Alphabet

Is AI changing the game?

Apr 25, 2023

Welcome to the 102 new subscribers from last week! If you haven’t already, join 9,676 other business lovers and subscribe for free:

Today’s post is sponsored by Stratosphere.io. It’s a financial data platform that we use and it’s awesome for gathering and visualizing financial data for any company you’re researching.

Get started researching on the Stratosphere.io platform today, for free. And you can use the promo code: investingcity for 25% off your first year.

Improve Your Process

Alphabet — no surprise to you — is the parent company of Google. We’ll spend most of our time focusing on Google as it drives the business.

We'll start with its advertising revenue, which is made up of three parts:

1) Search (used to be referred to as properties) - this includes any ads on Google-creations like Google (obviously), Android, Google Maps, Gmail, Chrome, and collaboration tools.

2) YouTube - the company started breaking out YouTube separately in February of 2020. This segment only accounts for advertising on YouTube, not subscriptions or YouTube TV which is accounted for further down the revenue segments.

3) Network - this is ad revenue from "Partners" or other apps/websites that allow Google to insert ads. AdMob is the mobile offering and AdSense is for websites.

So Google’s own properties still account for 72% of all advertising revenue. It’s clear that search is still pretty dominant.

You can see that Google Search hauls in about $50 billion more in revenue than Meta (I still have a hard time not calling it Facebook).

One thing I want to touch on is the cost structure for search compared to network properties. For search, Google has to pay browsers like FireFox to be the default search engine. In fact, Google allegedly paid $9 billion to Apple in 2018 to be Safari’s default search engine, which increased to $12 billion for 2019, rumored to have been $15 billion last year, and I’m seeing reports that it’s closer to $20 billion this year. Since Apple is the other main smartphone operating system other than Android, Google wants to ensure that it remains the default search engine.

While it’s hard, these days, to get the exact proportions of traffic acquisition costs (TAC) between Google-owned properties and network properties, the company gives publishers about 70% of ad revenue (if we use the last time it was broken out). So we can use that to figure out roughly how much Google spends to pay off mobile network operators, other browsers, and device manufacturers.

For 2022, TAC was $49 billion or 21% of advertising revenue and 39% of cost of goods sold. Network revenue was $33 billion so if we use the 70% number, that is $23 billion in costs. Put another way, Google keeps 30% of this revenue as gross profit.

That leaves $26 billion in TAC for search, which means that TAC only accounts for 16% of search revenues (26/162). If we do the math, that means that Google Search is roughly an 84% gross margin business after it pays Apple and all of the other companies to ensure that Google is the default search engine everywhere (that doesn’t include server costs so it’s probably a bit less than that).

This is why the company created Chrome; rather than paying another browser for the default position, it keeps the money. So Chrome isn’t really a money-maker, it’s a huge money-saver. If the company gives $26 billion to other companies to make sure that Google is the default, it makes perfect sense why Google tries so hard to improve Chrome, even going so far as to make Chromebooks.

And this is why the OpenAI/Bing stuff is so nerve-wracking to investors. If the ChatGPT4-embedded Bing can start nipping at Google’s heels and become the default search engine for a lot of folks, that diminishes the company’s position of dominance and likely inflates TAC. Further, it gives Apple more negotiating leverage as it pits Microsoft and Google against each other in a bidding war.

I actually wouldn’t be surprised if Google pays $20 billion to Apple these days. It was reported that Google paid the $9 billion Apple tax in 2018, when distribution TAC was $12 billion, or 75%. On $26 billion in TAC, $20 billion would be pretty consistent with that percentage. Put simply, Google’s distribution TAC has grown quite a bit over the past 4 years and I’m sure a lot of that has to do with Apple; Apple’s negotiating leverage is pretty strong if they can go from $9 billion in 2018 to $20 billion in 2022.

But back to the ChatGPT stuff before moving down to other revenue segments. From Microsoft’s perspective, they don’t have much to lose. Google already owns 90%+ of the search engine market so it’s really an opportunity for Microsoft to do something different to take market share. It actually seems like a classic innovator’s dilemma at first glance but I think it’s hard to argue that ChatGPT is inferior on any performance vector (maybe the fact that you can’t rely on it to get its facts right every time but then again, it’s not like Google is perfect in that regard).

Google does $162 billion in search revenue which is mainly from the blue links that you click —unknowingly sometimes — when you do a Google search. For each one of those that is clicked, Google gets paid. But if a lot of people start using ChatGPT to get answers, then Google loses a bit of relevance. ChatGPT gives answers which is increasingly what Google searches do, but that eats into the sweet high-margin revenue that Google is used to. In a major way, Google has benefitted from the explosion of consumer choices and then it just sits back as those companies compete for the highest value keywords. With more competition, Google makes more money.

