according to the letter of the law. In addition, it is unclear if they will turn a blind eye even if they don't let it because everyone seems to be doing it (though that may incentivize them to stop it as Meta is taking them on).

 

A more innocent way to do this is to use data gleaned from all their campaigns across their Network to predict real-time conversions from seeing what they have historically been from similar campaigns. 

 

In addition, they will use aggregated data from advertisers to see what is happening with users and engagement as they are launching campaigns to predict the effect they are having without linking users.

 

However, this takes a lot of engineering dollars, computing power, and data from both 1st parties and third parties in a privacy-first way, which will benefit the biggest. 

The one aspect that may hurt the more significant players is the limit on campaigns per Ad network and App per advertised App. We are still trying to find out how this aspect plays out.

Another aspect of ATT's effect is limiting how successful and valuable things like Meta Audience Networks will be. This applies to all small publishers that get aggregated by big Ad Networks. These publishers do not have access to much 1st party data and cannot access Facebook data due to limitations of ATT. In addition, Skadnetwork will not report certain reports like the publisher or campaign ID if there is insufficient aggregated data to stop publishers from linking the data to users. This causes publishers not to know their conversions, which due to limited Targeting, will probably be bad either way. This will severely limit smaller publishers' ability to monetize and send the ad dollars elsewhere.

 

All in all, our view is the following. ATT will not affect the broad Digital dollars spent in the ad markets. This is both because although some parts are ROI-focused, many are more the best of the worst option. This was seen in the recent quarter where although there were some misses, most did not have drops in spending. For some time now, it may cause growth in spending to go slower than it would otherwise. 

Over time, spending will likely shift to the best options and out of the worst. This will be a big boon to big publishers with the most significant budgets and the most Data, aka Meta. In addition, it will benefit search-based ads and intent-driven platforms.

 

Now our view on user competition is the following:

There are, in general, three things that social platforms do to get user time: Social points/status/accomplishments(likes), Entertainment, and Utility. 

Most social platforms take off from the first social aspect. It generally starts with teens being attracted to any service they perceive as the new hot Platform and come running. This novel of building a valuable social product is not something that capital can make willingly, like, say, a factory. This is because there is no blueprint for creating a resonating service, and even more importantly, they MUST TAKEOFF. This recipe makes it increasingly harder to create a new platform as you must find a new resonating use case and get that to take off like crazy.

 Many people criticize Facebook for not coming out with a new hot platform since Facebook. This criticism is perhaps, if not entirely misguided, at least not a big question. Many companies have launched a social platform, and most have failed. 

Facebook understands this, so they launched their NPE division (New product development team) in mid-2019. They launched multiple failures with Aux, Bump, and Lasso, to name a few. They are explicitly supposed to act substantially more nimble than Facebook corporate and launch and takedown fast when not working out. They also recently launched some international offices to try and launch in other regions. In addition, they have started to push into making small acquisitions to small companies that have taken off at a tiny scale.

This is all with this understanding that there are Apps that can take off, and most won't, with the capital being thrown at an app not going to be successful automatically. 

They have some outstanding apps, with the most promising being Tuned, an app for couples. They have about 20,000 reviews already, mostly five stars, and had about 30,000 downloads in November. Another promising app is Bars, a rapping app with 10,000 reviews, mostly five stars. If Meta were to hit on another app with a different use case that doesn't integrate well with Facebook or Instagram, the value would be massive. It would widen their moat in copying any successful app with a specific app that fits it.

 

This aspect of social points on a platform is very volatile. A big part of the start of these apps comes from them being on fire as the new hot App. But, when they either cool off or some new app takes the show over, they must fall back into more of the other two aspects: entertainment and Utility. Now, this doesn't mean they won't continue having a hot aspect for a while, but it is hard to rely on it, especially for growth in usage.

 

 The entertainment aspect comes in two ways: proprietary premium content and linear content. Social networks are generally more linear entertainment, with video being the best example. There will be original content on Networks with examples of Discover on Snapchat; these will lure users for entertainment to their Platform for specific content (very valuable). Then there are things like a Facebook watch, which are skewed towards linear watching where you come (or are there already) for general entertainment. This part is not as valuable because it is not explicitly minded, and most big Networks mainly have the same type of content. However, it is still valuable when getting users used to this form of entertainment; most big networks will gravitate towards this early on. Games are the same type of thing but are more premium.

 

Then is Utility; this can come from multiple sources, the first being Social Utility. This means that a network has become a utility for communication in whichever form is predominant on this Platform. Twitter is the biggest, with world conversations on current events being a utility communication. Facebook groups are another prominent example of this. Then there is more general Utility not based on social (but based around it usually); the best examples are Facebook/Messenger/Whatsapp pay and Facebook marketplace.

 

Now onto Meta specifically, the big question is how risky are their Platforms-the fundamental point of the bear thesis.

First is Facebook, which is under the most fire due to recent Wall St views of the non-relevance of the Platform among teens, which the recent whistleblower has confirmed. 

Frances Haugen recently leaked many documents documenting the teenage usage drop on Facebook. It reportedly cited that usage by Teens has been down 13% since 2019 and is expected by Meta to accelerate if nothing is done. The 20-30 aged cohorts were not that bad but were not exciting, mainly a tiny drop. A big issue was that young users were using it less as older people used it more, with those over 30 now using it about 25 minutes longer than those younger. This is causing the perception that Facebook is for older adults to worsen.

However, although we got some specific metrics, most aren't surprised at these data points. It has long (2012+) not been the cool platform for Teens. This is mainly due to a couple of reasons. First is that the Platform was made as mostly a messaging platform with word-based messages. Teen platforms these days are much more picture and video-based. In addition is the fact that many older people have come flocking to Facebook, which caused it to be less the hippy-cool Platform when your 85-year-old aunt surfaces in your feed or comments on posts. The older generation has led Facebook to optimize for its user base and therefore grew more Utility based services both in Messaging and out for their biggest crowd who is less receptive to a new cool feature to gain fame and friends. Then there is the Media that hates Facebook, who continuously attacks them by blowing up any story about them, which causes new users to have a worse impression of them as Evil.

 

Facebook will try to fix this over time but will not overhaul the entire Platform. Facebook is probably one of the least risky platforms due to its significant utility build-up. Most of its users use it for a mix of Messaging, Groups, Entertainment, commerce, Payment soon, and other aspects like news. It isn't the hot place to post and therefore has less risk in that aspect. They are building out these aspects with Groups starting community center-type things and Marketplace becoming better at closing the loop. They will not just focus on Utility for the older generation but ALSO focus on getting Teens back through new Teen features, with Reels being the first. 

