Latest / Deep Values: Value Investing Deep Dives / Google AI Killed The Click, Not Ads
Transcript
- 0:00Welcome to the deep dive for, I mean, what felt like two
- 0:03straight years. The loudest story in tech was
- 0:08this impending doom. Oh yeah, the consensus was
- 0:11clear. AI was going to kill Google.
- 0:14It was this huge existential threat to a company worth, you
- 0:17know, over a trillion dollars. The whole idea was that
- 0:21generative AI. These models that give you 1
- 0:23perfect answer would just make search advertising obsolete.
- 0:27And you can see the logic, right?
- 0:29The fear was, I'd say, pretty ironclad at the time.
- 0:32Google's entire business model is built on clicks.
- 0:36High volume, low friction clicks.
- 0:37You type something in, Google points you to another website.
- 0:40Exactly. It sends you away and that
- 0:42external site runs adsense ads. They share the revenue and the
- 0:44whole system works. But if AI just gives you the
- 0:47answer right there on the search page.
- 0:48You never leave. You have no reason to click.
- 0:50Away. And if there are no clicks or no
- 0:51external page views, and this is the critical part, there's
- 0:54nowhere to put those display ads.
- 0:55The whole narrative was framed perfectly by these competing
- 0:58models. You had Open AI coming out with
- 1:00their? Was it a $20.00 a month
- 1:02subscription? Right.
- 1:03And everyone said that's it. A finite subscription will
- 1:06always lose to infinite ad revenue.
- 1:09But what's so fascinating, and what are sources for this deep
- 1:12dive show, is that the argument was completely backwards.
- 1:16How? So that $20 subscription, well,
- 1:18it's finite. It has a hard ceiling on what
- 1:21you can earn from a user. But advertising, I mean
- 1:24advertising that stimulates actual shopping, new
- 1:27transactions, new economic activity that's functionally
- 1:30infinite. And the data proved it.
- 1:32We're looking at Google's Q 32025 earnings, and instead of a
- 1:35collapse, ad revenue was just booming.
- 1:39It's incredible. The core search ad revenue, the
- 1:40very thing everyone thought was on the chopping block, was up
- 1:4315%. 15% and YouTube ads, which are in this fierce fight with
- 1:47TikTok. They were also up 15%.
- 1:49Overall ad revenue jumped 12.6%. So they didn't just survive this
- 1:53thing. They thrived.
- 1:54But here's the most telling part, the smoking gun, if you
- 1:57will. There's one number that did go
- 1:58down. OK, Well, was it network
- 2:00advertising, Right. That's the Adsense revenue that
- 2:02goes out to all those thirdarty websites.
- 2:04It droed by 3%. Ah.
- 2:06So that small drop tells the whole story.
- 2:08It tells us Google made a huge internal decision, a strategic
- 2:12choice that some of our sources called, and I love this phrase,
- 2:16a violation of ancestral rules. A.
- 2:18Violation of ancestral rules. OK, let's unpack that.
- 2:21How do you get record advertising revenue while
- 2:23cutting off the cash flow to your partners?
- 2:26It was a complete philosophical U-turn.
- 2:28For decades, Google's sacred mission was to be the perfect
- 2:32traffic cop. Get people off Google as fast as
- 2:34possible. Exactly.
- 2:35If you stayed on the search page for too long, it was seen as a
- 2:38failure, a bad search result. Success was measured by how fast
- 2:42you clicked away. And now they've, you've
- 2:43completely inverted that idea with the AI search integration,
- 2:46this, you know, AI mode or chat interface.
- 2:49The new goal is to keep you there.
- 2:50It's all about retention. They want you engaged, solving
- 2:53problems, having these long conversations, all within the
- 2:56Google interface. So it's no longer about finding
- 2:58a door to another website. It's like they've built a fully
- 3:01furnished room right there on the landing page.
- 3:04With ads, of course. That's a perfect analogy, and it
- 3:06completely removes the need to send you to a third party site
- 3:09just to see an ad. User.
- 3:11Dwell time just skyrockets, and so do the opportunities to show
- 3:14an ad. Hold on, let me just push back
- 3:16on that for a second. If I'm spending an hour in a
- 3:18chat with Google's AI solving some complex work problem,
- 3:22that's an hour I'm not clicking on publisher websites.
- 3:24So while Google is keeping all that fat water for itself, isn't
- 3:28there a risk they alienate all those publishers who built their
- 3:31businesses on that traffic? That is the crucial question,
- 3:34and that 3% drop in Adsense proves that.
