Horizontal Topical Measurement
My Helpful Content analysis theory began with two years of indexation testing data and it revealed there was some horizontal topical measurement between the pages on the domain. The more topical variance, the less the indexation rate. The less topical variance, the greater the indexation rate.
And not by a little, indexation rates increased on primary smartphone sites by a whopping 113% after introducing less topical variance.
Did that mean there can only be one topical focus per domain? The answer depends on whether you approach it through a horizontal analysis of your topical nucleus—measuring breadth across all pages—or a vertical one focused on depth within a single subject.
Do Some Sites Play by Different Rules?
I began to wonder if I could find any site that was the opposite of a topically focused site. Was there any site that because of what it was lacked a topical center? What might that look like?
News sites by definition are sites that lack a topical center.
What Makes a News Site Different From Niche Sites?
A news site is different from other sites primarily by focusing on timely, fact‑based reporting of current events under editorial standards and a regular publishing cadence, but without a single topical center—even its “topic” isn’t news as a subject, it is simply news itself. This structural distinction also surfaces in how search systems read and classify content at the signal level—news sites benefit from semantic HTML signals beyond topical classification that operate independently of any topical nucleus requirement. That structural distinction has real implications. Sites without a topical nucleus do not have the same requirements as traditional niche sites do to semantic link architecture for topical coherence signals.
Once HC was rolled out fully in 2023, two things were happening simultaneously. On one side, the Helpful Content System (HCS) was decimating niche, affiliates, and blogging sites. And on the other, news sites appeared to be operating in an entirely different reality. That divergence is even more striking today, given why AI search and HCU now share the same foundation—the HC logic is the same logic AI-driven search is inheriting.
What made them different if not topic or rather the LACK of a central topic. The answer lies in how Google measures topical coherence across a domain—and why news sites never needed the semantic relationship links that counter topical variance that niche sites depend on for classification stability.
Helpful Content Is NOT Search Quality Rater Guidelines
Content quality defined by how comprehensive the content was written, is not the helpful content that the Helpful Content System is looking for.
Within the data, I could not find any to suggest any factor concerning the quality of the content pages themselves resulted in higher indexation rates.
No evidence found in indexation rate data that rates increased due to the quality of the content increasing.
Prior to setting a topical center on the test sites, I had moved from random alpha (e.g. Mozzwgdkqmaqzi wabhxuhcq fuxdvtvmnfyyw tmzvgchm qxeesuideeqkx) to standard english sentences and grammer with zero shift in indexation rates afterwards. This mirrors a broader pattern in content strategy: why unique signal beats average content at scale – generic improvement without structural differentiation does not move the needle.
However, each domain was comprised of every page’s topic being unrelated to any previous page(s). This is precisely the condition that modern AI evaluation frameworks are designed to detect—what signal engineers now formalize as topical variance and cross-entropy validation, the same measurement logic that surfaces when a domain’s pages share no coherent semantic thread.
It was only after the site was converted to pages of related subtopics organized around a single large topic that the response of Google Systems altered. The strength of the relationship between the entities used in topical shift in content appeared to create a new factor for consideration that led to successful indexation.
The News Site Exemption: A Structural Privilege in the SERPs
This News Site Exemption is a structural privilege that has fundamentally reshaped who survived in the SERPs regardless of topical coherence. Understanding why this exemption exists requires stepping back to consider how controlling your Google classification signal works at the architectural level—and why news sites were never subject to the same classifier in the first place.
During the March 2024 Core update, Olesia Korobka shared some of her notes from a streamed event which featured a conversation between Glenn Gabe and Barry Schwartz about the Core Update.
News sites were not hit with HCU because if you are a legitimate news website, you won’t have that much unhelpful content, not a majority of all your content compared to small niche websites.
I agree that news sites were NOT hit with HC updates. That immunity, however, creates an opening that bad actors are quick to exploit, understanding how fake news sites exploit induction-era trust signals reveals exactly why this loophole is so difficult for Google to close.
But the “get out of jail free card” is not because they have a better ratio of helpful content to unhelpful content. They get one because they are EXEMPTED from the HCS Classifier and its HC rules. Understanding how HCS domain-level scoring works makes clear why that exemption can’t simply be replicated by fixing individual pages.
Otherwise, why would the Site Reputation Abuse Penalty require Google to apply it manually? Helpful Content is algorithmically applied.
If HC was applied algorithmically to news sites, because of the horizontal content classification, that application would most likely have destroyed online news.
The Indexation Gap: 97% Coverage vs. Quality Limbo
Recently a study examining pages not getting and staying indexed by Google by Indexing Insight reinforces the alternate reality for news sites vs everyone else.
“Google is actively removing pages from its search results.”
“…they are actively being removed from search results because of page quality”
Marketplace and listing sites often struggle with index coverage as low as 70%, and even high-authority ecommerce sites rarely break 90%, news websites consistently maintain a 97% index coverage score.
