Keyword density is dead. If your AI writer is still counting how many times "best running shoes" appears in a 2,000-word article, it's optimizing for an algorithm that stopped existing in 2019.
Google reads text the way a linguist does — by identifying entities (people, places, concepts, products) and measuring how central each entity is to the overall document. That measurement is called salience.
Cluster Writer uses the Google Cloud Natural Language API to do exactly the same analysis — on your content, before you publish it.
What Entities and Salience Actually Mean
An entity is any distinct concept Google can identify in a text. "React," "Server-Side Rendering," "Vercel" — each is an entity. Google assigns each one a salience score between 0.0 and 1.0, representing how central that entity is to the document's meaning.
An article about "React Server Components" should have "React" and "Server Components" as its highest-salience entities. If Google's NLP reads your article and assigns the highest salience to "JavaScript frameworks" instead, your content is drifting off-topic — even if the right keywords are technically present.
That drift is invisible to keyword counters. It's not invisible to Google.
How Cluster Writer Uses Google Cloud NLP
Step 1: Competitor Entity Mapping
Before writing, the system pulls top-ranking pages for your target keyword and runs them through Google Cloud NLP. This extracts the exact entity profile Google rewards for the topic.
Step 2: Entity-Aligned Writing
The writing engine builds paragraphs around this entity map. Each heading targets specific entities at specific salience levels to ensure structural alignment.
Step 3: Post-Write Salience Validation
The finished text goes through Google Cloud NLP once more. If salience scores fall below benchmark targets, the system flags the text and prompts adjustments before publication.
Why This Matters for Helpful Content
Google's Helpful Content Update penalizes pages with low **information gain** — meaning they add nothing new to the search ecosystem.
Cluster Writer addresses this via a 3-tier entity strategy:
- Tier 1 (Must-have): Core entities present in 80%+ of top competitor results. Without these, your page is disqualified from the topic.
- Tier 2 (Expected): Supporting entities present in 40-79% of results, demonstrating topical depth.
- Tier 3 (Differentiators): Entities present in fewer than 40% of results or absent entirely — generating your structural information gain.
Analyze entity profiles, salience charts, and demoted keywords inside the console.