- 3:37Yes, the partners are seeing less traffic, but the massive
- 3:4015% increase in their own search ads proves the strategy is just
- 3:44overwhelmingly profitable for Google.
- 3:46They prioritize their own platform.
- 3:47They've decided the most valuable thing is the user's
- 3:50time spent with them. The big problem AI introduced
- 3:53wasn't, you know, AGI or sentience.
- 3:56It was a much more practical question, how does AI make
- 3:58money? And Google's answer is still
- 4:00advertising. Just a totally new kind of
- 4:03advertising. One that targets what you call
- 4:04the ultra long tail. Exactly.
- 4:06Keeping the traffic was step one.
- 4:08Monetizing it is Step 2. OK, so the long tail is a
- 4:12concept. We know platforms like Amazon or
- 4:15Taobao making it possible to sell niche products globally.
- 4:18Right. The Internet lets you sell, say,
- 4:21a very specific type of artisanal coffee bean to a few
- 4:24thousand people scattered all over the world.
- 4:27But AI is creating an even longer tail, an ultra long tail,
- 4:32meaning it's satisfying these tiny specialized demands that
- 4:35were just too complex or too low volume to advertise against
- 4:39before the friction was too high and.
- 4:41We have to remember how high that friction was.
- 4:42I mean traditional digital advertising was a brutal multi
- 4:45step process. Oh, it was a nightmare for small
- 4:47businesses. Just think about it.
- 4:48Step 1. You, a human, had to figure out
- 4:51your key selling points, your keywords.
- 4:53Then Step 2, create the assets and not just one picture.
- 4:56You needed videos, images, GIFs all in like 10 different sizes
- 5:00and you had to keep refreshing them constantly because of user
- 5:02fatigue. And the competition was just
- 5:05insane. In some markets, competitors
- 5:07were using AI to generate millions of ad creatives just
- 5:10testing constantly. This super extreme play they
- 5:14called it. How does a small merchant even
- 5:17compete with that? You can't, and that's before you
- 5:19even get to Step 3, which is setting budgets, CPC, CP/M,
- 5:24targeting by location demographics it.
- 5:26Was a full time job. And then Step 4, the platform
- 5:30has to collect all this user data, which you know comes with
- 5:32its own privacy headaches just to match an ad.
- 5:35It was exhausting. So this is where AI comes in and
- 5:38just wipes the slate clean. It eliminates all those
- 5:41barriers. For the merchant, it becomes
- 5:42almost laughably simple. Our analysis shows they just
- 5:45have to agree to an automatic product entry.
- 5:47And Google's system does the rest.
- 5:49It does all the heavy lifting, it scrapes their product
- 5:51descriptions, photos and that is brilliant, the user comments
- 5:55because that's where the real context is.
- 5:57The AI then figures out the key selling points on its own.
- 6:00Exactly. No need for a human creative
- 6:03director or that super extreme play.
- 6:06But the real genius is what happens on the user side.
- 6:09By keeping you in that chat window asking all these nuance
- 6:11questions, Google develops a long term memory of what you
- 6:14want. It's so much deeper than just a
- 6:17few quick keyword searches. And that deeper intent allows
- 6:20for what you're calling non standard matching.
- 6:22That's the key to the whole thing.
- 6:23That's the whole thing. Let's get into the case studies,
- 6:26because that's where you see it in action.
- 6:28OK, let's start with the small batch perfume, a niche product
- 6:30that would normally get lost. Totally lost in the old system.
- 6:34It's competing against massive brands bidding on keywords like
- 6:38perfume or Cologne. It doesn't stand a chance.
- 6:42But in the new AI mode, a user doesn't search for a brand, they
- 6:46search for a mood. Something like cool woodsy
- 6:49scent, not too sweet, smells like rain.
- 6:52Which is almost impossible to target with keywords.
- 6:55But the AI isn't looking at the manufacturer's description.
- 6:58It's pulling keywords like Woodsy and Rainy Day from user
- 7:02reviews for that perfume. So it connects the user's
- 7:04feeling to another user's description of that feeling.
- 7:07Precisely. And then it auto generates a
- 7:09specific ad for that exact mood. It's advertising the experience,
- 7:14not the product category. The targeting recision is just
- 7:18phenomenal. The running shoe examle is even
- 7:20more incredible. It's about synthesizing
- 7:22knowledge. OK, so a user searches for
- 7:24something really technical. 4 foot strike running shoes, wide
- 7:29toe box, no knee pain. That's already very specific.
- 7:32But then the AI also sees that this same user has been looking
- 7:35at map routes for half marathon training.