What Is The Driving Force In Losing Pages From The Index?
The distinction is not merely technical; it is systemic. Understanding topical coherence and domain topology helps clarify why this systemic pattern emerges across site types.
According to the Indexing Insight study, most niche sites that lost 10%–30% of pages are not lost to “crawl budget” or technical errors. They are victims of quality-based removal. They say that Google is actively “forgetting” these pages because they fail a relevance or quality test.
But I think its explained with a more definitive flaw than a vague description about “quality issue”.
The “quality test” they failed was the horizontal semantic analysis in comparison to the other pages on their site and were deemed “un-helpful content.”
Why News Sites Escape Topical De-Indexing
News sites, according to what we can see as well as the data, again appear to be exempt from this quality-driven or should I say topical-driven de-indexing.
When a news site faces an indexation issue, it is almost always technical. But when a niche site faces it, it is a verdict on their worthiness to be served.
Google Says They’re Only Looking for Unhelpful Content
Advice given since the arrival of HCS has said all we have to do is put more HELPFUL content on our sites. But in reality the only thing that we know is Google only looks for unhelpful content. They say so in their own Google Developer guidelines.
Our systems automatically identify content that seems to have little value, low-added value or is otherwise not particularly helpful to those doing searches.
Unhelpful Content Is Determined by Topical Distance Math
The Topical Nucleus Trap: Niche Sites and the “Death of the Edge
The core mechanism of the HCS is the measurement of a domain’s siteFocus. For commercial and specialized sites, Google calculates the semantic distance between every page on a domain to form a site-wide radius. The semantic analysis behind siteFocus scoring is what makes this radius calculable—mapping entity relationships across every page to produce a measurable topical boundary.
The Commercial Death Sentence
If a niche site (non-news) “dances on the edge” of its topic—for example, a roofing site adding content on general home insurance or broad DIY repair—it dilutes its topical nucleus. Without topical bridges to connect these ideas mathematically, the HCS classifier appears to flag the isolated topics as “unhelpful,” creating the potential for suppressing the entire site’s serving within the index, even if the individual pages in question are well-written and well-optimized on the page-level.
And its not just commercial sites, there are lifestyle bloggers who speculated on the future implications when the Helpful Content System was introduced as well. Elizabeth Tai, an essayist, sci-fi writer and digital gardener, said,
“What I’m most concerned about this update is how it will impact people who writes a diverse range of topics on their websites. Nicheless websites.“
This “nicheless” fear is in hindsight exactly what the Indexing Insight study’s confirmation of a 70% coverage score on non-news sites looks like in practice.
The News Immunity
News sites cover everything from local politics to global finance to celebrity gossip. By definition, they lack a singular topical nucleus. Instead of being penalized for this “all over the place topical” nature that other sites proudly pursued, news sites are judged differently outside of the range of reach of the Helpful Content classifier. For sites that don’t carry that exemption, the strategic answer lies in expertise architecture for topically focused sites—the structural approach that replaces scattered content with a coherent topical nucleus the classifier can recognize.
For sites that are subject to the HCS classifier, the next strategic layer is building topical authority beyond the HCS classifier—where AI’s autonomous discovery cycle begins to reward the sites that news exemptions never needed to earn.
Frequently Asked Questions
Why weren't news sites affected by Google's Helpful Content Update?
News sites are EXEMPTED from the HCS Classifier and its HC rules. If HC was applied algorithmically to news sites, because of the horizontal content classification, that application would most likely have destroyed online news.
How does Google determine that content is 'unhelpful' under the Helpful Content System?
The core mechanism of the HCS is the measurement of a domain's siteFocus. For commercial and specialized sites, Google calculates the semantic distance between every page on a domain to form a site-wide radius. The 'quality test' they failed was the horizontal semantic analysis in comparison to the other pages on their site and were deemed 'un-helpful content.'
What happens to a niche site's indexation when it publishes content outside its core topic?
If a niche site (non-news) 'dances on the edge' of its topic—for example, a roofing site adding content on general home insurance or broad DIY repair—it dilutes its topical nucleus. Without topical bridges to connect these ideas mathematically, the HCS classifier appears to flag the isolated topics as 'unhelpful,' creating the potential for suppressing the entire site's serving within the index, even if the individual pages in question are well-written and well-optimized on the page-level.
Does improving content quality help a site recover from Helpful Content System penalties?
No evidence found in indexation rate data that rates increased due to the quality of the content increasing. Prior to setting a topical center on the test sites, I had moved from random alpha to standard english sentences and grammar with zero shift in indexation rates afterwards. It was only after the site was converted to pages of related subtopics organized around a single large topic that the response of Google Systems altered.
What is the difference in indexation rates between news sites and niche or ecommerce sites?
Marketplace and listing sites often struggle with index coverage as low as 70%, and even high-authority ecommerce sites rarely break 90%, news websites consistently maintain a 97% index coverage score.