I was at WordCamp Asia 2026 in Mumbai last month, sitting in on a workshop hosted by Danny Sullivan, Google’s Search Liaison and adviser to the search division. The session covered good practice tips for content creators. Sullivan walked through examples contrasting content that performs well in the current system against content that gets absorbed into AI Overviews and stops driving traffic. The throughline across every example was the same distinction: content only you could have produced versus content anyone could have produced. He called the second category commoditized.

That word landed differently than the rest of the session. Commoditization is an economics term, not an SEO term, and Sullivan using it to describe a content problem pointed at something more structural than a checklist fix. The room took notes and moved on. I kept turning the word over.

The further I got from Mumbai, the more I kept seeing the same pattern in content programs I know well, including my own. Teams producing at volume, covering topics comprehensively, executing briefs cleanly, and still watching organic traffic compress quarter over quarter. The quality is fine. The strategy is the problem.

What Content Commoditization Actually Means

Let’s be precise, because the word travels as a synonym for “bad writing“, and that framing misses the real problem.

Content commoditization happens when the same information, framed the same way, produced through the same process, reaches the same conclusion across thousands of different URLs. It is a sameness problem, structural sameness baked into the production process itself.

A commodity is a product where buyers cannot distinguish one unit from another. Oil. Wheat. Standard steel. You buy it at price, and preference is irrelevant. When your content is a commodity, Google faces a buyer’s dilemma: if everything is interchangeable, it synthesizes across sources and surfaces the consensus directly in AI Overviews. Your page becomes input. A destination becomes raw material.

This is the current architecture. Present tense.

The industry has been building content factories since 2015. Topical clusters, content briefs, keyword coverage matrices, programmatic generation at scale. All of it was logical. All of it worked, until the assumption underneath it stopped being true.

The assumption was that comprehensiveness was differentiation. Cover every keyword in a cluster, own the topic. But when every domain runs the same playbook, comprehensiveness becomes the floor. You are expected to have it, and having it earns you nothing.

What Sullivan Was Actually Saying

To be fair to the volume-first school: content-at-scale was a rational strategy. If you understand keyword opportunity, conversion probability, and competitive gaps, deploying content broadly is sound capital allocation. I have run those programs. The math behind them is real.

But here is what that position misses at the architecture level: Google’s retrieval system has shifted its core question. It moved from “which page covers this topic best” to “which source has something to say that no other source has.”

When Sullivan said “non-commoditized,” he was pointing at a specific signal failure in how current systems evaluate content. Dense retrieval models, the kind powering both Google’s ranking infrastructure and AI Overview sourcing, are trained to find semantic variance, to surface the page that knows something different, frames something distinctly, or carries a perspective that resists synthesis. Entity-level authority. Authorial specificity. Experience signals that survive averaging.

That is the architecture. Non-commoditized content has low substitutability in that system. Commoditized content has high substitutability, and high-substitutability input gets absorbed into a generated response rather than cited as a source.

You become raw material. The destination disappears.

The Factory Problem

The uncomfortable part of the Mumbai workshop was the room.

Most attendees were content strategists, SEO leads, and marketing managers, the exact people responsible for content budgets and production pipelines. And the questions after Sullivan’s session were almost uniformly tactical: how do we optimize our existing posts, should we restructure our briefs, should we be adding more FAQs?

Execution improvements to a flawed strategy.

You cannot optimize your way out of a commoditization problem. Tighter briefs, better meta descriptions, additional schema markup, these are surface-level corrections. If the content itself is interchangeable with the next 40 results, technical SEO refinements leave the underlying problem untouched.

Most content programs were designed for a retrieval system that has materially changed. They were designed for a ranking race, a world where being the most complete, the most crawlable, the most link-backed page on a given topic was a winnable position. That world still partially exists. But a significant and growing slice of search traffic now runs through a system that rewards irreplaceability over comprehensiveness.

And organizations are still running factories calibrated for the old race.

What Non-Commoditized Content Actually Looks Like

Non-commoditized content carries a claim that can only come from you. A position derived from proprietary data. A diagnosis from repeated hands-on experience with a specific problem. An observation pattern built from working inside an industry long enough to notice what outsiders miss. It is the difference between “here is what the research says about X” and “here is what I have watched happen every time a company tries X, and here is where the published research misses the operational reality.”

The form can be anything, a case study with specific numbers, a documented failure and what it revealed, a contrarian argument backed by named evidence, an annotated process walkthrough that shows how the work actually gets done.

Sullivan put a table on the slide. Three industries. Two columns. It was the clearest articulation of the gap I have seen from anyone inside Google.

The pattern across all three rows is identical. Commodity content starts with a category and fills it. Non-commodity content starts with something that happened and documents exactly why it matters.

One column describes what the field generally knows. The other column describes what a specific practitioner learned, in a specific situation, with enough granular detail that a synthesis model cannot average it away.

That is the bar. Sullivan’s slide made clear Google is now calibrating toward it.

The Stakes Are Concrete

If your content program is running the playbook it ran in 2022, keyword research, topical cluster mapping, brief generation, volume production, you are producing commodities at scale.

AI Overviews already absorb a significant percentage of informational queries. The synthesis capability will improve. The share of search traffic that arrives on pages rather than answers will continue to compress. In that compressed environment, the pages that persist are the ones carrying something the synthesis model could not construct on its own.

The field has been debating tactics, structured data, E-E-A-T signals, first-person experience language, while the underlying strategy stays untouched. It is renovation work on the facade of a building with a structural problem.

Danny Sullivan said none of this in those exact terms. He was measured, as he tends to be. But the direction was unmistakable, and the implication was clear to anyone paying attention.

Google is moving toward rewarding irreplaceability. Most content programs are still optimizing for coverage.