The leaked documentation challenges Google’s public position that hosting content on a subdomain or in a subfolder makes no ranking difference. Fishkin emphasizes that the documents prove Google can analyze them differently, not that every subfolder will automatically outperform every subdomain.

Google publicly minimized the distinction between subdomains and subdirectories.
The documentation indicates that the systems can treat them differently.
Historical SEO testing often found subfolders outperforming subdomains.
The leak reveals capability, not a universal ranking outcome.

“And this documentation suggests strongly suggests that that is not true.”
Rand Fishkin · 08:00

“All the documentation shows us or proves is that Google does indeed analyze them differently and has the possibility to treat them differently”
Rand Fishkin · 09:00

#seo#subdomains#subfolders#google

Myth Buster44:00

Why AI Often Takes Longer Before It Saves Time

Fishkin cautions that AI frequently creates more work during initial adoption because users must verify results, choose stronger models, refine instructions, and measure accuracy. It becomes valuable after repeated iteration produces a reusable process, but disappointing zero-shot results can discourage mainstream users before they reach that point.

AI output cannot initially be trusted without review.
Model selection and prompt framing materially affect accuracy.
The first successful workflow may take longer than doing the task manually.
Reusable prompts create compounding value after the initial experimentation.

“which is using AI in a lot of cases initially takes me longer.”
Rand Fishkin · 44:00

“And that first time it was probably slower to use AI than it was to do classify the damn hundred things myself.”
Rand Fishkin · 45:30

#ai#adoption#prompting#verification

Hot Take· 1

Hot Take22:30

The Traffic-and-Attribution Era of Digital Marketing Is Ending

Fishkin argues that marketers are losing both open-web referral traffic and the ability to trace individual buyers across channels. Walled platforms suppress outbound links, browsers and privacy rules constrain tracking, and dark social removes referral data, weakening the attribution model that once legitimized digital marketing budgets.

Large platforms control a growing share of online attention.
Social algorithms commonly penalize outbound links.
Referral strings and keyword-level data have disappeared from many journeys.
Cookies, privacy regulation, and ad blockers undermine multi-visit attribution.

“Like this world is ending, and we're sort of grasping to the last few tendrils of attribution.”
Rand Fishkin · 25:30

“And even if you do manage to link out, you won't be able to attribute it.”
Rand Fishkin · 25:00

#attribution#open-web#privacy#platforms

Explainer· 4

Explainer02:00

What the Google Search Documentation Leak Actually Contained

Rand Fishkin recounts how search marketer Erfan discovered a large collection of internal Google Search API documentation exposed through GitHub. The material covered roughly 14,000 modules across more than 2,500 pages and appeared too detailed and internally consistent to dismiss as a simple fabrication.

The source initially approached Fishkin anonymously.
The leak exposed internal API documentation rather than Google’s complete algorithm.
The documentation referenced thousands of modules and ranking-related features.
Its specificity prompted Fishkin to seek independent technical verification.

“That documentation, perhaps not all of it, but a massive chunk. I think 14,000 modules across 2500 plus pages of documentation. That was leaked on…”
Rand Fishkin · 03:30

“And as he shows it on his screen, I am reading the names of the modules and the API calls. I'm looking at the documentation…”
Rand Fishkin · 04:00

#google#seo#search-leak#ranking

Explainer09:30

How Chrome Clickstream Data May Feed Google Rankings

Fishkin argues that the leaked material strengthens the case that Google uses Chrome-derived behavior in systems such as Navboost and Glue. Browser histories and click patterns could help Google distinguish authentic users from manipulation while revealing which destinations genuinely satisfy searchers.

The documents reference Chrome-related behavioral data.
Google can assess browser history and normal click patterns.
The same fraud-detection sophistication used for advertising can protect rankings.
Real-user behavior may help determine which pages satisfy demand.