In all probability, they will not fully catch fire in the Teen world unless they launch an innovative new feature that can give fame to users on their platform again. But, unfortunately, it also has the risk of overemphasizing a losing battle; this is one of the risks in this name. 

A more critical point is not to get Teens who they can optimize on Instagram but instead get them to come on and slowly build usage as they age. This has a more significant probability of happening, especially as they have made it a key focus for the company (though they have many).

 

Instagram has been a lot better at staying up with the Teen world because they have kept that focus, which Meta will doubtfully lose. For a new platform to kill another platform, it takes mostly laziness and slowness in social media. Instagram will never be able to kill Twitter because it has nothing in common; Snapchat had the power to kill Instagram but couldn't because Insta has the users already and was able to integrate the feature which works into Insta. Tik-Tok took off because both they bought Music.ly, which had some 200mm users, and because Meta didn't react besides Lasso. Lasso was never going to succeed because it was copying a platform that was big already and didn't come with a built-in user base. However, although it was foolish to let Tik-Tok go into the picture, they saved themselves from obsoleteness by integrating Reels. 

Instagram is getting attacked from two sides as they are a more general app. Snap is attacking by being the preferred way to chat with friends, and Tick-tock takes entertainment time away. However, Snapchat is not a significant threat to Instagram as they are not ''hot'' anymore. In addition, there is a new phenomenon on Instagram to either keep accounts private or create double Finsta accounts to use for friends in a private way. This is being encouraged by Instagram and will take on Snap.

TikTok is still hot, but Reels should brake on them and will probably end up being partially accretive to Instagram. 

At this point, Instagram is getting close to 50% of Facebook's revenue and even more of their market cap, so the whole issue with Facebook is likely overblown.

 

The thing that I think gets the least attention here, which is probably the most significant bull case here, is Shops (and Marketplace). Shops transforms two things: one is massively decreasing the risk in Metas Platforms, and the second drives massively higher monetization.

 

First, suppose they manage to make something out of shops(high probability). In that case, conversions will increase massively due to not needing to leave the platform and instead be integrated into the Instagram experience. This will cause Ad inventory to be way more valuable to advertisers. It will also allow Instagram to keep users on their platform instead of linking them away, especially at these most valuable times when they are in the middle of purchasing items. So that inventory of ads for those in purchase mode becomes super valuable to advertisers.

 

By getting users to buy things on their platform Discovery and purchase intent goes up massively on their platform again, making ad inventory more useful and valuable than other discovery platforms (Pins). This will help get more intent info on users and more commercial intent on their platform than now. 

In addition, because of Apple's new policies, those who get the most intent and shopping on their platform will be massively advantaged in having data on user interests. As explained, this is especially so with those who can get cross-platform data from the IDFV.

 

In addition, as has been explained, the most accretive type of advertising is those that take advantage of spending embedded on a platform that can essentially be bid for, like eBay, Amazon, and Walmart. By increasing purchases on their platforms, they will become a big-spending hub where as it gets bigger, there will essentially be a tax on that spending through ad bidding.

 

In addition, it will be valuable from a user qualitative standpoint.

This transformation will de-risk the platform from two aspects one and most importantly is more Utltity intent usage on Instagram that is significantly accretive to the everyday experience. As advertisers push for users to buy things and get used to it being a new platform, it becomes way less subject to the hot platform factor and increases user intent for that specific platform where they found super cool shoes last time. 

 

Over 1mm shops are already on Instagram, which generally leads over to Facebook. However, many do not have checkout enabled, and it still requires a link for those to actually shop. So the challenge is seemingly just to build up all the tools businesses need to run their business on Meta's platforms. However, more structural problems will be businesses who want to keep their customers captive on their websites and have all the customer info and mind share over a competitor. 

Shops have an advantage over other Marketplaces (amazon) because it is meant to be like shopping not in a general Instagram shop but rather in the retailer's online store. This will make it easier for businesses to build a share of mind and not get attributed to buying on Instagram, similar to Amazon shopping. This will probably get more onboard for Insta conversions than on other marketplaces.

Those who will hold out will probably have to do it at the end either way. As businesses get on the platform, two things will happen: First, those businesses will bid higher for ad impressions to convert more efficiently than non-native ads. This will take time until it scales, but those not offering a native shopping experience will get outbid due to their lower ROAS relative to native shops. 

There will be many who will take advantage of shops at first. This will be both general social celebs launching stores and other underdogs. Eventually, all the big players will have to join. The reality is that in the current online environment, most E Commerce businesses drive most of their traffic from digital ads. Meta being from the biggest they will not handle a loss of this platform. 

In addition, Apple's ATT rules will be instrumental in driving businesses onto the platform. They will have much more granular data from Meta on their ad performance relative to 3rd party advertisers. This is probably something Apple did not anticipate when killing off some of Social's revenues.

 

I believe that this will benefit Meta and perhaps the whole industry; however, it will probably help Meta the most. Now there will be many retailers who will be forced to join Meta's shops due to its size in the ad market and the amount of spending it will potentially have on its platforms. However, moving onto other social sites is less likely for some time, if ever (why Pins is not pushing ahead as hard as they should). In addition, Meta is so far ahead of competitors in really launching its shops and driving adoption that it will likely have a much bigger share of mind in this vertical and get the most spent per consumer, which will lead directly to ARPU growth.

 

Once this is successful, it will likely lead to many shopping verticals across their platforms by reconfiguring all the tools built for this and the consumer mindset. They will eventually be able to start selling all types of services on their platforms like paid Facebook groups and others. Essentially it is a springboard for what they call a super app.

 

Then, of course, is Meta's Whatsapp and Messenger, which are two of the biggest messaging apps across the world. A key issue over here over time is that they have been essentially unmonetized and have been eating into the company's bottom line with them not sparing any costs in improving the customer experience. Many investors have made a mistake with Meta, valuing it (cheaply) on current earnings, even with these apps eating into yearly profits. However, there is none who will doubt even after all the hardships in monetizing a messaging service that Whatsapp and Messenger are extremely valuable and not worth negative value. The challenge is figuring out where it will go in its monetization with certainty, especially since this has become a management focus (copying WeChat).

