Entity-oriented Search Understanding is an important part of Search Engine Understanding or Search Engine Communication. These terms might be new for the traditional understanding of SEO, but the process of understanding a search engine is a daily routine for any SEO to analyze the search engines’ decision trees that create their result pages. Entity-oriented Search Understanding is the understanding of a SERP Instance based on entities, their types, attributes, and connections to each other. A search engine might choose only certain types of pages that include a type of entity along with particular attributes and phrase variations for these attributes with the most related facts. Or, a search engine might filter the results according to the sources’ N-Grams for certain entities. If they don’t have enough numbers, or if they don’t have the relevant facts and external references for these entities, it might be outranked.

With entity-oriented search understanding, an SEO can have a better and deeper vision for the search engines. Knowing an entity from a topic, its contextual layers, and relation types with other entities are crucial to help a source (website) rank higher on the query results of a web search engine. In this article, entity-oriented search engine optimization, and its effect on a Broad Core Algorithm Update’s results, will be explained with a real-world SEO Case Study which is BKMKitap.com.

A section from the Entity-oriented Search of Krisztian Balog.

Before continuing, you can find the related information for the author of this SEO Case Study and examination.

Koray Tuğberk GÜBÜR
Koray Tuğberk GÜBÜR publishes SEO Case Studies, research and experiments, A/B Tests, and guides regularly. From Technical, Semantic, or Local SEO to Data Science, Web Development, he focuses on a variety of verticals. He is the owner of the Holistic SEO.

“By implementing entity-focused SEO perspective with innovation, and following the new SEO Concepts, we were able to regain the lost organic traffic due to the latest Google Broad Core Algorithm Updates. By focusing on Core Algorithm Updates to improve the authority of BKMKitap.com with Koray’s guidance, we have created more granular and detailed content for e-commerce and informative web pages.”

Haluk Bolaban,
BKMKitap.com Marketing Director

Background of the Entity-oriented Search Understanding Case Study BKMkitap

The last 5 broad core algorithm updates of Google were mostly negative until July 3, 2021, Broad Core Algorithm Update.

  • July 3, 2021, Broad Core Algorithm Update (Strongly Positive)
  • June 2, 2021, Broad Core Algorithm Update (Neutral)
  • December 3, 2020, Broad Core Algorithm Update (Harshly Negative)
  • May 4, 2020, Broad Core Algorithm Update (Neutral)
  • January 13, 2020, Broad Core Algorithm Update (Negative)
  • June 2, 2019, Broad Core Algorithm Update (Negative)

BKMkitap.com is one of the biggest book stores in Turkey, and it is one of the rare situations that a website has a strong search demand, lots of navigational queries but still loses all of the broad core algorithm updates despite there being no untrustworthiness or non-expertise signals on the open web.

Organic Search Performance Change Graphics
An All-Time Graphic of BKMKitap.com from Ahrefs.

However, BKMKitap.com has a tremendous amount of Technical SEO and Web Page Loading Performance Optimization problems. Below, you can see a general overview of the Google Search Mobile Pagespeed report for the BKMKitap.com.

Mobile Pagespeed Report
A summary of the web page loading performance time of the BKMKitap.com.

I try to keep only the “valid” section high in the Google Search Console Coverage Report. This is so search engines can focus on only indexing and ranking rather than understanding which URL is necessary on the SERP with which version; this includes how to parse, digest HTML, or understand the content, etc. Below, you can see the heavy canonicalization errors along with the other mixed indexing signals.

Google Search Console Coverage Report
Technical SEO Problems are not an obstacle for Semantic SEO success, despite it might make a site’s usability for the web harder.

Below, you can see the good URLs of BKMKitap.com in the Page Experience Report of the GSC, if you are lucky enough to see them one day.

Page Experience Report
From millions of URLs, even a single URL doesn’t have success for the Page Experience Algorithm.

To demonstrate the Technical SEO and pagespeed related problems of the BKMKitap.com SEO Case Project, you can check the list below.