“That is strongly suggestive that Chrome was indeed created, not exclusively, but primarily so that Google would have full clickstream data for billions of internet…”
Rand Fishkin · 09:30

“So they have to have very sophisticated ways of measuring real user versus not real user.”
Rand Fishkin · 13:00

#chrome#clickstream#navboost#google

Explainer11:30

Why Links on Popular Pages May Carry More Ranking Value

The discussion describes a system that may classify links according to whether real users visit and engage with the source page. A link from heavily consumed content could enter a high-value bucket, while a link on a page almost nobody visits may be discounted regardless of the publisher’s nominal reputation.

Link value may depend partly on source-page usage.
Chrome behavior can help distinguish genuine traffic from synthetic clicks.
Links may be sorted into high, medium, or low-value buckets.
Authority can emerge from real audience attention rather than a domain label alone.

“But one of the most fascinating ones is a structured system for essentially boosting counting or demoting and not counting the value of a link.”
Rand Fishkin · 11:30

“more popular authoritative content that is read by real human beings is going to pass more powerful links than stuff that's not.”
Rand Fishkin · 13:30

#backlinks#authority#click-data#seo

Explainer30:00

AI Could Scale the High-Touch Work That Never Used to Scale

Kip Bodner expects attention to become harder and more expensive to earn while AI makes personalized conversion experiences affordable. Instead of merely generating more traffic, companies can use AI to tailor videos, conversations, emails, and landing pages to each visitor once that scarce attention reaches them.

Website visits may become scarcer and more valuable.
AI can personalize conversion experiences for individual prospects.
High-touch interactions can be extended across an entire customer base.
Awareness still depends on strong research, messaging, and distribution.

“It's gonna be much harder to get a visit to your website.”
Kip Bodner · 30:00

“I think that to your point, AI lets things that don't scale scale.”
Rand Fishkin · 31:30

#ai#personalization#conversion#marketing

Story· 3

Story05:00

How Fishkin and Other Experts Tested the Leak’s Authenticity

Fishkin first asked technical SEO expert Mike King to inspect the material and connect it with patents, research papers, court testimony, and known Google documentation conventions. He then consulted former Google engineers, who said that a convincing forgery would have required extraordinary familiarity with Google’s internal systems.

Mike King compared the material with multiple independent technical sources.
Google Cloud’s documentation style provided an additional authenticity signal.
Former Google engineers privately reviewed the documents.
Reviewers found it difficult to imagine the material being fabricated.

“So I reached out to some ex-Google engineer friends on private messaging apps and asked them if they'd take a look.”
Rand Fishkin · 06:30

“It's very hard to imagine that this is a fake.”
Rand Fishkin · 07:00

#google#verification#seo#api-docs

Story16:30

The Terrible Website That Ranked First Through User Demand

Fishkin recalls an Adventure Time merchandise store with empty titles, image-only content, and exceptionally poor conventional SEO. Despite those defects, it later ranked first for relevant merchandise searches, suggesting that user journeys and Chrome activity helped Google recognize it as the destination people actually wanted.

The store initially had almost none of the expected on-page SEO signals.
It did not appear to have an overwhelming backlink advantage.
Users repeatedly searched for merchandise and eventually reached the store.
Behavioral demand may have helped Google surface the satisfying destination.

“And they had the worst SEO of all time.”
Rand Fishkin · 17:00

“And that Chrome data told Google, guys, this is what people want.”
Rand Fishkin · 18:00

#seo#user-behavior#ecommerce#case-study

Story36:00

Why Products Must Be Designed for Creator Amplification

Fishkin explains that indie game discovery depends heavily on creators rather than traditional search. His studio therefore has to solve two distinct problems: making an enjoyable game and making a game whose moments naturally invite sharing across YouTube, TikTok, Instagram, and other creator channels.

Search rankings offer limited leverage for discovering new indie games.
Creators account for much of the category’s discovery and sales.
A good product is not automatically a shareable product.
Product teams must consider amplification during development, not after launch.

