Monday 14 November 2022

Investment Thesis on Meta: Doom or Gloom?

The other day, I expressed my view on Meta on Instagram and invested 8% of my portfolio in the company. MJ from FIRL told me that he also shared his thought on Meta in his Youtube channel. Great video, you should watch it too and see how mine and his view coincide and diverge.

Anyways, I will try to lay down my thesis for Meta.

I still think that the 73% drop from peak to bottom (has since recovered a bit but still 67% YTD drop) is too much an overreaction.

Basically, Meta is facing 3 short term problems and 1 long term problem. 

Now, let’s dive deep.

Meta's Short Term Problem 1: Competition from Tiktok

In 2021-22, Tiktok has surpassed Facebook and Instagram to become the most downloaded app. Users are spending more time there from scrolling addictive short form videos. 

Also in 2021-22, Facebook was losing users for the first time in its entire history. 

These 2 bad news together are a combo, hitting Meta right in the face. This is because user attention is key to an advertiser's revenue and profit growth. The rapid domination of Tiktok shows signs that Facebook and Instagram are losing out.

However, Meta is not doing nothing. They have rolled out Reels aggressively and even faced heavy criticism from the biggest influencer in the world, the Kardashians

The problem with Reels now is that it is just NOT good. The recommendation algorithm built by Instagram sucks and is still way far behind Tiktok's recommendation algorithm, which can accurately recommend what users want to watch in just a few videos. 

And if the growth trajectory of Tiktok continues, Tiktok will eventually grab Instagram users' attention away and this will be the Beginning of the End for Instagram. Not good.

What MJ mentioned in the discussion for Tiktok competition is on point – Instagram need not outperform Tiktok, they only need to match it. This is because Instagram main goal is to keep users that haven’t tried Tiktok from giving Tiktok a try. 

And they can do this by making sure Reels improve to the level that is at least on par to Tiktok's experience, to prevent Instagram users from ever trying out Tiktok. (Reels is not there yet currently, more on this later)

The same thing has happened with Snapchat, whereby Instagram copied the whole approach of Snap and launch Stories. This kept Instagram users from trying out Snapchat. Overtime, the app that gives users more utility will dominate again. And I argue that Instagram will provide more utility to the users, when compared to Tiktok.

Meta's Short Term Problem 2: Apple's  New App Tracking Transparency Feature

In April 2021, Apple introduced a new feature, App Tracking Transparency (ATT) to all its iPhones in iOS 14.5 update. If you updated your iPhone to iOS 14.5, you most probably saw this:

ATT prohibits third party data sharing, unless you allow it. And obviously, no one will click “Allow”. Months after the introduction, Meta announced that the company will take a US$10 bil revenue hit from the implementation of ATT.

The question now is why ATT is so bad for Meta? Actually, ATT is bad for the ad industry as a whole, but most severe to Meta. This is because Meta is the most effective ad platform for business that are trying to build a brand by targeting niche markets. Normally, this kind of business have their own website (most probably powered by Shopify) and when they put ads on Facebook/Instagram, the target audience will be redirected to the website when they click the ads.

Post-ATT, whatever the users do on the website is now all categorised as 3rd party data, which Meta no longer has access to after ATT is implemented. Meaning, now Meta have no idea whether the user bought something from the e-commerce site or downloaded the app. Without all these information, Meta cannot build a strong deterministic measurement of which ad leads to which result. Meaning, advertisers on Meta's platform no longer can know which ad has the best ROI.  

2 consequences from this:

1. Business that use Facebook/Instagram to advertise will see less predictable ad performance. Last time, advertisers on Facebook can more or less expect a certain number of revenue from an ad campaign based on the ROI. 

Now, they no longer have a clear picture on what revenue can expected to be generated from their investment in ads. Because Facebook/Instagram advertising has been so advanced for the past few years, business have also built a system to auto allocate advertising budget based on ROI, hence if the system detects that Facebook/Instagram’s advertising ROI is dropping below their threshold, the advertising budget will be automatically reallocated or even stopped. 

The peculiarity here is that Facebook/Instagram’s ability to convert sales have not changed, it is just the measurement of the ads efficiency that changed, hence artificially lowering the ROI.

2. Users of Facebook/Instagram may see ads that are not so relevant, compared to before. This is because Facebook/Instagram will face some difficulties in user profiling. Meaning, without the data of whether an ad click leads to sales conversion – Facebook/Instagram might face problem understanding what really matters to a user.

But remember this, this is not only a Meta problem, this affects the whole ad industry – just maybe other advertisers like Amazon and Google (not Youtube) face smaller impact because most of their users’ activity happened in their own app/website.

And yes, Tiktok will be impacted by this ATT feature too.