 

WeChat ARPU is probably around $20+ from multiple places, including payment processing financial products (3rd party), minimal ads, and others. This is a rough guess due to them not breaking out WeChat revenue and, in reality, is probably more. Whatsapp and Messenger with 3-4 billion Mau, which we would estimate cost between $3-5 a user, is costing Facebook probably about $10 billion a year. Yet, at this point, it is barely monetized, limited to charging businesses a limited amount to use the WhatsApp business app. 

The company has recently (Messenger in late 2018 slowly) started tap to message ads on FB and Insta for businesses; this is an indirect way they are beginning to monetize with these ads being higher in conversions. These will especially take off with the Apple measurement challenges as it will be easier to track conversions in real-time. They have also recently launched limited ads on Messenger, which will take time to speed up. They will always launch new features on Messenger pre-Whatsapp to see engagement and issues. 

 

Facebook Pay is expanding to Whatsapp and is already done in India and Brazil. They will soon be in Mexico, the Uk, and others. This and Facebook Pay opening up to merchants off Facebook and Shops will cause FB Pay to start scaling in the Global payment processing industry and monetizing a lot through processing fees soon. People do not realize how big FB pay is already at over $100 billion of payments processed or about 10% of Paypal core. This will lead to processing fees and the ability to offer 3rd party financial products that will both monetize and drive platform utility. 

 

People think that Facebook's payment future is dependent on Diem getting regulatory approval which is inherently unsure. As David Marcus said, they could easily focus on making FB pay a global wallet used in many free ptp transactions and monetize on regular processing revenues+. Not only can they, but that is what they are doing. However, they are taking a bet on trying to overhaul the system partially and make a cryptosystem that can accommodate many use cases, including lower international transfer fees (estimated at 7-20% WB) and mini transactions becoming economical. In addition, the underbanked will have access to wallets without using a third party if they manage to figure out the regulatory issues with this (limiting transactions). There are many other use cases for the system, including making it much more interoperable and accommodating Federal Reserve crypto when they become available. However, the reasons cited above will likely make it take off from a consumer standpoint, which is always needed first.

 

If it is successful, it will not just benefit Meta, especially since they cannot and are not taking control of the Diem Association and will be under a lot of scrutiny to not make much money off it. So the most they will gain is that it will benefit the entire Social ecosystem and perhaps them a bit more for integrating it first.

 

The regulatory issue is that it is in style for Congress to stand up to big tech, specifically Meta, which will cause them to slow it down. This is especially so as they have moved strategies from pursuing Swiss certification to focusing on the US to launch. Senator Brian Schatz put out a letter with others urging Meta to shut down Novi (confusing it with Diem) now. They are biased against Meta as many of their complaints are unfounded (I won't get into it now).

 

What is becoming clear is that Management is trying to use India and other big non-developed countries that use Whatsapp a lot to launch many new services and then move it over. For example, they have recently partnered with Uber to get Ubers on Whatsapp in India besides payments. In addition, they are reportedly looking to launch insurance products in India soon with third-party providers. 

There is also talk that they are planning on Launching Reels (tremendous if successful) to Whatsapp though they will have to do that in a roundabout way not to ruin the experience.

 

All in all, there is little reason to believe that they cannot move towards getting to higher monetization on their messaging apps. They doubtfully will get close to the $150 US ARPU in FB and Insta but can cover their single-digit costs plus a lot. This is especially so as they managed to make Status Updates such a big thing in Whatsapp even while keeping its messaging identity which will be a crucial part of monetization.

 

In addition, as they scale new Utility options in their messaging apps, they will make it substantially safer from new entrants like Telegram as it is significantly harder to launch a messaging app like WeChat than Whatsapp. Besides, Messaging apps generally have the most robust social network effect in the industry.

 

The next significant cost in their cost structure that gets penalized as a reduction of earnings power is their VR/AR costs which are projected to take out about $10 billion of corporate profits in the current year. 

There are two points: first is how successful it will be. The second is that even assuming no success, this will most probably be clear in about five years or something. They will likely stop investing in this new area by then, and it is not a fixed cost structure of their social business.

There are two distinct strategies in this segment built on relatively the same type of technology. First, their VR segment is building an ecosystem like the Book/Movie Ready Player One. There will be an entirely new Metaverse world where people can hang out, do different activities, have different digital experiences, and many new use cases. This is more likely to succeed through the degree is very unclear. They have initially launched on some easy-to-sell use cases, primarily social (most popular) Gaming and exercise. By doing this, they have been relatively successful with having sold about 10+ million Quest 2 headsets since coming out in October 0f 2020.

Besides building out the ecosystem from their studio, they are investing heavily to build out 3rd party developers. Best Saber has already grossed over $100mm showing how even early developers can develop apps and be profitable doing it. 

The Metaverse is substantially different from the smartphone ecosystem as it is being built at least from Meta as a highly interoperable system. The idea is that buying something on one App will be transferred to any app across the ecosystem, making it much more limited in having many different ecosystems. There will be a minimal amount of competitors, mostly probably Apple and Meta though the other tech giants will try. Meta is perhaps the most invested and later in the game, having started a while ago.

Although the platform took off already, it will gain much traction as more people use it; developers keep developing, and Meta makes the Hardware and the use cases better. For example, they are coming out with Cambria, an add-on feature used for two use cases. The first is that they have developed a mix of Virtual and a real-world API, allowing mixed case VR experiences. The Cambria device will fully reconstruct what you are looking at in the real world and incorporate it into the virtual world. In addition, it is being built to give over an accurate eye contact and expression experience as in the real world. 

The exact amount of success here is unclearly ranging from the Metaverse taking a life of its own to many notches down from that. 

They are already far into the investment and way past proof of concept. I would recommend watching the most recent Facebook connect event for anyone that didn't.

 

Their AR ambitions are inherently riskier from a Tech standpoint than from a consumer adoption standpoint. There are many challenges in their Nazare project, which is building out full-blown AR glasses. This will be less of a whole new concept but rather a hugely new way to interact with technology while being in the world. 

 

To understand the risks, I'll  quote Michael Abrash, who is leading this for Meta, "It's going to take about a dozen major technological breakthroughs to get to the next-generation metaverse, and we're working on all of them: displays, audio, input, haptics, hand tracking, eye tracking, mixed reality sensors, graphics, computer vision, avatars, perceptual science, AI, and more." 

And Mark Zuckerberg "We have to fit hologram displays, projectors, batteries, radios, custom silicon chips, cameras, speakers, sensors to map the world around you and more into glasses that are about five millimeters thick. So, we still have a ways to go with Nazare, but we are making good progress."