This is clearly why investors are spooked. AI changes the economics of search. It becomes more answer-focused than option-focused. And since Microsoft doesn’t care about the margins on this business, it could really eat into Google’s margins as they increasingly try to retain users and provide answers instead of more blue links.

At the same time, Bing has a long way to go before reaching parity with Google. Embedding ChatGPT4 is helpful and things will only get better but one question I have with an answer-focused paradigm is what about trust?

Imagine a future scenario where everyone uses Bing and it uses ChatGPT version 22. You tell it to book you a 5-night stay in San Diego under the cheapest option possible. What happens when Microsoft can make the decision for you? How does the monetization work there? Do you trust it to truly give you the lowest-priced option even if Booking.com is paying them to push people towards its own site? There is probably some happy medium between searches that give you options to choose from and searches that give you the answer. There may end up being categories of searches where you use Google and where ChatGPT is actually more helpful. Google is obviously working on its own version, called Bard, but it does beg the question about margins in the future. Giving the people what they want might be worse for margins but it’s likely what Google must do or else people will start migrating to Bing. I never thought I’d say that. At the same time, I’m confident that Google perceives Microsoft as a threat and is taking it seriously. The sheer amount of data that Google has is an advantage in itself, but lower incremental margins against a competitor who is able to attack with nothing to lose (heads, I win market share, tails, I get Azure revenue) makes for an interesting situation.

It’s hard to truly imagine Google losing significant market share. I see plenty of potential scenarios where, years from now, we sit back and feel silly for thinking that Bing could actually provide a threat to Google. But it’s always a good idea to think about terminal risks and monitor the real-time data. Google investors should take this whole thing seriously but I think we also need to give Google the benefit of the doubt as they navigate their own response and attack. Google very well may lose a little bit of market share but Bing has a LONG way to go to put a dent in the company’s dominance. It’s fun to see a situation that will be a Harvard case study in 10 years play out in real-time. Google shouldn’t be dismissed but ChatGPT poses a real threat that investors need to think about.

Ok, that was a lot! But I think it had to be discussed.

When we move further down the revenue segments, the company also brought in $29 billion in “other” revenue. Only a big tech company like this could fill an under-the-radar segment with $29 billion in revenue, almost an entire Netflix.

This segment consists of Google Play, hardware revenue from Fitbit, Nest and Pixle, and then non-advertising sales from YouTube Premium and YouTube TV.

Then we have the cloud — Google Cloud Platform — and “other bets” which include Waymo, Verily, Calico, X, and CapitalG (DeepMind is actually allocated under corporate costs now in an attempt to be more clear about the company’s AI investments).

So in total, the revenue segments look like this:

So search is the big line item, making up 57% of revenue, which is significantly lower than it has been over the past few decades. The company has made a concerted effort to diversify a little bit from the cash cow that is Google Search. YouTube and GCP (cloud) have grown incredibly fast. You can see below that YouTube has grown at nearly a 30% CAGR over the past five years while GCP has grown more like 45%. Combined, they now account for $55 billion in revenue.

YouTube has been an important part of the company’s story. In 2006, the company purchased YouTube for $1.65 billion in stock. Now, the company does $29 billion in revenue. Talk about an amazing investment!

Here is a chart comparing YouTube and Netflix over the past five years. You can see that YouTube has actually grown a little bit faster, closing the gap on the streaming giant.

Though Google Search only makes up about 57% of revenue, I’m sure it accounts for a much larger percentage of gross profit. Though the cash cow lives on, it is being hunted. Microsoft and OpenAI pose a possible threat but I don’t think Alphabet’s AI capabilities should be taken lightly. After all, the entire reason OpenAI exists is because Elon Musk thought that Google had too much of the world’s AI talent, especially DeepMind, and they weren’t taking AI risks seriously enough. Google is certainly no stranger to AI but if it truly does eat into margins, will the company have the capacity to suffer and heavily invest in the consumer experience? Time will tell but I would be pretty surprised if Google doesn’t come roaring back.

If you enjoyed this, please turn that ♡ into a ♥️. This will help others find this newsletter. Thank you so much 😁

What Else We Offer

For self-directed investors: Dynasty Membership
Long-biased fund: Infuse Partners (intro call)
More content (The Investing City Podcast, online course, and blog)