<img height="1" width="1" style="border-style:none;" alt="" src="//googleads.g.doubleclick.net/pagead/viewthroughconversion/907525871/?value=0&amp;guid=ON&amp;script=0">

Power Digital Marketing Blog

    seo-3.jpg
    Nicole Grodesky
    by Nicole Grodesky |

    What Does ‘Semantic’ Mean & Why Does it Matter for SEO?

    Google has made adjustments to their algorithm over the years with a focus on the user. They want the best possible results to be returned to the user after they complete a query on  Google’s search engine. One of the most recent and complex updates was incorporating a machine learning aspect known as RankBrain.

    The problem for many business owners and marketers is being able to understand what semantic search is and how to optimize for it. In this post I will cover what semantic search is, what tools to use, and how to leverage the tools to develop a full fledged semantically search optimized content marketing campaign.

    What does Latent semantic analysis mean?

    Latent semantic analysis (LSA) draws on a couple different concepts that analyze the relationships between a set of documents (or webpages) and the terms they contain by understanding a set of concepts. LSA uses a technique called natural language processing which is a field of computer science, artificial intelligence, and computational linguistics that compares the interaction between computers and human languages.

    The main goal of this is to process a natural language corpora or samples of real world text. The challenges that come up with processing the natural languages comes down to of course a machine learning and understanding the meanings and context of a specific set of terms.

    For example, if you’re a person from the UK researching cars in the US and you type in “auto” an LSA would know that a car and an automobile (auto) are the same thing.

    Does Google use latent semantic analysis in their algorithm?

    The short answer is yes. Google uses machine learning to help process all the queries and match the best possible results for that query. Google has named this machine learning feature “RankBrain.” Andrey Lipattsev, a Search Quality Senior Strategist at Google came out and said that one of the top three ranking factors are content, links, and RankBrain.

    Google is very secretive about their algorithm and what they incorporate, which is understandable. This makes our job as SEOers harder of course, but who doesn’t like a challenge.

    Understanding RankBrain and the importance of content are key to building web pages that add value to the user and are indexed properly.  

    How to optimize for semantic search

    There are a lot of different tools out there that attempt to optimize web content for semantic search. Our favorite here is MarketMuse but there are many others that you can use to text mine as many other long-tail keywords as possible. The whole goal here is to wordsmith your content so that you’re hitting all the so called “buzzwords” that are associated with your topic.

    You might be wondering, how do these tools know what keywords to pull in. These tools scrape the top ten results for whatever search query you’re optimizing for. They pull in the top rated content and try to mimic the LSA. These tools sort and bucket the top phrases within that content. Sometimes the text can be as simple as saying “sales representative” more often than “Salesman” or “Sales Rep.”

    The Moz Keyword Explorer Tool pulls in the top results based on relevancy and offers the option to drill down into a topic based on the following criteria:

      • Include a mix of sources
      • Only include keywords with all the query terms
      • Exclude your query terms to get broder ideas
      • Based on closely related topics
      • Based on broadly related topics and synonyms
      • Related to keywords with similar results pages
      • Are questions (Great for blog topic ideas)

    Screen Shot 2017-09-01 at 11.08.04 AM.png

    Another sorting feature is to group keywords by the following:

    • Yes with lexical similarity
    • Yes with medium lexical similarity
    • Yes with high similarity

    Screen Shot 2017-09-01 at 11.08.42 AM.png

    The term lexical similarity refers to keywords that are related to the search query you’re researching, in this case it’s “digital marketing agency.” This “lexical similarity” that Moz is showing you is all based of the concept of Natural Language Processing through a natural language corpora.

    The lexical similarity includes sub-words with compound words and phrases that all correlate to each other. The idea for optimizing a web page for semantic search is to drill down into the semantic networks  and discover lexical semantic trends within the top ten SERPs. You in an essence giving Google more information on a topic and saying, you’ve indexed this page for this keyword and I’m adding more text on the topic that is high in lexical similarity. You are basically feeding Google’s bots with the data they are looking for to compare against their current results.

    Another term that is frequently used for this within the SEO and content community is “topic modeling.” The definition of topic modeling is a “type of statistical model is a type of statistical model for discovering the abstract “topics” that occur in a collection of documents. (Source) An older terms used for this concept is “word cloud.”