This new project from Meta will allow them to de-risk the company by not relying on 3rd party hardware and opening up new massive monetization channels. Additionally, it will enable them to break into the Hardware market to build more devices that potentially interact well with their ecosystem. 

Another big perhaps underrated plus is that they will hopefully be able to separate their brand from being associated with Evil Facebook to having a whole new identity-making, their core, and new platforms better.

 

Using Horizon as its Metaverse Social brand, they are building the first full-blown social ecosystem from building out a house where one lands in the Metaverse (similar to Apple building out Imessage with special APIs). They will meet people there and travel together into a different world. In addition, they are building out a workplace ecosystem that allows people to work together in this world. The actual Horizon Apps that have come out of beta recently are relatively limited but would have been expected from new technology. 

 

Just to reiterate, I am not saying that their VR/AR unit will change the world. Instead, they will probably spend about $50-100 billion on building this out (much more taking in 3rd party investment) by the time it becomes clear how successful it will be. For VR, the real risk is not the technological feasibility; the core technology is mostly in place, with them having to make it a much richer experience. The real question is consumer adoption, where with any new technology, especially one that attempts to change the world, there is a considerable risk of consumers not fully adopting it. 

However, there is little doubt that it will become a relatively big platform, perhaps at the end, mostly in Gaming. Therefore, they will recoup their investment to some degree if it fails to become what they believe it will be. It is already in its early days in well over 10 million houses which should continue to increase. If 50-100 million people adopt it, it would probably look like the Microsoft Xbox division. They sold some 80 million Xbox 360s with gaming revenue of about 15 billion in 2021, primarily attributed to Xbox; essentially, they should not have such a massive loss to this investment. 

Their AR investment which is in a big part related to their VR investment, probably has less of a risk of more considerable adoption not happening. However, it has more of a chance of being useless if it doesn't take off without a gaming/cool social experience. But, even more, there is still a risk of the tech not being feasible.

 

The point is that investors are discounting this investment currently as a fixed part of their cost structure even though it isn't and will probably not sustain a loss even if not successful.

 

Meta expects to finish the year with about $116 billion in revenue on the lower-end guidance range. In addition, they are expecting to have total expenses of $71 billion on the higher end, leaving over $45 billion in operating profits for the year. They will spend about $9 billion more than GAAP DA on Capex, which leaves over $36 billion of free cash flow. They are projecting a loss of $10 billion from Meta reality labs. In addition, they are probably losing about $5-10 billion on their messenger apps as they are minimally monetized as of now. Assuming only a $5 billion loss from messaging, they will make about $51 billion this year on an EV of $870 billion or at about 17 times the current year's earnings. 

 

The critical pushback here is that they plan on scaling expenses and Capex in the coming years. This year, they plan to scale costs about 28-38%, likely outstripping revenue growth by some amount though unclear precisely as it is a relatively volatile environment here. In addition, they plan on scaling back up Capex here to about $30 billion, further reducing reported profits. 

There is a fundamental difference between companies like Meta and its cohorts and many other companies that Value investors are used to. In a manufacturing company ( and most others), a company's cost structure is primarily what it costs them to earn that year's revenue and is, in actuality, an actual cost structure of the company. There are two types of costs in social media. First are the fixed costs. For example, the general hosting costs are relatively variable on user growth and usage. Based on usage, this cost is fixed and stays stable even as monetization increases from better conversions. Then are the marketing costs to service advertisers. These are variable per advertiser (especially big ones) but less than hosting. 

Then there are the engineering costs that build new features and the system's back end. This part is where there is a lot of uncertainty. The Social space works on two dynamics first is relative ROAS and absolute. Relative primarily drives short-term revenue shifts in the space, and absolute drives longer-term growth. When companies increase costs to increase ad spend (or user growth), it will appear to be a less profitable company.

 

Take Twitter as an example who is generating about a 10% margin relatively these days. They are investing a lot of their cost structure in driving higher ROAS, especially as they try to get DR market share. Their current cost structure lets them generate their current revenue plus growth. If they decide to scale costs 20% higher over revenue growth with a clear plan, most people would say that an investment in Twitter is much less an opportunity than they will say now that you can put a PE on them. However, if they do that, there is a good chance the firm's value goes up rather than down as it will allow better ad market share in the future. Manufacturing companies don't have this problem as they invest in new factories; it is relatively straightforward what investment capital versus maintenance is; however, it becomes more challenging in social. 

 

Many know this effect regarding Amazon, which historically has not posted massive Gaap profits. However, even Amazon has this more minor than the social tech space, which has almost their whole cost structure in engineers and AI processing abilities which drives better results over time against distribution channels.

Meta has this substantially the most with by far the most investment in RD outstrips competitors by multiples (taking out the need for relative investment); this becomes even more important in the ATT world. In recent years, much of their cost structure increase has come from security and content takedown. This is primarily an automated function where they are now having well over 90% of content taken down being done by AI, which will increase as time goes on. Those engineers go to different apts if they aren't needed; it isn't a fixed cost structure.

 

In reality, when a company like Facebook grows by, say, 20% from more ad demand, it is almost entirely incremental. However, they are in a very lucrative industry, and there is little reason not to reinvest. 

There isn't an answer to how to value these companies, but these ideas must be considered when thinking about them. 

To somewhat quote Joel Greenblatt on this (from memory), "we don't quibble with valuations on the Faang type companies; they are incredibly phenomenal one of a kind companies."

 Meta has historically spent double D''A on Capex, mainly for servers and offices as they grow worldwide. However, recently it has jumped up because of AI investment. There are two reasons for this overinvestment. First, because Meta uses their own servers (against Twitter and Snap who use AWS and GCP), it costs in the short term more. For example, say the average server lasts ten years; if their user usage goes up 20%, they would have to invest in new servers equal to double their current depreciation besides the maintenance/replacement Capex.

 

In addition, they are investing much more into AI and ML to understand users and site content better to serve ads better and make the platform more valuable and personalized. This takes a lot of computing power but will make their services better across all aspects and become a significant competitive advantage against smaller players. 

Onto Snap.

Snapchat was believed to be left for the dead in 2017-2018 when the company slowed materially in user growth due most substantially to Facebook's launch of stories on Instagram. The company was at that point still burning substantial amounts of cash and was supposed to be in a growing stage where they were financing growth. 

