What Is Semantic Search and What Should You Do About It?

Posted on: November 15, 2016 by in Local SEO Strategies
No Comments

SEO specialists spend massive amounts of effort trying to get Google on our side — to see the brilliance of our content, the ingenuity of our meta elements, and the genius of our organic strategies. We spend so much time treating Google as a metaphorical friend that we sometimes we lose sight of the overarching picture: that Google’s (sometimes magical) results are built upon an algorithm seeking conversation.

Algorithms can (and do) solve many problems, but having one match the conversational level of human beings presents an enormous challenge. Engineers at top Mountain View companies have pushed and pried to move computer science into the realm of artificial intelligence. Their wins in the world of artificial intelligence and machine learning have been impressive, resulting in a new champion in Go, appearing in local stores as pseudo-employees, and has been predicted to drive even more fruitful conversations with our now personal assistant phones. Search engines have always been at the forefront of driving the AI initiative.

Since its beginnings, Google has been pushing search results into the realm of natural conversation, and a huge component of its strategy has been categorized under the umbrella of semantic search and, subsequently, machine learning algorithms (think: RankBrain).

So, when it comes to showing up in Google, what does it take to rank #1 now? Many of the following elements will come back to the idea of a simple conversation.


It comes as no surprise when Hollywood fantasizes search engine founders as the creators of AI in Ex Machina.

What is semantic search?

The word “semantic” refers to the meaning or essence of something. Applied to search, “semantics” essentially relates to the study of words and their logic. Semantic search seeks to improve search accuracy by understanding a searcher’s intent through contextual meaning. Through concept matching, synonyms, and natural language algorithms, semantic search provides more interactive search results through transforming structured and unstructured data into an intuitive and responsive database. Semantic search brings about an enhanced understanding of searcher intent, the ability to extract answers, and delivers more personalized results. Google’s Knowledge Graph is a paradigm of proficiency in semantic search.

Why do engines pursue semantic search?

From an engine’s perspective, it’s not hard to imagine why Google would want to pursue a more connected world: more data, less spam, a deeper understanding of user intent, and more natural language (i.e. conversational) search. Understanding all of this data maximizes the possibility of their users getting the best search experience possible.

With the world’s data doubling every two years, big data has become the norm for players in the online realm. All this data creates an overarching concern of “What does this mean to me?” The process of organizing, structuring, and semantically connecting data is a coveted role for search engines.

One of the ways that semantic search helps Google is by identifying and disqualifying lower-quality content. Methods like article spinning and keyword stuffing are more easily flagged due to advanced systems such as latent semantic indexing (LSI), latent Dirichlet allocation (LDA), and term frequency-inverse document frequency (TF-IDF) weighing schemes, which use term frequency and their predetermined weighted relationships to determine quality. This means that search engines have a good idea of what words statistically occur together and make semantic correlations, which can be used in the war against spam.

Using semantics and entity-based search, engines can gain a better understanding of what users may want. For example, the image below shows a simplified illustration of what the data in an entity-based search algorithm would contain. It includes entities (people, places, things, concepts, or ideas) which are represented as nodes, and connected by their relationships as the arrows. The diagram shows how entity-based search seeks to connect various entities, in this case the individual Simpsons characters, which creates more depth to search responses.

Illustration of a semantic taxonomy of entities (as nodes) and attributes (relationships).

Knowledge Graph answer driven by the entity-based search associations.

Semantics help to understand more completely what our searches mean today. For example, a search for [Jennifer Lawrence] is most likely related to the American actress, star of the Hunger Games, and fashionista. Google provides news, photos, facts, social media accounts, and movies all related to Jennifer Lawrence. Through understanding entities, and coupled with the perplexing amount of data behind the habits of the 7.4 MM searches for Jennifer Lawrence, search engines can gain a better understanding of what the next user will want. Google’s invention of the Knowledge Graph is a golden example, aiming to understand things, not strings.

Search for “Jennifer Lawrence” appears with only topics related to the actress.

Google, and other engines, have become very adept at recognizing different entities and formulating answers to questions. And it’s through this connecting of data that search becomes stronger. Answers to questions are algorithmically understood and displayed when, for example, one searches “who is the dancer in the chandelier video?”. Google “knows” that it is Maddie Ziegler. The idea that a search engine can connect the keywords to an entity and reply with the accurate answer makes Google’s search much more constructive for its users.

Knowledge graph answer with entity name not mentioned in the query.

SEO implications

For SEOs, understanding semantic search has some major benefits. A large part is the ability to remain ahead of the curve. Search engines are moving forward and as SEO experts we need to make sure to stay at the top of our game. Semantic search is going to become especially important as voice search gains more traction.

The method of integrating semantic search signals has huge implications about how we approach our SEO strategies. If we could know all of the topics and keywords associated with a particular entity, we could create perfect content and achieve the optimal rankings for our clients. Although we live in an entity-not-provided world, there are a few tried and true strategies that can enhance your semantic search strategy.