  1. Half of the website URLs don’t exist in the sitemaps
  2. There are millions of cannibalized URLs.
  3. There are thousands of duplicate product URLs.
  4. More than 30000 internal 404 pages. (From full data)
  5. Blocked URLs within the Sitemap.
  6. Hundreds of 5XX errors daily
  7. Submitted URLs with Noindex
  8. Redirection Errors
  9. Submitted but 404 URLs
  10. Indexed but blocked URLs (Tens of thousands)
  11. Indexed contents without actual content
  12. More than half a million robots.txt excluded pages.
  13. Nearly 100.000 URLs are crawled and not indexed.
  14. Nearly 53.224 pages are currently discovered and not crawled.
  15. Over 600.000 duplicate with canonical and submitted URL is not selected as canonical.
  16. The site has millions of URLs, but even a single URL doesn’t pass the Core Web Vitals.
  17. Most of the website has poor scores from the PSI.
  18. Thousands of AMP Related issues such as referenced AMP URL is not an AMP, or custom javascript, etc.
  19. Thousands of structured data errors, and missing information for the related products.
  20. Hundreds of thousands of products without stock information, or stock existence. These last two subjects also affect the search engine’s confidence to rank the specified e-commerce page, since the stock information, or brand, reviews, prices are not clear enough for the evaluation algorithms.
Coverage Report of Google Search Console
A colorful Technical SEO look for BKMKitap.com SEO Case Study.

During the BKMKitap.com SEO Case Study, I also optimized lots of technical SEO and pagespeed related tasks. However, most of the problems are caused by BKMKitap’s external developer company TSoft as they don’t have enough expertise, experience, or understanding of the SEO to finish these tasks. When the company grows, the problems and technical necessities, and task handling obstacles grow too. That’s why having a Holistic SEO perspective is important. Since the technical SEO wasn’t an option to change the status, I had to use the Entity-oriented Search approach.

SEMRush Organic Search Performance Graphic
General graphic of the BKMKitpa.com on SEMRush.

Topical Authority is the relevance of a source to a topic by proposing link value to the Search Engines by satisfying the users via the queries that seek answers for certain entities with a certain context. Topical Authority is a fundamental term for Semantic SEO to describe the value of the Natural Language Understanding of the search engines.

Entity-oriented SEO Case Study Stats
General Change Graphics of the BKMkitap.com from Ahrefs.

How to Improve Topical Authority with Entity-oriented Search Understanding?

To improve the topical Authority, a source should cover all the related details to a topic with a certain context, query, and intent template by satisfying the related and possible search intents. To improve the topical authority, the “keyword gap” is not as important as the “information gap”. The facts which exist within a source for a topic, their accuracy, and clarity, position are important to increase the contextual signals of a web page.

To improve the topical authority of an SEO Project, in the context of entity-oriented search, the methods below can be used.

  1. Compare the entities within different web pages.
  2. Compare the context and content angle for these entities.
  3. Compare the facts, prepositions, and semantic role labels for these entities.
  4. Compare the questions on the competing web pages.
  5. Compare the Site-wide and Page-level N-Grams of the web pages.
  6. Compare the web page layout of the web pages (web page design can affect the meaning and context of the entities within the web page)
  7. Compare the anchor texts from outgoing, and incoming links for these web pages.
  8. Take all of the attributes of the specific entity, and give them an order based on the relatedness of the attribute for the source, and popularity of the attribute to generate better questions.
  9. Use a clear sentence structure for all of the prepositions.
  10. Do not dilute the context of the web page with irrelevant opinions, or analogies, and other types of entities.
  11. Process the same entity or same entities from the same type with the same context from start to end.
Google Search Console Change Data for Entity Oriented Search
Change graphic only for the biography of the authors with “who” questions.

If you want to learn more about the contextual search, you can read an SEO Case Study that uses question and answer generation for certain topics. Besides these methods, there are many more things that can be used to improve topical authority, contextual relevance, click satisfaction possibility. I will be creating a course to teach Semantic SEO and Semantic Search Engine’s nature to detail all possible steps together.

How to Use Topical Authority for E-commerce Sites with Entity-oriented Search Understanding?

