“Commodity” content by Google (an allusion to Marx’s Commodity fetishism)

Why Google Started Calling SEO Content “Commodity” 

What Google meant by making an opposition of commodity vs non-commodity content? 

Commodity content itself is a fresh, new term — it has never been used in relation to the content before, especially in official Google documentation. The part of recent documentation entitled “Apply foundational SEO best practices to generative AI search” suggests: 

1. A distinction between commodity vs non-commodity content itself
2. Where resolution lies, i.e. in application of the common knowledge.

So, unique expert or experience content takes — falling under the notion of non-commodity one — are the way to optimize for the gen AI search.

Why does Google official documentation choose the “commodity” term? 

Well, this could simply be a pragmatic vocabulary choice: "commodity" may just be a convenient shorthand for "generic”. Google's documentation teams iterate terminology constantly. Previously, Google most often referred to terms “helpful”, “reliable” and “people-first” in relation to the content.

“Helpful” is used to emphasize the “problem - solution” idea rooted in content:

• providing the answer to user search intent and 
• being exhaustive, i.e. does not command further search or going back to search results. 

“People-first content” often refers to the basics of SEO, i.e. creating the content for people, not bots. Now, commodity terms seems to have expanded this triad of helpful - people-first and reliable (meaning trustworthy) with something new, i.e. unique insights coming from those who are in position to produce this uniqueness. 

Who Google thinks can produce the non-commodity content? 

In a recent speech by Danny Sullivan, he used the following slide to demonstrate the difference between the two types of content.

I acknowledge that the linguistic markers Danny Sullivan uses in his slide are signals, not definitions, yet, these signals suggest non-commodity content might be: 

• based on personal experience — uses pronouns “I” and “we” — and is 
• process-centred (uses terms like “deep-dive” and “breakdown”). 

So, this type of content can be called maker-specific in the sense that it

1. pertains to those who actively do and report and, also,
2. is deeply rooted in what you do — a specific angle, position or problem that’s connected to the maker’s person. 

This is meant to emphasize a contradiction thereof to the maker-agnostic or generic content:

• pertain to anyone who jump on keyword research tool to discover “a niche”
• is shallow in experience because the maker simply “have not been there” or “have not seen it”: they just produce something to fill in the niche found.  

So, producing non-commodity content that’s beneficial for gen AI optimization takes personal experience and a deep orientation in the process itself — this is why it may be called maker-specific, i.e. only this maker can produce this content.  

To my taste, the very selection of the term “commodity” deserves additional analysis from the standpoint of gen AI search: after all, it is their (gen AI) bet to discourage content makers from commoditizing their work — passing from a product to a commodity.

How exactly does a product become a commodity?

Let’s reverse-engineer a commodity: how it emerges (on the market)? Here I’ll import the “commodity fetishism” term by Carl Marx to explain this. Carl Marx insisted that capital subsists in two forms:

1. use-value or consumer value — where capital is employed as means of production — and as 
2. objects defined as capital, e.g. gold, etc. 

In the manuscript of "Results of the Immediate Process of Production" (c. 1864), an appendix to Capital: Critique of Political Economy, Volume 1 (1867), Marx said that:

In consequence, the product embedded in this mode of production is equated with the commodity, by those who have to deal with it. It is this that forms the foundation for the fetishism of the political economists.

This means gold or money make “emerge” products that are embodiments of capital and that can themselves be used to produce even a greater number of products. So, commodities are, in the first hand, objects of exchange and, secondly, their use-value is a foundation for new products. 

In the case of content the role of capital is to enable emergence of the content using AI, human labour or a mix of it. And these content pieces must then work as building blocks of their user acquisition strategy in organic search. If content is made for this, it has 

• an external aim (to rank higher and attract traffic) and 
• external locus of control (it starts with bots, not humans). 

In other words, commodity content is a transactional strategy of exchanging content units (URLs or chunks) to search traffic at a certain rate. 

Acting on this illusion — as if there’s a real market for such content — is commoditization. If radicalized, this indicates that the market is in a different place (this isn't the real market; the real market is undergoing formation).

Why may gen AI features bet against the commodity content?  

There can potentially exist numerous reasons while Google started betting against commodity content, e.g.:

• AI Overviews degrading in quality, 
• reacting to advertiser pressure, or 
• simply updating vocabulary 

But to me, a strategic bet against commoditization is rooted in the very essence of how gen AI search works, i.e. repackaging the publishers’ content. AI generated search uses Retrieval-augmented generation (RAG) technique that generates an AI overview or response on the basis of the information existent the page’s content. 

Commodity content that’s written on the basis of keywords analysis does not benefit RAG technique because its proliferation decreases the variety of responses, thus, information gain from each incremental URL in the set (they tend to be all the same) — commodity content may reduce marginal informational value for AI retrieval systems.

Excessive commodity content is likely to reduce the marginal informational value of additional documents in retrieval systems.

So, if the aim is to be able to repackage the URLs into AI answers, then commodity content is something that prevents this repackaging: it is not maker- or case-specific and is too synthetic by itself (in the sense that relies on other sources). 

This evokes a term “commoditization” in another sense: doing so by other market agents, including AI search itself, as a sign of shifting markets.

This sign may be interpreted as a permanent process of market shift — a function of a capital over the knowledge (embedded in the content) — that occurs permanently by incentive of different market agents: competition or search engines themselves. 

This new tendency shows that a maker-specific content is protected (at least now) since it carries the hallmark, i.e. specific angle, personal (indissoluble) expertise or a unique experience about a certain process. Simply saying, commoditizing such content requires at least citation — creating another market of search. 

Possibly, the other side of a content being maker-specific is a relatively low volume of traffic from search. If content makers will become very much maker or case specific, they are likely to have a defensible place in AI answers, but will not this position mark the limits of their potential traffic (like in traditional SEO)?

In other words, does defensibility reduce scale? Can we say that conformity to the requirements of gen AI search, including producing non-commodity content may reward:

• specificity,
• experience-led expertise,
• defensibility,

but these same characteristics:

• narrow addressable demand,
• reduce broad keyword coverage,
• constrain scale.