“They are they are essentially 70% of the discovery and sales market in the indie video game space.”
Rand Fishkin · 36:30

“But make a fun game that people will want to share on their YouTube channel and their TikTok and their Instagram reels and their Twitter,…”
Rand Fishkin · 37:00

#creators#product-design#distribution#gaming

Takeaway· 4

Takeaway20:30

The Leak’s Clearest SEO Lesson: Build Demand Beyond Google

Fishkin finds little evidence that marketers can reliably manipulate modern search through simple technical tricks. Because Google can evaluate rich patterns of authentic behavior, the durable advantage comes from building a brand that people actively seek, visit, and recommend outside the search engine.

The leaked systems appear resistant to basic click manipulation.
Authentic user histories are difficult to reproduce at scale.
Technical SEO alone cannot manufacture genuine audience demand.
Off-Google brand building can create signals that later benefit search visibility.

“None of it suggests that you can do a lot of dark arts manipulation, right?”
Rand Fishkin · 20:30

“The way is seeming very obvious, which is build a brand people want outside of Google.”
Rand Fishkin · 21:00

#brand#seo#demand-generation#marketing

Takeaway27:00

Why Digital Marketing Is Returning to Audience Research

The guests predict a return to research-led marketing resembling the advertising world of the 1960s. Marketers will need to understand audience attitudes, craft resonant messages, appear in the channels people already trust, and evaluate results through geographic or time-series lift rather than individual click attribution.

Deep audience understanding becomes more important as direct attribution fades.
Messaging must reflect customer attitudes and emotional motivations.
Distribution should follow where audiences already read, watch, and listen.
Campaign measurement will increasingly rely on aggregate lift.

“I believe that if you want to have success in the long-term future, you need a deep understanding of your audience's demographics, behaviors, and attitudes,…”
Rand Fishkin · 28:30

“And you have to measure it based on time series or geographic-based lift.”
Rand Fishkin · 29:00

#audience-research#messaging#measurement#strategy

Takeaway33:00

Empathy Will Matter More Than Technical Marketing Leverage

Fishkin identifies empathy as a marketer’s enduring advantage: understanding how customers feel, what frustrates them, and why certain product choices or messages repel them. As easy traffic and technical leverage decline, weak products with strong distribution will struggle against brands that genuinely understand their audiences.

Marketers must recognize customer emotions, not only behavioral data.
Product decisions and messaging both reveal whether a company understands its audience.
Distribution can no longer reliably compensate for a mediocre offering.
Brand preference, familiarity, loyalty, and trust become survival factors.

“And I think that is a marketer's greatest strength to be able to feel the feelings of your customer, to put yourself in their shoes”
Rand Fishkin · 33:30

“those types of businesses that have that but don't have brand preference, loyalty, familiarity, trust, they're gonna die.”
Rand Fishkin · 34:30

#empathy#customer-insight#brand#marketing

Takeaway41:30

AI Will Change Marketing Tasks, Not Marketing’s Core Purpose

Fishkin expects marketing roles to change gradually across sectors over the next decade or more, much as search transformed marketing between 2001 and 2016. The tools and daily tasks will shift, but marketers will still need to understand people, shape compelling products and messages, and reach audiences where they pay attention.

AI-driven change will arrive unevenly across sectors and problem spaces.
Future marketers may spend less time on bidding and more on customer research.
Human understanding remains central even as execution becomes automated.
Marketing’s purpose is stable even when its methods change.

“I think that the change will happen little by little, sector by sector, problem space by problem space.”
Rand Fishkin · 41:30

“That's been the job of marketing for you know a thousand years. It'll probably be the job of marketing for the next thousand years. It's…”
Rand Fishkin · 48:00

#future-of-work#ai#marketing-careers#empathy

Frameworks in this episode

AI-Enabled High-Touch Funnel

AI Trust Calibration Loop

Audience-Message-Place-Lift System

Click-Qualified Link Authority

Curiosity-Test-Revise SEO Loop

Empathy-to-Resonance Marketing Loop

Leak Authenticity Triangulation

Product-Amplifier Fit

Task-Level AI Automation Audit

Insights & moments

⚡Myth Buster· 2

Google May Treat Subdomains and Subfolders Differently