To use the topical authority for e-commerce sites, a source should cover product, brand, services, and dimensions, attributes of these with related information. This information includes price, color, size, availability, shipping, or refund policy, usage guideline, and product-related questions, reviews, or comparisons. Definitional, informational, comparison-based, opinion-based, review-based, factual, and commercial content should exist at the same time within an e-commerce site to make a source topically authoritative for certain types of entities with different contextual layers.

InLinkes Entity Oriented Search
InLinks can help to understand the main topic and entity within a content.

Informational content and commercial content support each other to satisfy every possible, and relevant search activity from the same source. A source should process the related attributes by answering the questions that can be generated from these attributes. If an e-commerce site is about “electronic bikes”, the source should cover “mountain electronic bikes”, “fat-tire electronic bikes”, their parts, invention, maintenance along with similar “bike types”, or its “alternatives”, and “electronic bike brands, products, facts, tips, usage, obstacles, advantages” and more. All these knowledge domains will include different questions with different word distribution possibilities. Moreover, the possible search intents, correlated queries, sequential queries, entity-seeking queries, and query themes should be used to cover these knowledge domains with the best possible contextual vector.

Semantic SEO for E-commerce
The last three months change the graphic of the BKMKitap.com

If you don’t know what a “query theme” and “query template” are, you should read the “Indexing SEO Case Study” that I have written to conceptualize the cost of document for indexing along with the cost of not indexing the document. With simple steps, the usage methods of topical authority for e-commerce sites are described with the list below.

  • Understand the dimensions of the product that you sell.
  • Find all of the relevant entities for the product, including its brand, material, inventor, alternatives, similars.
  • Generate the best proper questions for these dimensions of the product, brands, related entities, and their attributes.
  • Give the questions a proper order based on the web page layout, and web page purpose.
  • Match the query and answer format with NLP convenient sentence structures.
  • Use information redundancy, and unique value opportunities for the products.
  • Connect all the entities based on their ontology for commercial purposes.
  • Understand the popularity of entity attributes and relatedness of entity attributes.
  • Try to use entity relations, relation types, semantic role labeling, entity resolution from the eyes of the search engines.
  • Use phrase templates, phrase pattern taxonomies, and create a prominence hierarchy without diluting the context.
  • Search Engines’ perspective for a topic and central context of the topical map should align with each other to make search engines understand the website easier, and faster.
InLinks Entity Recognition for Internal Links
Performing a NER for a webpage, and using them for potential internal and contextual links is possible.

How to Understand Which Entity Attributes are More Important for a Context?

The prominence of the attribute, relatedness of an attribute, and popularity of an attribute have different importance levels. An entity attribute can be popular, but it might not be prominent. It can be related to the specific contextual domain, but it might not be prominent enough. And, entity attribute popularity can be seen in two different ways, one is the relatedness of the attribute with different synonyms and query patterns, the other one is the total search demand for the specific attribute. To improve the contextual relevance of a document to a certain query template, a source might need to process lots of different entities from the same type with the same attributes, questions, and answer formats.

Entity Oriented Search Case Study
Non-brand Queries, and All of the queries for change over the last 3 months, and last 28 days.

To understand the attribute’s importance for an entity, the entity-oriented search analyst should focus on the source’s context, purpose, and the common attributes of the entities from the same type. For instance, if the source is about Formula One, the important attribute of the car will be the “driver, constructor, engine, top speed, weight”, if the source is about history, the main attribute will be the “inception of cars”, or “inventors”. According to the source, the most common attributes from the same entity type will be most important.

To find the attributes that matter for an entity in order to generate questions, a search analyst can focus on the relatedness of the attribute, and prominence of the attribute. For instance, if the source’s knowledge domain is Formula One, the car’s driver, and race circuits will be more prominent than the lap count, or circuit’s viewer capacity. Some attributes have better popularity, and these search-demand waves or trend changes can protect the ranking of the document during newsworthy events. This can improve the news-focused documents from the same source.

InLinks Entity Oriented Content Writing
Via InLinks, having a content brief creator, and entity-oriented keyword clusterer is possible.

