--- title: "Optimize for entities and topics so Google's language systems understand what your page covers" description: "Learn how BERT, MUM, and neural matching read meaning, then map entities, build topic clusters, and write pages that cover a subject in full." canonical: "https://scalewithsearch.com/articles/semantic-seo-nlp" date: "2026-03-20" modified: "2026-09-25" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # Optimize for entities and topics so Google's language systems understand what your page covers. An SEO specialist still follows a density rule: "use the keyword 5 times per 1,000 words." Google left that model behind a decade ago. Since RankBrain in 2015 and BERT in 2019, Google's ranking systems evaluate what content means, not only which words it contains. Semantic SEO optimizes for concepts, entities, and relationships instead of exact-match keywords. A page that targets "best running shoes" does not need that phrase 15 times. It needs the related entities (Nike, Adidas, trail running, marathon training, cushioning, pronation). It needs answers to related questions, such as how to choose running shoes and what makes a good pair. It also needs enough context to show real authority on the topic. The natural language processing (NLP) models behind Google search understand synonyms, context, entity relationships, and intent. They judge whether a page covers the meaning around a query, not whether it repeats the query word for word. That changes the work. Entity maps replace keyword lists. Topic clusters replace isolated pages. Contextual relevance replaces mechanical keyword insertion. This guide covers how Google's language systems work, entities and the Knowledge Graph, and semantic keyword research. It then covers on-page entity work, topical depth, natural writing, and how to measure the result. ## Learn how Google's language systems read content ### BERT Google launched BERT (Bidirectional Encoder Representations from Transformers) in search in October 2019. BERT reads each word in the context of all the words around it, in both directions, not only left to right. That helps with queries where word order and small words such as "to" change the meaning. Google's own launch example was the query "2019 brazil traveler to usa need a visa". Before BERT, results focused on "brazil", "usa", and "visa", and could return pages about US citizens who travel to Brazil. With BERT, Google read "to usa" as a Brazilian traveler who goes to the US. It returned the correct visa pages. The lesson for content: answer the real question precisely. A page that mentions the keywords but drifts around the need rarely ranks. ### MUM Google announced MUM (Multitask Unified Model) in May 2021. At the announcement, Google described it as 1,000 times more powerful than BERT, trained across 75 languages, and able to work with text and images. It can compare concepts, carry knowledge across languages, and handle questions that need several steps of reasoning. Google has said it applies MUM to specific search tasks rather than as a general ranking system. Google's example question: "I've hiked Mt. Adams and now want to hike Mt. Fuji next fall. What should I do differently to prepare?" MUM can connect the two mountains, the season, and the unstated question about differences in gear, weather, and permits. The lesson: a page that addresses the several facets of a question and its related concepts serves this kind of query better than a narrow keyword page. ### Neural matching Neural matching, which Google introduced in 2018, connects queries to concepts even when the query's words do not appear on the page. Google's example is the query "why does my TV look strange". Neural matching links the vague description to the technical concept called the "soap opera effect". It returns pages about motion interpolation settings, even if they never use the word "strange". The lesson: explain concepts fully in natural language. Do not force exact phrases into the text. ## Understand entities and the Knowledge Graph An entity is a distinct real-world thing that Google recognizes: a person, place, organization, product, or concept. Google's Knowledge Graph maps the relationships between entities. | Entity type | Examples | |---|---| | People | Serena Williams, Marie Curie | | Places | Paris, Mount Everest | | Organizations | Google, Harvard University | | Concepts | machine learning, photosynthesis | | Products | iPhone 15, Tesla Model 3 | Relationships connect them. Tesla manufactures electric vehicles. Electric vehicles relate to battery technology, charging infrastructure, and climate policy. When a person searches "Tesla CEO", Google does not only match the two words. It follows five steps: 1. It identifies "Tesla" as an entity, a company. 