 

What ended up happening was a real turnaround where they managed to reignite user growth even in developed markets, which they have kept up since. This resulted from emphasizing original content and shows and the company executing exceptionally well on quickly fixing the android issues it was having and keeping to innovate in the face of Facebook. They launched  Maps, Highlights, Discover,  Minis, and Games. Those points allowed it to keep growing and were aided with its continued substantial usage among the Snap generation 13-24, which gave it a continued niche.

 

 At this point, the company has essentially said that they are at the point where they are not actively working on adding new features to their platform ''with the structure basically finished'' and will instead focus on enhancing and monetizing the whole ecosystem. So the key questions here are will they be able to hold their own against the new platforms and Facebook copycatting and the degree of monetization possible from their platform.

 

The company currently has an enterprise value of 75 billion to put things in perspective. In addition, it is presently on pace to do about $4 billion in revenue this year after substantial growth over last year. To grow into its current valuation, it would probably need to earn pre-tax $2-3 billion depending on what you believe the risks inherent in the industry are. To get there, we would estimate they need a top line of about $10 billion or 150% growth from where we are today. 

 

The rationale behind this 20-30% margin prospect is the following:

  At this point, Snap seems to be increasingly leaning on original and Paid revenue share content. The motivation is that they are both trying to differentiate themselves from the bigger scaled competitors that keep copying their ideas and increase usage with older demographics. This is seen with Snap citing the substantial growth of Discover content in 2018 with the separation from Snaps, and in 2019 they called out 50% growth of 15+ minutes users, and in 2020, they mentioned 70% growth in Discover usage. This has coincided with seemingly stable use at about 30 minutes per Management's comments. In the last two calls, Management started pointing out how Discover is cannibalizing time spent with Snap postings and views going down. They cited that the lockdowns are affecting them-not a very good argument as Teens are not likely to be in lockdown mode at this point.

 

To use original content to get users to come back to your platform for the SPECIFIC show they want is a brilliant strategy. It turns the platform from a distribution platform that is easier to substitute to a platform with content that is getting sought out. However, that's if it doesn't become the main attraction and is used to get usage up across the platform. 

 This would make the service's economics substantially different from the historical, social media companies with the substantial amount spent on original content/revenue share. Yet it is still unclear what the longer-term ramifications would be on gross margins. Snap is interesting because they do not store much data with Snaps off their servers within 30 days. This is perhaps the cause of Twitter's Cogs per user being 40% higher than Snap, even with seemingly more minor usage in original (and highlights) content by Twitter than Snap. From a longer-term perspective, it is a con on Snap's business model that they have less knowledge about their users over time than competing networks, especially Facebook. However, this was partially offset by having a more robust theoretical margin profile when scaled anywhere near Facebook. The leaning on Discover will take away from this advantage if over needed.

I do not mean to say that it is an uneconomical business, just that the 40% margins that Facebook had at about 10 billion of revenue becomes less certain for Snap. If you model an extra 8% of cogs and 7% of RD  above 2014 Facebook, you hit about 25% margins. This is not overly conservative given a higher need to offset Apple challenges and pull away from incremental Facebook and higher safety costs, causing them to scale higher. 

There is also another question here regarding monetization in each country. Snap currently monetizes Europe substantially less than Facebook did in 2014 at somewhere near a third (complex to calculate given different reporting). This may take away margin potential if they do not get that up. However, this may be offset by ROW ARPU close to Facebook's back then and will likely be surpassed soon, especially with India's recent rapid growth.

 

If you agree with those assumptions, the critical question now will be how they get to that $10 billion. There are two options: monetizing use better and increasing usage through more prolonged use or more users.

 

Our thoughts on their monetization prospects are the following:

This year they should do somewhere near the mid 30 ARPU on US users, still substantially lower than Twitter which should be about $70+ this year. Despite Snap seemingly close to Twitter in average usage (Emarketer though others put Twitter lower). Twitter has been around longer than Snap, but historically, due to multiple reasons (discussed by Twitter write up), it had not captured share from the DR market, which is more significant than Brand. So there is very little reason to believe that Twitter's prospects are better than Snaps, especially with Twitter being so bad at executing across product development and DR targeting/measurement/self-serve. The only question will be if you believe Twitter's older male-skewed audience is a better target than the Snap younger generation female skewed one. If Twitter was more direct response focused, perhaps there is room to believe that, but given that it is Brand, that becomes a hard argument. DR may work better with people who have more money to spend, i.e., Twitter's audience. Brand marketers tend to be more long-term-minded and be big global brands. They would find it extremely valuable to get the young Snap generation who is habit-forming. In addition, given Snap has about double the audience in the US than Twitter, it would make us believe Snap can get more marketers on board. This is because it is more economical for big advertisers to create a system including teams/measurement/targeting tools for a bigger platform.

Twitter's only thing is that it is more a differentiated advertising platform than typical Facebook/Insta. It is more conversation-driven, leading to more relevance to advertising there. An example is a recent debacle with Peloton/HBO Max with their just like that show. After HBO Max showed an actor having a heart attack from a ride, Peloton released an Ad on Twitter where there were tons of conversations about it, leading to the soaring Search for Peloton and good feedback, something no other platform could do.

 

On a more granular note, Snap does not yet monetize Highlights as they are trying to improve the experience beforehand. They under-monetize Maps which is at the beginning, and they still have room to build on their Games/minis platform, which has only recently launched. Their CPM is also significantly lower than Twitter/Facebook/Instagram having 2-3 times upside on that metric. This came as the way these things work because they do not release ad inventory until there is more demand. It isn't easy to get marketers onto their platform from others, and it takes a cheaper rate and hard marketing. They have, therefore, slowly released ads as the demand came forward, which resulted in either stable or sometimes dropping CPM. This is even more pronounced as they get better at targeting/relevance and engagements; CPM should go up as CPE go down but didn't. The lack of tools/newer platform/ basically caused it to lag. 

 

From this perspective, all it takes is a 50% increase in CPM and a 50% higher ad load (highlights/maps, general), and they get to the $10 billion targets without any increase in audience.

Their prospects for user growth are less than for higher monetization.

 

 They had substantially slowed down when Instagram started stories for really two reasons. First, those who specifically went onto Snap because of that feature were less driven to go on now that they had it on Instagram. This, in turn, caused a lot of the hype in the ''new hot platform'' to go down. That hype is critical to getting a real-scale platform, especially with better competitors. Especially with Tik-Tok's rise, Snap faces even more significant issues, with the big driver of its platform being that it is primarily seen as the Gen-Z App Tik-Tok is taking away from that. This is not an exact comparison as Tik-Tok is used more broadly than Snap, with about 50% of its US users under 30 against 60% of Snap 24 and younger.

