Google is About to Enter the Field of Battle — Part 1: Bad Demo Day
Forth, and fear no AI challenge!
Arise, arise, Googlers of Mountain View!
Mice shall be shaken, keyboards shall be splintered!
An AI-day, a Code Red day, ere the sun rises!
Code now, code now, code!
Code for victory, code for Google, and the world’s transformation!
(Inspired by Théoden, King of Rohan, and produced via ChatGPT)
The Search Wars
Google has been seen by some as failing in the “Search Wars.” What an unusual defeat considering that the contest concluded in a week with the result of Google still controlling over 80% of the search market.
The narrative goes something like this, “Microsoft released an AI search product before Google did. Is something fundamentally wrong at Google? Either they didn’t see this coming, don’t have the resources to compete, or are incapable of executing the plan.”
I don’t buy any of those arguments.
Smoke vs. Fire
This perspective is further muddied by two somewhat-related but non-causal events: a mistake in a demo and a $100 Billion stock sell-off. For example, Reuters reports, “Alphabet shares dive after Google AI Chatbot flubs answer in ad”. While both of these events occurred, one didn’t lead to the other.
A Tale in Three Parts
Part I will tackle those events, set context, and learn what really drives these two companies. Part II: Here Comes the Hammer will address Google’s capabilities, challenges, and opportunities. Part III: Predictably Unpredictable is why Bing’s unpredictable behavior was predictable.
Bad Demo Doesn’t Mean Bad Company
The demo snafu. If there has been any act of incompetence on Google’s part, it was their rushed response to Microsoft’s Bing announcement. One part, of one screen, of an unreleased, non-public demo had an incorrect answer. Let’s take a big breath in and let it out.
Ever Demoed Software Before? 60% of the time, it works every time. New technologies have bugs 90% of the time and demos have unexpected behavior 90% of the time (“It was working last night!”). The word “demo” translates from the First Nations Salish language Klalaam to mean “unexpectedly nightmarish experience.” It is an immutable Law of Nature. Just ask any of your software pre-sales friends.
Context: The Money Making Machine
Google Is an Online Ad Seller First and Foremost. We know Google does all sorts of things — search, ads, smartphones, watches, thermostats, cloud computing for enterprises, YouTube, and on. The lion share of its revenue, a full 57% of all money in the door, comes from selling ad placement within its properties. That is primarily search results. If one were to group all advertising revenue from its search properties, network members, and YouTube, it tops out at 81% of revenue ($147 billion in 2020).
Source is Finshots synthesis of January to December 2020 revenue.
Online advertising is very profitable. Today, when Google runs a query to fetch search results, it costs them very little — likely less than $5 per 1,000 searches. They can sell ads for $100 to $300 per 1,000 searches. I’m no mathematician, but when you can do that billions of times, you may need to start burying cash in the backyard. Alternatively, you can buy other multi-billion dollar businesses. Google serves 8.5 billion searches per day.
That’s like, a lot. To put Google’s 57% margin in context with other tech companies, we hear how Apple has big, beautiful, ergonomically-designed margins. Apple has a quite respectable gross margin rate of 43% yet Google significantly exceeds that at a 57% gross margin rate (in the same period).
Here is a VisualCapitalist rendering of quarterly results from June 2022.
Reasons for the Sell Off
The cause for the sell off was two-fold, but neither related to the demo.
Interesting competitor. For the first time in a very long while, Google Search seemed to maybe, perhaps have the possibility of a credible threat. The other reason is Google’s cost to run search queries may increase, thereby decreasing margins.
Costs more to run the machine. Google has chosen to compete against Bing to avoid marketshare loss. These new searches are more expensive than the prior type of searches. If Google switches to use the Bard language model, instead of traditional search results, the per-request cost materially increases. Estimates range from 25% to 10x increase. Whatever the actual answer, the outcome for all scenarios is “costs more than before.”
Can’t pass along costs. While this might surprise some of you, Google doesn’t control the price of its ads. The price is set during auctions by customers. Google only controls the quality of the product and the cost to serve. Under the current auction pricing model, that means higher costs cannot be passed along to customers. Further, it is not entirely clear how paid links will or should, from a user experience standpoint, be interleaved with direct chatbot answers.
Investors are not gruntled. That would lower the overall gross margin rate from the quite high 57%. The result of higher costs and potentially lower revenue makes investors…not happy.
Honeybadger just don’t care. While Microsoft will also incur higher search costs, they process a fraction of the search volume that Google does. At the end of last year, these were the respective marketshare:
- Bing’s search market share: ~9%
- Google search market share: ~84% (but who’s counting, right?)
Here is a look at the search market share in the U.S. as of December 2022.
Microsoft has balanced revenue streams. In contrast to Google, Microsoft’s revenue and costs are not centered around search. It is proportioned between Office, LinkedIn, Azure, Windows, Gaming and Devices. I dare you to find Bing advertising revenue on the revenue stream chart below. What is immaterial for Microsoft is of existential-level importance to Google.
Let’s Talk Stonks
The storyline for institutional investors. Microsoft CEO Satya is playing this up in the press: “This new Bing will make Google come out and dance, and I want people to know that we made them dance.” [Fortune]
He is likely hinting that investors should sell Google and acquire Microsoft stock because, at a minimum, by Microsoft’s actions, Google’s profitability will decrease even if Microsoft doesn’t gain any marketshare in search. Satya wants to see investors shift their portfolio allocation to Microsoft, thereby driving up the share price. Remember, CEO Satya serves at the pleasure of the board. And the board cares about stock price.
When $100 Billion isn’t big money. The day after Microsoft’s announcement, Google’s stock sold off 9% representing roughly $100 billion in market cap. That’s a hard week but the sky has not fallen. Google still maintains a $1.2 trillion dollar valuation.
Good for the Gander? While researching, I found something unexpected — Microsoft’s stock also fell in February. Guess by how much? $100 billion dollars. That’s the same amount of market cap decline Google experienced when the tech press was losing their minds in February. Huh.
Bring It All Together
To summarize what happened in February:
- Microsoft announced and released a material improvement to Bing by infusing it with AI capabilities from OpenAI.
- Google is likely to match those capabilities to foil Microsoft’s attempt to re-energize their search business. That may add cost to Google’s search business.
- Investors recognized that Google might lose some market share and is likely to post lower future margins while battling Microsoft.
- While both Google and Microsoft lost at least $100 billion in market capitalization from their respective February highs, they retain very high valuations ($1.2 and $1.9 trilllion respectively).
- As for the demo, some unfortunate Google demo person is now likely updating their LinkedIn profile. (also, LinkedIn is owned by Microsoft. Ouch).
- Microsoft’s rollout of Bing is going poorly in the press. Like, really poorly.
Next Up!
Now the ball is in Google’s court. Why did it allow a competitor to release an enhanced search product and what are they going to do about it? On to Part II: Here Comes the Hammer.