- 7:38OK, so it's not just someone who wants trendy sneakers.
- 7:41This person is a serious runner. This is a training runner with
- 7:44specific biomechanical needs. The AI knows that, so it
- 7:48synthesizes information from technical reviews and runner
- 7:50comments, finds a shoe with, say, a low drop and stiff
- 7:53cushioning. And it serves an ad that focuses
- 7:56on those performance features, not an ad that says 20%.
- 7:59Off. Exactly.
- 7:59It's an ad that says this shoe solves your specific training
- 8:02problem. The conversion rate on that must
- 8:04be astronomical. In that final pet food example,
- 8:07it really shows how AI can map symptoms to solutions the user
- 8:12searches. Cat frequently scratches ear
- 8:14soft stool after changing food. Those aren't shopping keywords,
- 8:18they're medical symptoms. The AI understand that
- 8:21distinction. It maps scratching ear to
- 8:24product features like low allergen or single protein, even
- 8:27if the merchant's website just says pet stop scratching.
- 8:30So instead of a generic coupon for cat.
- 8:33Food the user gets an ad for a low allergen single protein
- 8:36trial pack. It delivers the exact solution
- 8:38they need right in the chat and that has a huge economic impact.
- 8:42Right, because the conversion rates are so much higher,
- 8:44advertisers are willing to pay more.
- 8:46Google's cost per click the CPC, it rose 7%.
- 8:50And since total ad revenue was up over 12%, you know, a big
- 8:53chunk of that came from advertisers paying more for
- 8:56these incredibly precise placements.
- 8:58It's a positive feedback loop. Higher conversion means you can
- 9:01bid more, which makes Google more money, and it also means
- 9:04more small businesses can compete.
- 9:06You don't need a huge marketing team anymore.
- 9:08You just need a good product. The AI handles the rest.
- 9:12We saw the big picture effect of this in the 2025 Black Friday
- 9:15report. That was fascinating.
- 9:17Orders were actually down 1%. Which sounds bad, but the
- 9:20average price of each order was up 8%.
- 9:22And when you filter out external factors like tariffs, which were
- 9:25about 5%, you're left with a 3% price increase that's attributed
- 9:29directly to higher ad spend. It shows that profits are being
- 9:33funneled right back into this high precision AI advertising.
- 9:36It's just a market shift. We're moving away from mass
- 9:39market impulse buys. Toward targeted high value
- 9:43consumption. The market isn't chasing volume
- 9:45anymore, it's chasing precision. Which means Google has
- 9:48completely redefined its value. It's built on three interlocking
- 9:51layers of strength now. OK.
- 9:53Layer one is the tech. Their Gemini model is leading
- 9:55the pack and their custom chips the TP use are seriously
- 9:59challenging Nvidia's dominance. Layer 2 is the ecosystem.
- 10:03They're rebuilding everything they own Android, YouTube,
- 10:06Gmail, Chrome around Gemini that keeps you inside their system,
- 10:09constantly exposed to these new ad formats.
- 10:11And layer 3 is what we've been talking about this whole time.
- 10:14The core logic AI changed the method, it killed the simple
- 10:18click. But the goal of commerce is
- 10:21still the same. Transaction and delivery.
- 10:23Google is just the new hyper efficient middleman for that
- 10:26transaction. Which sort of reaffirms that
- 10:29basic product economics are still the engine of everything
- 10:32people need to work to buy things.
- 10:34Right. And since our consumption of
- 10:36physical goods, steel, clothes, food, is ultimately finite, the
- 10:40value that absorbs all this new growth has to be something else.
- 10:43Content and stories. Content and stories AI let's
- 10:47creators make much more differentiated, specialized
- 10:49content. If AI can find one person who
- 10:52wants a rainy day cedar perfume. It can find a few 100 people who
- 10:55want to read a very niche AI assisted sci-fi novel.
- 10:59And suddenly that becomes a viable business.
- 11:01The old model of making money from high volume clicks is dead.
- 11:05Google One because it embraced AI's power for non standard
- 11:08matching. Attaching new value, new stories
- 11:10to products. Exactly.
- 11:12So the final thought for you, the learner is this.
- 11:15The market isn't about being the biggest anymore, it's about
- 11:17being the most precise. AI is incredible at eliminating
- 11:21the friction and creation and advertising.
- 11:23So the question becomes, how can you use it to create your own
- 11:26highly specialized small scale service piece of content?
- 11:29How can you create that unique value that can now be profitable
- 11:31in this new global AI mediated system?
- 11:35The door to the ultra long tail is wide open.