MBI Daily Dose (July 10, 2025)
Companies or topics mentioned in today's Daily Dose: Commercial query growth in Google, Social commerce, Generative UI
I have mentioned about UK CMA’s report on Google before, but one astute reader highlighted a couple of data points from the appendix of the report that I feel compelled to mention.
CMA report mentioned total monthly search query in the UK increased by ~50-60% during 2017 to 2024 period which implies ~6-7% query growth CAGR.
My initial impression to this MSD+ query CAGR was positive since persuading people to query more tends to be challenging. This quote by Sridhar Ramaswamy, who is a former SVP of Google's Ads business and current CEO of Snowflake, is quite apt in this context:
...It turns out that persuading people to query more is next to impossible. All of us have a certain propensity to use search, and it varies from person to person. There are people, and you can observe this among your friends, as soon as there is a disagreement about something or a question about something, they'll be like, "Let me take my phone out," and they will do this in front of you.
...In general, it's really hard to change this number.
However, it was the another data point that was more interesting. CMA later also reported that Google’s commercial queries increased from 30-40 Billion in 2015 to 40-50 Billion in 2024, which at the mid-point implies a meager 2.8% CAGR. Since the report also shared that total monthly queries hovered between 10 Billion and 25 Billion, let’s take the mid-point of that and multiply by 12 to get to total queries in the UK to be 210 Billion in 2024. This means every one in five queries is currently commercial in nature.
Of course, this is just query, but how about the revenue? CMA reported Google’s real revenue (so not nominal) increased by 130-140% during 2015 to 2024 period, which at mid-point implies only 3.4% CAGR during the period. (Note: some correction here)
Google didn’t report search only revenues for the company in 2015, but their overall advertising revenue in 2015 was $68 Billion. Since they reported $198 Billion search revenue in 2024, the company actually enjoyed ~12-13% CAGR in search during this period.
I have three takeaways from these data points:
a) Even if we add ~2-4% inflation to Google search’s revenue growth in the UK, search revenue growth in the UK almost certainly materially lagged the overall company’s growth in search revenue. Therefore, we need to be cautious not to extrapolate too much from UK numbers to the overall company. (Note: some correction here)
b) I was admittedly surprised at the anemic commercial query growth which grew at less than half the overall query growth. If I had to guess, I would have predicted the opposite i.e. commercial query growth outpaced overall query growth. Data points like this often turn out to be bit of a Rorschach test, but I think it is likely an incrementally negative data point for Google investors. With AI, we will almost certainly ask more questions which will be a natural tailwind for commercial query growth but the extent of such tailwind is likely more limited than anticipated. Most of the queries with commercial intent are already being asked on Google, and it will be harder to move the needle in commercial queries even with AI chatbots. Moreover, commercial queries in traditional search are already super high intent queries, so I’m not sure how much a conversational query will increase click-through rates or CPMs.
c) Of course, chatbots such as ChatGPT are considered preeminent threat to Google’s traditional search. If the nature of commercial query is inherently not as open ended as overall queries, it may be easier for chatbots to gain share in such queries eventually and if the pie of commercial queries grows at anemic rate, encroaching market share in commercial queries may lead to a more pronounced impact to Google search.
Overall, Google remains a pretty difficult stock to bet in size in my opinion. Looking at their multiple, I can tell I am not the only one whose mind gets muddled every once in a while.
Social networking companies have been trying to make social commerce happen in the West for quite some time, but it hasn’t quite worked so far.
Admittedly, I was more optimistic about its chances after looking what my own sister-in-law could do with her Facebook page. She opened a Facebook page right around Covid and started live streaming to sell women’s dresses online. My elder brother thought the business had potential and took a risk by leaving his job to help his wife scale the business. Today, their business makes ~30-40x of what they were earning in their old jobs pre-Covid (they live in Bangladesh, so the jobs don’t pay as well as it does in the US). It’s a bit funny that both my brother and I went into Covid with deep uncertainty about our respective careers and we both ended up building our own businesses.
Anyways, it turns out Eric Seufert was also quite positive about social commerce’s potential, so he revisited why it didn’t quite work in the US:
why are TikTok and Meta reversing course on social commerce? I believe there are three possible general explanations and one that is specific to Meta.
First, the commerce tools offered by TikTok and Meta are inferior to those provided by other retail platforms, specifically Shopify.
Second, Western audiences never acclimated to social commerce, preferring to transact on dedicated retail platforms (such as Amazon) or directly on brands’ websites. This could be a function of a lack of trust, a lack of product discoverability, or simply habit.
And third, retailers simply preferred to operate and drive traffic to their own storefronts, given the benefits of owning the relationship with the consumer.
In addition to "Daily Dose" like this, MBI Deep Dives publishes one Deep Dive on a publicly listed company every month. You can find all the 60 Deep Dives here. I would greatly appreciate if you share MBI content with anyone who might find it useful!
Google recently published a very interesting blog post about simulating an OS itself with Gemini 2.5 Flash-Lite (h/t Liberty’s Highlights). From the post:
In traditional computing, user interfaces are pre-defined. Every button, menu, and window is meticulously coded by developers. But what if an interface could be generated in real time, adapting to a user's context with each interaction?
Our prototype simulates an operating system where each screen is generated on the fly by a large language model. It uses Gemini 2.5 Flash-Lite, a model whose low latency is critical for creating a responsive interaction that feels instantaneous. Instead of navigating a static file system, the user interacts with an environment that the model builds and rebuilds with every click.
By default, our model generates a new screen from scratch with each user input. This means visiting the same folder twice could produce entirely different contents. Such non-deterministic, stateless experience may not always be preferred given that the GUI we are used to is static.
Reading this post reminded me of Hugo Barra’s interview with Ben Thompson in October, 2024. It’s a very intriguing interview and I highly recommend it if you have missed it earlier. Barra had some provocative ideas and he did mention about generative UI and apps as real possibilities for AR glasses. We may be gradually heading towards such possibilities.
Current Portfolio:
Please note that these are NOT my recommendation to buy/sell these securities, but just disclosure from my end so that you can assess potential biases that I may have because of my own personal portfolio holdings. Always consider my write-up my personal investing journal and never forget my objectives, risk tolerance, and constraints may have no resemblance to yours.