How to Connect Entities to Each Other to Strenghten the Contextual Signals and Relevance?

Creating entity connections based on context is possible by using the perspective of ontology. Every entity will have a mutual part with another entity to create a triple. These triples (one object, two subjects) can be used to form a knowledge graph. Information Graphs that include facts can signal the factuality of the content by improving the relevance of the content for a specific query or need behind the query. In this context, co-occurrent phrases, and co-occurrent entities can shift the context of the document. To connect the entities to each other based on a context, Semantic Annotations should be understood as a concept from Named Entity Recognition. Semantic Annotations can be used to label a document for a specific context. The labeled document for a context can signal a weighted attribute of an entity that is a subject of another object. And, the same object can be a subject for another entity, in other words, an attribute.

These subject-object or entity-attribute switches will change semantic annotation of the text span, and these semantic annotation changes can be used as “internal links” with a definitive relevance.

Entity oriented search query earning
Since every query and entity within it is connected to another one based on facts, and context, the new query count is way much higher than main competitors.

To give a little bit further concrete example to deepen the process of entity connections, you will find 5 different named entities below.

  • Germany, type country.
  • France, type country.
  • England, type country.
  • Turkey, type country.
  • The United States, type country.

All these 5 entities are from the same type which is country, thus they will have the same main attributes. Attribute hierarchy or semantic dependency tree of the attribute can be used to understand the priority of the attribute to define an entity. If a search engine sees these 5 attributes on a web page with the attributes “currency, bank, finance”, it will understand that the topic is international finance. From this contextual layer, possible queries, search intents, and related search activities, indexed documents, questions will be retrieved from the storage of the search engine to satisfy the user. If the terms from the page include “education, school, classroom”, it will understand that the topic is the educational situation, and programs from these countries. An extra phrase or related attribute here can shift the context, and if context shifts, the attributes will be shifted too. To create the entity connections, these entities should be used with the mutual attributes for a specific context, so that the page can experience ranking signal consolidation.

Google Entity Storage
A Ranking Signal and a Ranking Factor are not the same things. A signal can be used to understand a factor, or multiple signals can be used differently to form a new factor. The schema above is to understand the “sentiment for reviewable entities”. Google can store reviews, entities, review owners to understand the general sentiment for a reviewable entity which is a product, or service. Thus, use everyone’s facts, not your opinions for an entity-oriented search engine optimization project.

Germany can be connected to Turkey based on its currency exchange rate, or its populational mutual features. All these entities have millions of different possible connections and connection variations between each other. At the same time, Turkey can be connected to England for external debts, while the United States can be connected to Germany for the dollar index. All these different connections, connection permutations, and relation types will define the page’s relevance, and factuality to the user’s query, and represented queries.

How to Define Entities by Specifying the Context?

An entity can’t change its definition based on languages, but its attributes’ prominence can. Or, an entity’s prominence can change based on context. The general information about an entity can switch, the sentiment can be contrary, and the definition of the entity might alter its shape. A tree as an unnamed entity can be a plant in the context of city planning, or biology. For these two contexts, a tree is completely different. A tree can be a decor for home-balancing, or it can be a figure from a painter’s interpretation. A tree can be a mythological creature or a material for bridges. From architecture to ship construction, or from biology to paper prices, an entity can switch its definition, based on context.

Google Entity Question Answering
Question answering is an important part of the entity-oriented search. Google can choose multiple possible answers for a specific question. A contextual domain can be determined by a single qualifier, such as a year, place, or demographic group. In this context, the search engine can choose the best web page with the best answer coverage.

To improve the entities’ precision, and factual information redundancy of the source, the entities should be defined with their functions, importance, usage, benefits, and effects for the specific knowledge domain. If its difference, unique and similar sides, alternatives, and advantages are absent within the web page, or if they are not being able to select easily, the web page might dilute its context, relevance, and informational value for the search engine’s re-ranking, and initial ranking algorithms.

How I Used Entity-Oriented Search Understanding for BKMKitap.com SEO Case Study?