2. It identifies "CEO" as a relationship type. 3. It asks the Knowledge Graph who holds that role at Tesla. 4. It returns the person entity. 5. It shows that person's knowledge panel, not only pages that contain "Tesla CEO". A page that identifies, defines, and places its entities in context helps Google. Google can then read the page by meaning, not only by words. Clear mentions, supporting detail, and Schema.org markup all add to that signal. The [entity SEO and Knowledge Graph guide](/articles/entity-seo-knowledge-graph) goes further into how to get your own brand recognized as an entity. ## Research topics, not single keywords ### Map the whole cluster The old approach targets the keyword "running shoes". The semantic approach maps the topic around it: | Subtopic | Coverage | |---|---| | Types | trail, road, racing, minimalist | | Brands | Nike, Adidas, Brooks, Hoka | | Technologies | cushioning, pronation support, heel-to-toe drop | | Use cases | marathon training, casual jogging, sprinting | | Selection criteria | fit, arch support, terrain | | Maintenance | cleaning, when to replace | Then build a pillar page, "Running shoes: complete guide", and a supporting page for each subtopic. Link each supporting page to the pillar, and the pillar to each supporting page. ### Find the entities and related terms Several tools extract entities and related terms from the pages that already rank: - MarketMuse analyzes top-ranking content for a topic and lists the entities, subtopics, and questions it covers, with the gaps in yours. - Clearscope and Surfer SEO score a draft for semantic relevance and suggest related terms and concepts. - AlsoAsked collects People Also Ask questions and shows how they connect. - The Google Cloud Natural Language API returns the entities and sentiment it finds in a text. It shows how one Google model reads your entities. It is not the search ranking system, so treat its output as a check, not a verdict. You can also do the analysis by hand: 1. Search the target keyword. 2. Review the top 10 results. 3. List the entities, subtopics, and concepts they share. 4. Note their depth and structure. 5. Build content that covers the shared elements and adds something they lack. ### Cover the variations naturally Three kinds of variation matter: - Synonyms and related terms. For "automobile", also use "car", "vehicle", "auto", "sedan", and "SUV". - Co-occurring entities. For "electric vehicles", also cover Tesla, Rivian, charging infrastructure, battery range, regenerative braking, and tax credits. - Question clusters. For "keto diet", also answer "What is the keto diet?", "How does keto work?", "What foods are keto?", "Is keto safe?", and "Keto vs paleo?". Use these where they fit the text. Do not force exact-match repetition. You still need the primary keyword: put it in the title, the H1, and a few places in the body. Then give most of your effort to the related concepts. If the top 10 results all mention the same 15 to 20 brands or concepts, your page needs to cover similar ground. ## Mark up and mention entities on the page ### Use Schema.org markup Schema.org markup tells Google which entities a page contains. An Article can list the entities it mentions: ```json { "@context": "https://schema.org", "@type": "Article", "headline": "Complete Guide to Running Shoes", "author": { "@type": "Person", "name": "Jane Author" }, "mentions": [ { "@type": "Brand", "name": "Nike" }, { "@type": "Brand", "name": "Adidas" }, { "@type": "Thing", "name": "Pronation", "description": "The natural inward roll of the foot during running" } ] } ``` A Product can link to its brand: ```json { "@context": "https://schema.org", "@type": "Product", "name": "Nike Air Zoom Pegasus 40", "brand": { "@type": "Brand", "name": "Nike" }, "category": "Running Shoes" } ``` Markup helps Google separate one entity from another with a similar name and places the content in context. A site without it misses the most direct way to state what a page is about. For type selection and required properties, see the [schema markup types guide](/articles/schema-markup-types-guide). ### Make the first mention unambiguous The first mention of an entity sets its identity. After that, pronouns and short forms work. Weak: "The company released new features last week." Strong: "Google released new search features last week." Link important entities to an authoritative page, such as Wikipedia, the official website, or a knowledge base entry. The link helps confirm which entity you mean. Define technical entities briefly where they first appear: "Photosynthesis, the process by which plants convert light energy into chemical energy, is essential for..." The definition gives the entity semantic context. ## Build topical authority with depth Sites that