 

They accelerated at the end of 2018 due to the fix on Android and them executing well. However, they did this for only a bit more than a year before covid hit, making it unclear how sustainable it is. In addition, they are at a structural issue at this point. They have effectively captured much of the developed western world 13-34-year-olds at 75% in the US, UK, France, Australia, Canada, and 90% of 13-24 with engagement a lot higher there; 20% of their users are under 18. To scale upwards as these users age will be a big challenge. A big part of Snap's current success lies in their focus on the younger age groups, with much of the product focused on the pictures only and AR being a lot more for the two specific examples. It may take time for these people to start churning off as they age and may give them a boost for some time, but it will be a challenge over time. Suppose they decide to pivot and accommodate all age groups (Instagram and even more with Facebook). In that case, they face losing their leading competitive position against Facebook and may destroy their image as the younger generation's App. 

Their developed markets will probably see some small mid-single-digit growth in the next few years (Emarketer). Germany is the only big Western Europe market with lowish penetration at about 18% and has and will grow a bit faster than the average. However, it will not be that much incremental profits in the bigger picture. They have lower monetization, with a doubled usage, probably leading to some $75 million total earnings from rough estimates.

 

 A key question then is whether they will expand well internationally and how profitable will that be. They are currently huge in India, with around 30% of their users. However, India likely provides minimal incremental profits as monetization is substantially lower than gross (though it must be kept in mind they probably cost less to service). However, this will likely change soon as both Snap, which as of now has reportedly had a hardish time cracking the market, should take share as their recent surge makes them more viable there. However, this will take longer from both monetization and a user growth point of view to move the needle.

In many Asian countries, China being the first, Snow that has just focused on the AR capabilities of Snap has made a decent penetration. However, they seem to have slowed down with a fraction of Snaps downloads recently, the cause being them seemingly focusing on a Selfie App over social.

 

On the riskiness of its Network, I would vote Snap as having from the highest. Snap is mainly dominated by the young Gen-Z crowd, with older people just a hangover from more youthful years. This causes that its usage comes most substantially from social factors, with it being the best place to get so-called social points on the web. They tend to use networks less for Utility and instead for a mix of social use and less substantially for entertainment. This heightens the risk of stagnating in User growth or risking losing its status as the Gen-Z App. I would believe they would not accept slowing growth and try to keep older users with new features. They will have a tough time doing that both because they are more similar to Insta, which will do a better job than them on anything that goes viral and fits its platform. The only thing Snap may be used for in the older generation is the disappearing messaging for things that are generally not a basic universal use.

 

One last point is that the Apple privacy issues will probably affect Snap the most out of the whole industry. They are currently substantially skewed in revenue terms towards DR; they have considerably less knowledge about their users than competing networks, and not much purchase intent is shown, unlike Pinterest. Then most importantly, besides the fact that they have substantially more data than Snap from in-house users, Facebook also has considerably bigger Rd budgets, which will be critical, as explained earlier in getting marketing dollars.

 

The one thing Snap talks about a lot is their innovative AR capabilities. The main attraction of this more substantially than users putting smiley faces on is the ability to use it to try on clothing pre-purchase and engage users on trying different brands pieces. The problem is that you can not rely on Snap to have an innovative technology against Facebook for long if successful Facebook revenue scale and marketing share would allow them to substantially copy this feature and get most of that revenue to themselves.

 

All in all, we have held back for the time being on investing in Snap. We believe there is more uncertainty here, especially from recent management comments. Snap is a lot more social-based than entertainment based like a company like Instagram. It is more about Teens communicating and keeping streaks going. There is less browsing for content like Tik-Tok and Instagram. Recent moves to highlights and shows may de-risk the business if successful but create more uncertainty, with Snap having a more minor edge there. Especially from the recent comments by Management that Snaps and Snap views have gone down to be replaced by regular viewing of Highlights and shows, exacerbating our uncertainty. 

The stock is not overly expensive at this point but does not hold enough upside for it to be worth the risk. This is especially so as we find other opportunities in this space, and one must be mindful of overconcentration. However, a significant enough drop may open opportunities.

 

Onto Twitter.

Twitter went public in 2013 under high expectations and had since really disappointed investors. The company had mainly disappointed investors with its user growth which slowed substantially and even dropped soon after its IPO. The company has also been criticized for being slow on rolling out both consumer products and slow monetization products. 

 

Recently, the company has been very vocal about its slow past and how it's now focusing on innovating. To quote Jack Dorsey-" It comes down to three critiques: we're slow, we're not innovative, and we're not trusted." The company has been pushing that they have been dead on innovation because of mainly them rebuilding and enhancing their core product, making it more relevant to users, and improving core ad products by rebuilding the ad server completely. However, most believe that it was also a failure in execution, probably because Dorsey was less than a part-time CEO. This is generally attractive as an investment when a company is trading based on lousy performance caused by execution and not structural issues.

 

Twitter is substantially more interesting in this space than a typical social platform due to its nature as more of a Utility/entertainment than a social platform focused on getting the most social points like Snapchat. This is seen without discussing its use cases from its demographics. Over 75% of Twitter users are 25+, and about 70% are male. This is substantially different from most social-based platforms, which are generally more skewed towards Females. 

 

Jack Dorsey describes the platform as a discovery platform. This is probably only partly true. We like Dantley Davis's description- "Twitter is what's happening, and what people are talking about right now. It keeps people plugged into conversations they care about and topics they want to learn more about." 

This includes investors who follow many prominent investors and industry participants to hear real-time about different opinions and news. It is also used as an instant news outlet for many (about 20% of users use it for news higher than most platforms besides Facebook). A stat I recently heard how every social media platform goes down in usage by significant events (Super Bowl) besides Twitter which goes up, brings out the point.

This is very important as it significantly de-risks the platform and allows its Network effect to work more robustly than is typical in this industry. This is seen in the past; in general, a social media company is hard to take down because, given their scaled audience, anyone who makes a similar platform with something extra to take down the old one will lose as the old one can quickly replicate that use case. Myspace allowed Facebook to beat it at its own game, which caused it to go down. Twitter has been pretty bad at executing on new products and innovation/copying other platforms, yet it has stayed up through all that. 

