SEO traffic is falling and small website owners want answers. The honest answer is uncomfortable: Google owes you nothing, the algorithm is math, and the post-SEO era has already begun. Here is what that means and what to do about it.
SEO traffic losses are accelerating, and for the first time in years the cause is not technical SEO problems, content quality, or backlinks. The losses are driven by structural changes inside Google itself, chiefly the rollout of AI Overviews. Losses that were once recoverable are now permanent for many sites, and the complaints filling every SEO forum share one theme: this is unfair. It is not. This article explains why, and what comes next.
What Google’s Latest Updates Actually Did to SEO Traffic
Google’s Helpful Content Updates (HCU) and subsequent core updates did not kill traffic by penalising bad technical SEO. They reshuffled ranking signals at a systemic level, rewarding demonstrated expertise and brand authority over individual page optimisation. We have covered the mechanics in depth in our HCU recovery guide and our core update research piece, so this article will not repeat that ground.
What is new this cycle is the AI Overviews feature. Where previous updates changed how Google ranked existing blue-link results, AI Overviews change how many users ever see blue links at all. An answer delivered inside the search result page is an answer that never generates a click. That is the mechanism behind the traffic losses that feel inexplicable to site owners who have done nothing wrong technically.
Why Calling the Algorithm Unfair Is the Wrong Frame
The loudest complaint from site owners who lost traffic in 2024 and 2025 is that large media outlets gained while independent creators lost. That observation is accurate. The interpretation, that Google is being unfair, is not useful and arguably not accurate either.
A ranking algorithm is a series of mathematical operations applied to signals. It has no preferences, no grudges, and no awareness of your business model. The question worth asking is not why is the algorithm punishing me but rather which signals does the algorithm weight heavily right now, and do I produce them?
Consider the algorithm’s perspective. Given two pages on the same topic, one from an independent writer with modest credentials and limited external recognition, one from a major publisher with named expert contributors, an editorial process, and thousands of external references: which page is statistically more likely to be accurate and useful to the majority of searchers? The answer is not personal. It is probabilistic.
The algorithm was never designed to be fair to content producers. It was designed to be useful to searchers. When those two goals aligned, independent creators thrived. When they diverge, independent creators lose. That divergence is now structural, not cyclical.
Google owes no website organic traffic. If you genuinely object to how Google operates, a robots.txt disallow directive will remove you from the index entirely. That option exists. Most people will not use it, which reveals that the traffic is valued more than the principle.
Traditional SEO No Longer Works the Way It Did
Search Engine Optimisation, by definition, means optimising for search engines. That objective has not changed. What has changed is what search engines now return in response to a query. When the answer lives inside the results page, optimising for a ranked blue link is optimising for a diminishing outcome.
AI-powered search engines, including Google with Gemini, OpenAI’s ChatGPT Search, Perplexity, and others, are built on licensed content from established publishers. The partnerships below are not rumours. They are signed agreements that determine whose content trains these systems and whose content gets cited in AI-generated answers.
Confirmed OpenAI Content Partnerships
OpenAI has secured licensing deals with the following publishers, giving ChatGPT and related products privileged access to their archives and real-time content:
- News Corp: Wall Street Journal, New York Post, MarketWatch, Barron’s, The Times, The Sunday Times. Five-year deal reportedly worth over US $250 million.
- Reddit: Agreement to incorporate Reddit content into ChatGPT and related products.
- Vox Media: The Verge, Vox, Eater, The Cut, New York Magazine, SB Nation. Includes content licensing and co-developed AI products.
- The Atlantic: Licensing plus product collaboration through Atlantic Labs, with citations in output.
- Associated Press (AP): Archive access dating back to 1985.
- Axel Springer: Politico, Business Insider, Bild, Welt.
- Financial Times: Licensing agreement allowing ChatGPT to cite FT content, reportedly valued at US $5–10 million.
- DotDash Meredith: Multi-outlet licensing and collaboration on AI-driven editorial features.
- Condé Nast: Vogue, The New Yorker, Vanity Fair, Bon Appétit, WIRED. Multi-year agreement.
- Guardian Media Group: Attributed summaries and article extracts from The Guardian.
- Axios: Content licensing plus funding for local news expansion.
- Prisa Media, Le Monde, and select French and Spanish publishers: Regional licensing agreements.
Confirmed Google Gemini Content and Device Partnerships
Gemini’s content partner list is shorter but the device distribution agreements give it a structural advantage that no amount of content licensing can replicate for a competitor.
| Partner | What was agreed | Status |
|---|---|---|
| Associated Press (AP) | Real-time news feed integrated into Gemini, Google’s first live news integration into its AI chatbot. | Confirmed |
| Approximately $60 million per year to use Reddit data for AI training, expanding Gemini’s conversational knowledge base. | Confirmed | |
| Stack Overflow | Licensed content powering Gemini’s coding and technical Qu0026A capabilities. | Confirmed |
| Samsung Galaxy S25 | Gemini set as default AI assistant, with Google paying Samsung for exclusive placement across select Android devices. | Confirmed |
| Apple iOS | Reported talks to license Gemini for iOS to enhance Siri and on-device AI features. | Rumoured |
The device placement deals are the detail most SEO commentators overlook. Even if Gemini loses the content quality contest against OpenAI, it will be the default AI assistant on the majority of smartphones globally. Default placement is a distribution moat that content quality alone cannot overcome.
What Post-SEO Actually Means for Your Strategy
Post-SEO does not mean no traffic from search. It means that treating SEO as the primary goal of a digital presence is a structural error. SEO is a distribution channel, and like every distribution channel in history, it has a lifecycle. It grew, matured, and is now being disrupted by AI-mediated search. Channels do not die instantly; they shrink relative to alternatives and eventually become a minority share of a diversified mix.
The sites that will continue to grow are those that treat search visibility as one output of a broader strategy: entity authority, structured data, answer engine optimisation, generative engine optimisation, and genuine topical authority built for machines and humans alike. We explored the full framework in our post-SEO marketing strategy guide.
The practical implication: stop measuring success by keyword rankings alone. Measure citation frequency in AI answers, brand search volume, direct traffic, and conversion from every channel. A site that ranks third for ten queries but appears in AI Overviews for a hundred queries may be winning even as its GSC click data falls.
What Independent Sites Can Still Control
The argument here is not that independent publishers should give up. It is that they should stop competing on the terrain where large publishers have an insurmountable structural advantage and start competing on terrain where they can win.
- Specificity: Large publishers cover everything broadly. Independent sites can cover narrow topics with a depth and speed that editorial processes at scale cannot match.
- Entity building: Establishing clear authorship, credentials, and consistent topical focus helps AI systems identify and trust a source. Our entity-based SEO service addresses this directly.
- Structured data: Schema markup makes content machine-readable at a level that most large publishers apply inconsistently.
- AI answer inclusion: Optimising for inclusion in AI-generated answers is a distinct and learnable discipline. See our AI answer inclusion service for the current approach.
- Topical authority: Concentrated expertise in a defined subject area remains one of the strongest signals an independent site can build. Our topical authority service covers this in detail.