Google has been working on finding entities since it introduced the Google Knowledge Panel. Unlike a semantic index or adding synonyms, entity-based search needs to scale and adapt to new or shifting concepts. Almost every noun is an entity, and the most popular entities are capitalized in the titles and headlines of web pages.
Entities can be found in every sentence, in every paragraph, and in every title. But the ones that are relevant for SEO are the ones people are searching for and the related entities to the entity/topic.
The competitor's entities can be found using AI
Google finds entities in public databases
Bert finds entities in search queries
Entities are found when searching for the breadth of topics
While Google was developing its entity dataset, people doing SEO noticed that semantically related terms were affecting their SEO even though these terms did not have the keyword density to make the page relevant. Google stated they did not use a semantic index, and Bert was announced. Google has not explained the details of what related words it considers as important, but they did leak clues.
"Google’s Knowledge Graph isn’t just rooted in public sources such as Freebase, Wikipedia and the CIA World Factbook. It’s also augmented at a much larger scale—because we’re focused on comprehensive breadth and depth," -- Admit Singhal."
Google essentially crowd-sources entity knowledge from the internet at large. Using certain sites as public data and authorities to maintain the integrity of the dataset.
Trusted Sources Google finds entities at include sites like
wikipedia.org and wikidata.org
linkedin.com and imdb.com for people
annualreports.com for companies
onetonline.org for professions
Google My Business for local businesses
Trusted directories for local businesses
Websites are considered an authority about themselves and schema is used
Trending topics are also used to create knowledge panels not otherwise discovered.
Knowledge Panel Information can be queried Query Knowledge Panel use the "try it" option or developers can use an API for larger volumes.
Related entities are the entities or questions people ask in regards to their entity of interest. People add these to their search queries. They place these in their reviews about businesses and products. User generated forums such as reddit maintain that the comments need to stay on topic; Or, otherwise be related entities.
Google's released information on Bert, illiterates how they use natural language to identify the topical entity.
Bert (Natural Language Processing) - User Intent
Google's Bert processes the user query for entities to understand the user intent. Algorithmically, it selects the dataset index for the results. Matching entities with user intent is critical for entity-based SEO.
Here’s a search for “2019 brazil traveler to usa need a visa.” The word “to” and its relationship to the other words in the query are particularly important to understanding the meaning. It’s about a Brazilian traveling to the U.S., and not the other way around. Previously, our algorithms wouldn't understand the importance of this connection, and we returned results about U.S. citizens traveling to Brazil. With BERT, Search is able to grasp this nuance and know that the very common word “to” actually matters a lot here, and we can provide a much more relevant result for this query." -- Pandu Nayak
Entities, while they are nouns, people, places, and things, are also the subjects of sentences. When scanning a page for entities, we are looking at the subject of the sentence and the topic of a paragraph.
When Google user intent is for a breadth of information, Google will present related entities. The entities may not be the best ... rather AI is testing for what entities are the most popular. Adding factors, types of, practices, etc to the query will signal intent. For example: "factors for painting a building."
Looking at the source code on Google's search engine results page shows the Google Knowledge Panel Entity ID for these related entities.
Entity Drift - Entity Deserves Freshness
Over time, related entities will change as new topics emerge. Google's system updates or rotates entities to match those that might interest users.
Gain of Knowledge and Diversity of Interests
You can’t please everyone—there’s no single set of related entities for any topic. People prefer different pages based on their interests. Google’s Navboost algorithm selects the most popular topics, but diversity is maintained because people engage with content matching their intent and preferences.
Entity-based search has the potential to personalize results by focusing on fields or concepts that certain people may find more interesting. While there is no evidence that entities are currently being used for personalized search, this is an evolving technology.
Competitive SEO
If two pages match the same entities, the original page is usually ranked higher, and the second page doesn't add value to the search results.
To rank for competitive terms, you need to use entities that your competitors aren't using, which is called gain-of-knowledge. The content can either explore a topic in depth or cover a broader range of aspects related to it.