cover a topic area in full tend to rank better than sites with scattered, shallow coverage. ### Use the topic cluster model A pillar page covers the topic broadly, at 3,000 to 5,000 words. Cluster pages go deep on subtopics, at 1,500 to 2,500 words each. Every cluster page links to the pillar, and the pillar links to every cluster page. For example, a pillar called "SEO for e-commerce: complete guide" could anchor four cluster pages. They cover product page SEO, category page optimization, e-commerce site architecture, and schema markup for product pages. The structure tells Google that the site covers e-commerce SEO in depth. The [topical authority content strategy guide](/articles/topical-authority-content-strategy) shows how to plan the full map. ### Show depth on each page A complete running shoes guide covers the types, how to choose by foot type, and brand comparisons. It explains the technology: cushioning, drop, and stack height. It also covers care and replacement, common mistakes, and the questions buyers ask. Add images, video, and diagrams where they explain something the text cannot. Update the page when facts change, and show the date of the last update. ### Put related entities together Pages that mention related entities together can rank for queries about the group. A page that compares Asana, Trello, Monday.com, and Jira fits the query "best project management software" better than four separate pages. In a comparison, cover every major entity in the category, not only the one you recommend. ## Write in natural language Keyword stuffing is a named violation in Google's spam policies. It also reads badly: Weak: "Our running shoes are the best running shoes for runners who need running shoes for running." Strong: "Our shoes give distance runners the support they need for marathon training." Vary the words where the meaning allows. Instead of "electric vehicle" 20 times, use "EV", "electric car", and "battery-powered vehicle" where each fits. Google's language models learned from natural text such as books, articles, and conversation. Text that reads the way people write fits what the models expect. The [guide to content optimization without keyword stuffing](/articles/content-optimization-without-keyword-stuffing) covers the edit pass in detail. Structure the page around the questions people ask. Headings such as "What are the best running shoes for beginners?", "How do I know if I need stability shoes?", and "What is the difference between trail and road shoes?" match how people search. Add sections that can win a featured snippet, even on a long page: - a definition of 40 to 60 words for "What is X?" queries; - a numbered or bulleted list for "how to X" and "types of X" queries; - a comparison table for "A vs B" queries; - numbered steps for procedures. ## Measure semantic SEO Six measures show whether the work lands: | Measure | How to check it | |---|---| | Coverage against competitors | A content score in MarketMuse, Clearscope, or Surfer SEO compared with the top-ranked pages | | Entity coverage | The entities in the top 10 results, compared with the entities on your page | | Rankings for related queries | Rank tracking for semantic variants and related queries, not only the primary keyword | | Featured snippets | The share of target queries with a snippet that your page holds | | Cluster completeness | Subtopics in your map that still have no page | | Engagement | Time on page and pages per session, which should rise as pages satisfy intent better | The simplest test is in Search Console. If your pages rank for related queries and variants you never targeted directly, Google reads them semantically. Pages that win featured snippets show strong semantic relevance too. Content can rank without the exact keyword when it covers the topic fully, because neural matching connects the query to it. The keyword in the title and headings still helps. ## Avoid five common mistakes | Mistake | Why it fails | |---|---| | Exact-match repetition | Awkward repetition reads as stuffing, and Google's spam policies name it | | Entities without relationships | A list of product names with no comparison or connection gives no context | | Keyword checklists without depth | Google's systems recognize thin content even when every keyword is present | | Keywords that miss the intent | Semantic SEO needs the page to match the searcher's goal; see [search intent explained](/articles/search-intent-explained) | | No structured data | The most direct statement of what a page is about goes unused | Semantic SEO does not replace older SEO practice. It includes keyword work and links, and adds entity work, topic clusters, and structures that language models read well. ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```