Given the above, new ''hot'' platforms like Tik-Tok have a more muted effect on Twitter than a social/entertainment platform with more specific-minded intent when coming to the App over others.

 

For context, the company has an EV of about $30 billion and nearly $5 billion in annual revenue. With a projected 15% medium-term margin target which has been within the range recently, the stock isn't a screaming bargain from a glance but isn't too expensive either. The way we frame this type of thing is that the company would need about 1.-1.2 billion of earnings power to support valuations without fantastic growth prospects. The current estimated earnings power from the company is 15% margins on about $5 billion in revenue or about $750 million. This is consistent with 2018-2019 margins and management comments. Recently margins took a hit, mostly from cogs driven by higher usage and insufficient ad revenue to offset this. 

 

As in all social media, the key questions are how well they will increase users without risking current use cases and their prospects for growing monetization for existing users.

 

In regards to user growth first:

 

In 2015 soon after going public, user growth started slowing and then declining slightly (perhaps), not just in the US but in international markets too. This was incredibly concerning because of the low penetration into Developed economies and the inability to grow internationally, which should have been easier. This has burdened the stock. However, in 2019 Growth started accelerating, which carried significantly into the pandemic. This was happening in both developed economies like the US, where many competitors were not seeing growth and emerging markets. In addition, engagement has gone up significantly in emerging markets where it was closer to 20% (MAU/DAU) in 2014, rising to 40% in 2018, and has since not been reported. 

 

Obviously, increased usage during Covid must be discounted and not seen as a key sign of changing long-term prospects. However, this picked up already in 2019 where DAU's relative growth doubled from the year before. Without a specific reason to believe usage should continue to increase, we would generally not put so much weight on a single-year performance (as is the case with Snap); however, Twitter is a more peculiar case. 

The company has a massive top-of-the-funnel user base with about 2 million people coming to the site every day that have not visited at least in 30 days. That is about 600mm prospective long-term users every single month. In addition, the company's MAU/DAU is lower when last reported at about 40% than the average social service at 60%+. From that perspective, they can get another 100mm users, many of the US (most valuable), if they get normalized. The question is why they have not converted all this into DAU.

Given the unique aspects of Twitter where it is not used to follow random people with cool pics/videos and instead is used as a network to get good news/conversations/opinions on things people are interested in, it makes conversion harder. If you ask some people, they will tell you Twitter is super interesting/helpful, and others will tell you it's nothing. The core reason for this is it is hard to make the service valuable because it takes a long time to curate a beneficial feed by adding people and taking out the bad ones. 

This has two effects: it takes longer to convert people into longer users. The second is that those users who work on it become valuable longer-term users who find the service very valuable.

The main thing which makes Twitter very interesting now is that late in 2019, they started Topics. This started with about 6,000 and kept expanding to more granular topics. This allows people to follow specific topics they are interested in and allow Twitter algos to keep learning what gets people interested in particular topics and what doesn't. This will create more and more value over time on a never-ending cycle as they grow the granular topics and learn more and more about how to feed them.

This will essentially drive conversion way faster if done well by making it supremely easier to get a valuable feed. It will also help find the most helpful people to follow when put into the feed. This will also create more value for people posting good material by increasing followers similar to Tik-Toks algos.

Twitter is also now launching Twitter communities. Although this is more a copycat of Facebook groups, it fits nicely into Twitter's use case. It will allow people to have more granular conversations on specific topics they find interesting, allowing small topics not to get squashed by more popular Twitter topics. 

You always must differentiate when a site is launching a new feature that may be an additive feature but is not solving a platform's pain point and one that solves it. For example, Insta Reels does not solve an Insta pain point but is a well-integrated future. These two new Twitter products specifically solve pain points on the platform and solve this long-term conversion issue on Twitter.

 

   In addition to the above specific top-of-the-funnel opportunities, Twitter has a structural driver to get new users. There is no doubt that many more people benefit from Twitter than are Signed in DAU. This is a crucial point over most other social media companies. Even if many say they get all the news-worthy Tweets from watching the news, this is still a long-term structural driver of new users to their site, which causes them to be prospective long-term users. Other social sites have this attribute also, but Twitter is the strongest here. This will allow over time if Twitter solves the conversion issue well enough to drive significant long-term user growth.

What is also interesting is that Twitter is relatively more prevalent in more specified regions, most notably the US and Japan. Since it is somewhat less penetrated in the US, we believe they can continue growing nicely there and, in extension, other developed countries. Its growth at this point is not about India, which monetizes a lot lower than the US/Japan. This is not to say that it isn't valuable to grow in international markets, which Twitter can improve on; rather, their growth prospects are worth more than other social sites because it has room in developed markets. 

 

Now onto its monetization prospects:

 

There are a couple of crucial questions to answer over here, which is the following:

 

The consensus on Twitter is that it has lagged competitors on monetizing its platform. We disagree partially; Twitter in 2020 was monetizing at about 1/3 of Facebook Blue on DAU and is actually at about 50%+ higher than Snap's monetization of its users. So although User metrics across platforms are not perfect, it probably isn't far off.

 

 Internationally they have lagged competitors with a more significant spread at about 12% of US ARPU against Snaps 20%. However, it is even worse than this simple metric. The Snap number does not include the EU, higher at closer to 30%. Twitter does have EU mixed in there. Twitter is dominant in Japan; if you take out Japan, which is at about $20, it is even worse. This is even more surprising given that it usually takes time to tackle the international markets, which are generally done after really making headwinds in NA markets. Twitter took off about four years earlier than Snap and would have been expected to move faster there. This, we believe, comes from three different areas. One is pure execution, where it takes a lot of sales investment to break in with ad agencies and work hard to explain and help them optimize their campaigns. Twitter generally has been quite bad at execution since Dorsey took over again and was focused on other areas rather than building out its international sales force. 

Second is Twitter's inability to crack the DR market, with about 85% of ads sales coming from Brand against a 50/50 split for Snap and Pins and a 70/30 split for Google and Facebook. This has essentially caused more prominent corporations who focus a lot on brand advertising which is inherently harder to measure and requires scale, to focus on Twitter's platform. These companies have their measurement tools to optimize TV and Media ads used for the Twitter platform.