SEO semantic search strategies:

1. Provide value.

Google is looking towards AI and envisioning conversation as the next evolution of search technology. Google CEO Sundar Pichai even mentioned during the Google Assistant reveal, “We think of it [Google Assistant] as a conversational assistant; we want users to have an ongoing two-way dialog.”

Google needs a source of information for all of its conversations, a reference point, an expert friend, its trusty companion in the Wild West of the WWW. Become authoritative in your discipline, become the expert source that Google will reference in its conversations. Become that valuable source that drives connection, exchanges information, and provides visitors with some value.

  • Recommendation: Determine what you want to be known for. Answer the following questions, then create a killer organic search strategy based on your findings.
    • What are the types of keywords that you want to rank for?
    • Who is currently in that space?
    • What are they doing that makes them the expert?
    • How can you be 10X better?
    • Who is interacting with your content?
      • Are they existing customer or prospects?
    • How are users interacting with your content?
      • Are users converting?
      • Is this content targeting users early in the funnel (awareness and consideration) or later (conversion)?
      • Are they getting what one would anticipate out of the content (i.e. are they finding the answers they sought out by clicking on your content)?
    • How can you improve your users’ experience with your content, including their customer journey throughout the site?
    • How can you reach your target better at every organic search touch point your customer encounters?

2. Develop targeted content that answers your customer’s questions.

Create targeted non-brand content, which doesn’t interfere with your acquisition-focused online assets (Think: Don’t cannibalize traffic from your product pages). The idea is to create content related to the entity of your product line, which interests users and fills gaps in organic visibility. Become a valuable source of information for your customers, build your semantic authority in the “eyes” of search engines, and become search engines’ go-to guru on the topic via building robust, informational content using mixed media (images, graphics, and videos).

  • Recommendation:
    • Prioritize non-brand content with strong question/answer focus.
      • Tip: Google appears to prefer numbered lists or bulleted step-by-step instructions that succinctly answer questions.
    • Perform keyword research to determine opportunity for queries that are being searched with “how to,” “why,” and “what is” questions.

3. Structure sentences clearly and answer-based.

SEO writing is natural language writing. Content should use natural language. This simply means that content should make sense. With Hummingbird’s improvement on precision and semantic search, along with RankBrain’s machine learning ranking factor incorporated in 2015, also throwing in the growing popularity of voice search — natural language is necessary.

When creating content it is important to write in terms of entities which means more noun-focused sentences. Simple, subject focused sentences provide engines with more information. Try to structure sentences as Subject Predicate Object (SPO). This will make the content easier for users to understand, as well as for search engines to parse the information. The key here is to sound natural and construct your sentences with purpose, writing content that directly answers a question.

  • Recommendations:
    • Write in natural language.
  • Structure your sentences so users and bots can easily understand what you’re trying to say.

The image above is a screenshot from the KDD 2014 Constructing And Mining Webscale Knowledge Graphs by presentation (presented by a Facebook and Google engineer). It shows how strong sentence structure can support bots’ grasp of content. The first two structures are able to be parsed and data organized. However, the last sentence offers no data.

4. Structure your data to help bots parse content.

Structured data markup annotates information, which is already on web pages, to add clarity and increase confidence for search engines. Using structured markup not only enables search engines to better grasp content, but also can be used to signal a desire for rich search results. These snippets provide users with additional information about the contents of the page and can improve click-through rates (CTR) from organic search.

Make sure that all marked-up content is visible on the page per Google’s structured data policies. Google lists all active and in-use structured data markups with examples (check them out regularly, because they update all the time!).

  • Recommendation:
    • Implement structured data that is relevant to your site.
    • Examples:
      • Ecommerce:
      • Brick and Mortar Store:
      • Blog:
      • All Sites:

5. Leverage internal linking.

Internal linking has long been a method of indicating topicality, supporting the user experience as they navigate throughout your site. Remember to use internal linking sparingly and only when it is in the best interest of the user.

  • Recommendation:
    • Identify thematically relevant internal linking opportunities to target landing pages.
    • Important pages should be referenced in your main navigation or global footer
    • Important pages should be referenced in your site’s XML and HTML sitemap
    • Add contextual links within pages to important pages on the site
    • Fix any links that lead to pages returning 3XX or 4XX status codes
    • Always link to end-state canonical URL and not a URL with a parameter


  • Search engines are incorporating semantic signals in their results. This change requires webmasters to integrate synonyms and related content for each target topic.
  • Semantic search provides additional meaning for engines: data, spam, answering user questions, establishing more personalized results, and providing a more conversational user experience.
  • Semantic search high-level strategies: Provide value to your visitor, answer your customers’ questions, create content with structured sentences, and implement structured data.

Comments are closed.