Meta's Short Term Problem 3: Declining Revenue

In MJ's video, they mentioned that Meta has basically captured every market except China. Where else can they grow? Globally the are 3.7 bil people already using Meta's app (Facebook, Instagram, Messenger, Whatsapp) and the world population is only 7.8 bil. 

This is also reflected in the revenue growth for Meta. In June 2022, that is the first time Meta experienced negative revenue growth in its history. It also matches the narrative that Meta has reached maturity and has limited potential for growth, right? 

I believe this is only partially true

We can observe Meta's revenue growth started its downtrend since Sep 2021. That was 3 months after Apple rolled out ATT, so we can say that the effect is actually because of ATT and not because Meta has faced limitation on growth. 

The fact that Meta is able to maintain their current level of revenue (flattish or slightly negative) even after ATT was implemented means that their growth trajectory are still strongly intact.

Also, Meta is purposely taking a revenue hit themselves on promoting Reels. For now, Reels does not monetise as well as the other forms of content. When Meta choose to promote Reels, Reels will canabilise the other forms of content that can make more money for the company. 

So, what if the capital expenditure is successful and Meta did improve their recommendation system to match Tiktok? Then, the disclosed US$3 bil annual run rate for Reels is not the cap for Reels' potential, but just the beginning. 

What actions are Meta taking to solve these short term problems?

The immediate solutions to all that Meta need right now is to:

1. Build probabilistic ad model that can bypass ATT in figuring out who should be targeted and which ads converted, all without the use of 3rd party data.

2. Improve Meta's recommendation algorithm that is capable of delivering what users really want to watch with the whole of internet as an input. 

Previously, Facebook/Instagram only recommends contents generated by people we follow - the inputs are limited. Moving forward, all contents on the platform can be recommended to users - unlimited input

Because of the unlimited input, Meta needs to build a better AI system that can understand what each content (video, image, text) is about and recommend it to users what they really want to watch. The difference between the processing power needed for limited and unlimited input is day and night.

Note that these two solutions are converging into one - building the next-gen AI capability

This is exactly what Meta is doing right now. To quote Meta's CFO words,
"We are significantly expanding our AI capacity. These investments are driving substantially all of our capital expenditure growth in 2023. There is some increased capital intensity that comes with moving more of our infrastructure to AI. It requires more expensive servers and networking equipment, and we are building new data centers specifically equipped to support next generation AI-hardware. We expect these investments to provide us a technology advantage and unlock meaningful improvements across many of our key initiatives, including Feed, Reels and ads. We are carefully evaluating the return we achieve from these investments, which will inform the scale of our AI investment beyond 2023."

Currently, Meta's data centres are CPU heavy, built for the good old days when deterministic ad model is sufficient. At that time, Apple hasnt implement ATT and Tiktok's competition is non existent. But today is no longer the good old days, Meta's future will be heavily dependent on AI and that requires a lot of GPU power.

To power the future of Meta's AI transition, they have put in US$33 bil in 2022 and US$39 bil in 2023 capital expenditure for moving into GPU heavy data centres. The ultimate goal is to build a super strong AI in-house.

By the way, many investors mistakenly see this US$33 bil (2022) and US$39 bil (2023) figures as the spending on Metaverse. That is actually not true. 

Currently, all the spending on Metaverse is categorised as Research and Development under Operating Expense. I still dont understand why they put it there because that is not accurate and will artificially pull down the profit figures. 

Long term problem 1: Metaverse

Meta is investing US$100 bil in developing the Metaverse over the next 10 years, they even changed their company name from Facebook to Meta. 

Everyone knows this, everyone is making fun of Meta’s metaverse investment and everyone is expecting Meta to completely fail on this. This has also been reflected in Meta’s valuation.

This is to no fault the investors' problem. Meta (or Mark Zuckerberg) is the one to be blamed here. The problem with the Metaverse investment is that Mark Zuckerberg talks a lot about his visions of the Metaverse and how he is going to build it, but there is no mention about the business model around it. 

Most investors invest in Meta for their social media advertising business, not for the Metaverse. As a result, the share price reflects what investors think – they think Meta’s Metaverse investment will all go down the drain.

Conclusion / Valuation

I have no idea how the Metaverse investment will turn out to be, but no matter how I look at it, the US$10 bil investment every year is still small considering Meta is a company that can generate US$30-US$40 bil cashflow every year. 

And as mentioned in the previous point, the R&D for Metaverse is directly recorded as an operating expense, meaning the operating profit and net profit of Meta were artificially lowered down

Considering that in and the high likelihood that Meta can solve their short term problems mentioned above, Meta is just trading for around 7 PE at the market cap of US$250 bil (around $US 100/share). 

Am I too optimistic, or is the market too irrational?

Soon we will see. 

1 comment:

  1. Congratulations if you have bought it at time of posting