 In addition, both that and the underinvestment in Self serve ad platforms caused SMB to be only about 15% of ads sales. Twitter underperformed Facebook on this metric is not hard to understand. Facebook has more info on users, a grander scale internationally, bigger RD budgets allowing better Self serve and targeting, and measurement tools. Its difference against Snap is the biggest question and is attributed to two things. First is again execution; it takes a lot more focus on building tools to get very good at DR response, and Self serves to get SMB in scale. More prominent companies targeting big audiences will use Twitter to get good returns. Twitter has said in the past (backtracking now) that they have worse intent signals than some other platforms, which causes DR again to lag. 

One last point is that Twitter's scale in international markets besides Japan causes a more challenging time driving marketers to its platform as they do not have the required scale needed. For example, Twitter runs at less than half of Snaps users in France and Germany, about 35% less in Saudi Arabia, with the UK the only one they are close to. 

 

It is unclear that Twitter will drive monetization much higher than the industry average in the US. Their CPM is relatively in line with others, and CPC is seemingly much lower. This comes from the fact that Twitter is 85% brand advertising and only 15% Dr, primarily CPC. It is unclear if the Brand CPM has cannibalized the outstanding ad inventory with not much left to go towards CPC. Baird analyst Colin Sebastian asked this question to Management at the recent analyst meeting but didn't get a detailed answer (do they ever) besides that, no, they will be able to drive more accretive DR revenue. However, it would seem this is true as MAP revenue and impressions are substantially up recently because of Management ramping up its offerings (it's half DR revenue).

 

On international, they are mostly solving all the problems cited above. Execution is a more objective aspect; however, there are multiple reasons to believe it is improving. First, they have been ramping new products faster than they had in the past. In addition, employee attrition has been down 60% recently, even with hiring up. Most importantly is Jack Dorsey leaving. Whatever you think about Parag (we can't comment), it is better than before. Parag has reportedly mostly been running a big part of the company, which is inefficient given Jack's official CEO. 

 

Their scale in international markets is getting better. They have been growing international DAU at double digits in recent years, which should only improve from here because of better conversions and increasing scale. These things work because as scale increases, there is a flywheel effect until they get penetrated (relatively). So Twitter still has more room to grow in many international markets.

 

In addition, their new DR initiatives should help push internationally and SMB (a product of improving DR). This aspect is a bit more objective but the recent focus laid out by management points in the right direction. There is no reason to believe they will come to par with Facebook, but Snap should not do this that much better. Twitter lagged in driving this, which was partially a product of mismanagement. This is seen from multiple places; for example, they have also only recently launched targeting on their promoted Tweets and their pixel only recently launched (ridiculous). 

DR seems especially promising now that they are pushing their topics that seemingly were not used for targeting. This becomes very important because Twitter has always lagged in info on users and Apple's recent privacy issues. Management has said they will start integrating this into targeting abilities (crucial for DR).

 

The risk that Apple's privacy policies pose to them is perhaps the least in the entire industry, with 85%+ of revenue coming from Brand and not DR, which is not built on the number of measurement tools DR is. It affects their future ability to target relative to what would have been, but we believe it may be a net plus. With increasing limited ability to see outside of the App, Companies will have to utilize their info to develop competitive ad products. Twitter historically has been at a disadvantage on this because there is a lot less purchase intent shown on the platform, even relative to, let's say, Snap. However, in recent years, the companies have pushed into topics and communities where people get more granular on things they are interested in against general tweets; this puts the company in an excellent position to know interests. Now that companies must turn to internal data to create the ability to target, it will make significant investments in ML to take apart pics and messages to better understand interest and purchase intent. Facebook will dominate at this, as has been explained due to their scale. Twitter's ability to directly drive these signals because it is a learning platform will be essential to go INCREMENTAL DR ads. 

 

A fascinating monetization tactic by Twitter from subscriptions has recently come to the forefront. They have launched two primary services: Twitter blue, a premium offering, and second super follow, allowing creators to monetize big followings. This is interesting from two aspects:

 Twitter competes on user service and Ad service to get ad dollars from competitors. By opening up subscriptions, they have found a way of monetizing without that part of the competition. Second and perhaps more interesting is that it will probably cause usage to go up. Many investors say that they should give all Blue features to all users to enhance the experience. If it were just a way of getting new monetization, perhaps correct, it would drive better UX. However, by doing this, they have done something clever. 

A paid service will always get used by users more than a non-paid service because of a phycological aspect of getting your money's worth. However, Twitter could not start changing its model to a paid model as that would have been impossible. So by doing this, they are still essentially keeping open the free model but give those that don't mind a premium experience for pay. This will have the effect of stronger retention due to higher usage and higher monetization due to more ads and the sub-payment. This can lead to ARPU of US users doing this to about double from where we are without improving the ad business.

 

It will be hard to say whether it will be successful in driving subscriptions without management comments. However, we do not believe there is much need to speculate because the thesis does not rest squarely on this success.

More granularly:

 What they do seem to be doing is targeting different user patterns that can get value from it. First are the prominent Tweeters, and they give them the option of undoing a Tweet. Then are the big readers, and they give them multiple tools to save threads and better format them. Then are the big news guys where they give them ad-free access to partner articles (with a revenue share model with third-party publishers). These do not need to be overwhelmingly valuable, but rather, it is a relatively cheap way for someone who found these as pain points if they don't mind the money. Given that Twitter demographics are wealthy and more educated than the average American, they can get many who don't mind the $3 a month.

Their super follow is probably, even more, an engagement play than a monetizing play. They allow $50,000 to go mostly free and take 20% after that. They will need a massive amount to be processed to move the needle on this one. Super follows also increase the risk that the high-value Tweeters start charging, lowering the average user's value. However, all in all, we don't view that as a significant risk. There are two types of Tweeters: First, those who like having a voice; these users will probably not charge as it isn't their purpose. Then there are the focused monetary ones. They will also be driven not to alienate too many users to monetize other ways without this. They also want to have that funnel to get more subscribers and will probably focus on giving value to both.

 

An aspect of Twitter makes it perhaps a bit riskier. Content is more concentrated in the top Tweeters because it is more an information/blog service than a social one. This is seen with Pew research that suggests that 97% of Tweets come from the 25% most active account. This causes they are more dependent on the top users rather than all in all. However, this also has advantages as their Network effect does not work as much in reverse as others do.

 

The recent resignation of Jack Dorsey as CEO is something most investors and we are pleased about. However, he was not putting his most into this company. He had substantially less of an actual and monetary interest against Square, besides the issues of running two public companies. Parag Agrawal appointment is yet to be seen how it will turn out. Some are upset that an insider was chosen. We are unsure if that is a bad thing.

 

Catalyst

Earnings growth