Since there is no way to fix the technical SEO, web page layout, web page loading performance, and user experience-related issues on BKMKitap.com as a result of TSoft issues, I had to use entity-oriented search and optimization. To use entity-oriented search understanding for BKMKitap.com, I have focused on the source’s context which is “book e-commerce”. To cover the informational, and e-commerce related contextual domains from the same knowledge domain, I need to find the most related attributes for both of the angles.

During the SEO Case Study, to connect the e-commerce, and informational, definitional contextual domains based on books, I have used the “ontology” and “taxonomy” of these things. When it comes to books, as a product, it has “size, material, author, ISBN Number, price, page count, an image or visual for cover, editor”, when it comes to booking as a literature value, it has an “effect on the subject that it processes, a topic, unique sides, differences, authors, characters, genre, era, style, school and more.” When an SEO understands these two sides of the entity as a product, and artwork, the next step is search intent understanding.

Entity Query Association
Query and Entity Association can be used to change the documents on the SERP or re-ranking the query results by understanding the relationship between the named entities. A query can be connected to the multiple entities or mutual attributes of different entities. These query templates can be used to cover a topic with high categorical quality.

In the context of search intent understanding which is an entirely different topic, an SEO should know that a web page should have a dominant context. In other words, a web page can’t be an e-commerce web page and an informational web page at the same time at the same level. One of these options dominates the other one for the specific web page, and the anchor texts or the web page layout should align with this option. Thus, in the BKMKitap.com SEO Case Study, I have created two different web page types, one is for the books’ e-commerce side, the other one is for books’ literature value along with their authors. An e-commerce web page can have an informational content piece, but if this content is about “buying the book”, and “using the product” along with “refund and delivery policies and conditions”, it would improve the search intent coverage for the related web page.

How I Improved the Contextual Relevance and Topical Authority with Informational Content for Commercial Intents?

To improve the contextual relevance, and topical authority with informational content for commercial intents, an SEO should cover the informational, definitional, and factual hinterland of the topics for the specified products. In this context, I have created a different web page group for the things (entities below).

  1. Books Genres
  2. Books from Different Geographies
  3. Books for Different Geographies
  4. Authors from Eras
  5. Authors from Geographies
  6. Authors from Cultures
  7. Authors from Ideologies
  8. Individual Author Biographies
  9. Author and Book Connections
  10. Author’s Similarities, Differences, Thoughts, Childhood, and more.

The other important section is question generation, creating a context sharpening entity-oriented search document, and connecting all the relevant facts to each other. 

Context Sensitive Ranking
A section that I like from Google’s own “Context-Sensitive Ranking” patent. A word or preposition can have a different weight for relevance based on context for reranking purposes. An entity, an attribute, a fact can differ the rankings based on context, and content’s contextual optimization.

Last Thoughts on Importance of Entity-oriented Search Understanding for SEO

Entity-oriented search understanding is not popular as AI Text Generation. But, whether it is Natural Language Understanding or generating text, the entities, and entity-oriented search is the center and heart of these processes. The person who understands the entities, their nature, and their possible connections, search engines’ perspectives for these entities, possible functions, actions, and definitions for these entities will create the difference between conventional SEO and Holistic SEO. To implement the entity-oriented search principles, an SEO should perform experiments with the entities, and their attributes by creating hyper-structured data. To make the information extraction easier, different sentence structures for different questions can be used to implement entity-oriented SEO A/B tests.

SHARE:

Koray Tuğberk GÜBÜR is the founder and owner of the Holistic SEO & Digital. Koray Tuğberk publishes SEO Case Studies, Researches, and Detailed A/B Tests along with their results. Koray regularly attends webinars, conferences, and SEO-related events.

Comments (4)

  1. Benu Aggarwal

    Reply

    Great coverage. Love to show you what we did using Innlink, SEM rush and our own product Schema Manager. Let’s connect. I will ask Dixon to introduce us.

  2. Reply

    Great article Koray! I love the idea of using the pages the client already had in place and just bridged them together with all the pages by “book genres”, “books from different countries”, etc. Seems like a very effective way to improve topical authority and make connections between all those entities.

Leave a comment

Your email address will not be published. Required fields are marked *