    Text mining Semantic Tools include:

    MarketMuse Webinar - In Content

    How to leverage semantic search

    Most every business owner knows that they need a blog, but often times what to write on the blog is total questions mark dangling in outer space. It’s often the blog strategy isn’t a priority and because of this, articles don’t move the business into a positive direction when posting articles to the company blog. If you don’t have a strategy for content with measurable goals, your content marketing campaign could fall flat. If you think this might be happening on your company blog then you will want to audit and improve your content.

    One way to really leverage optimizing for semantic search is  to build out long-form articles with a strategic topic ideation structure we like to call “contextual content clusters.”

    Long-Form Articles

    What does it mean when we say “long-form articles?”  In an article by Backlink.io they have found that “Content Length:  Content with more words can cover a wider breadth and are likely preferred to shorter superficial articles. SERPIQ found that content length correlated with SERP position.” This means that if you are going to be posting content on your blog, you’ll want to make sure it meets a specific length in word count. Having thin lower quality content on your site can actually hurt you.

    Contextual Clusters

    Google loves relevance and their bots love to correlate data. So we say, give them what they want. The first step is to research the questions people are asking about your specific industry's product or service. Then you’ll want to find similar themes and build out full length articles and link them internally to each other using keyword rich anchor text. Using the keywords you found doing your semantic text mining would also be a great place to implement internal links.  

    Optimize Your Blog Articles

    We like to use a tool called MarketMuse to optimize our web page content as well and blog articles. This tool will score the content based on the top results in the SERPs. The tool has a feature that allows you to analyze your content against a particular keyword you’re trying to rank for. The tool scans the top results and pulls in the average content word count, number of mentions of the keyword in the content, and an overall content score. You’ll want to compare your content to the top performing content and add suggested keywords to increase your word count and overall content score if your score is under the average content score.

    Screen Shot 2017-09-01 at 11.10.17 AM.png

    Putting all these pieces together are a great way to leverage Google’s semantic search algorithm to your favor.

    Quality Updates & Semantic Search

    The first main quality update that rocked the SEO world was the Panda Update in February of 2011. I started working in a the SEM industry before the major updates hit. I can remember overhearing people in content talking about how nobody reads the articles and they are just for the bots. I was sitting at my desk and I cringed.

    I came from a publishing background and wanted to stand up and shout, NO! That’s not why you write content. But I was new in the industry and was still learning. Sure enough the Panda update rolled out and everyone started scrambling and i was sitting back all cool, calm, and collected enjoying the sweet feeling of “I told you so” (even though  I didn’t).

    So what was the Panda update? Panda was an adjustment ( or update) made to Google’s algorithm that  focused on eliminating low quality content from creeping up in the top results.  Before Panda hit SEOers were creating tons of pages with content that wasn’t focused on the user. The thought behind this old tactic was build a page, rank, and win. That’s not the case anymore as not only has Google’s algorithm changes, users behavior  has evolved. So as users evolved, Google made improvements to their search engine and made quality content a ranking factor. This meant entire shift happened where people weren’t focused on the content and quality, to being all about the content. Journalism students everywhere took a big sigh of relief!

    The Panda update gets refreshed from time to time and you’ll see fluctuations in your keyword rankings and traffic.

    The most recent update focused on quality of content is the Hummingbird update. Hummingbird is said to be the update that included Google’s semantic search addition they call RankBrain. The idea behind Hummingbird was to increase the algorithm to produce quick and efficient results. This included some knowledge graph testing and roll outs.

    With all these updates Google is making to their algorithm, it’s really important to have high quality content on your site. And since page one is so competitive, there are ways to increase your chances of ranking in the top results by feeding Google more relevant keywords by mining topic modeling phrases using a variety of tools mentioned earlier.  

    What you need to know about semantic search and voice search

    The main goal behind Google incorporating machine learning AI into their search engine algorithm is to try and make searching for something more conversational. We’re seeing voice search becoming more and more popular especially since laws have been passed to be hands free as well as people adopting the Siri technology. As more people use voice search it’s really important to apply that when writing content for a web page or resource article.

    Voice search is definitely not going anywhere and will be evolving so it’s highly relevant and important for SEOers to keep this in mind when optimizing webpages. Power Digital’s Senior SEO Strategist, Kendall Brennan recommends to sues “wordier” phrases.

    “One of the biggest differences with semantic search and its connection to voice search is that the keywords are oftentimes “wordier,” meaning they include certain filler words,” Brennan said. “The way someone searches for something on their desktop is much different than the way they would search verbally. The voice search “keywords” are more question based and conversational.”

    Wrapping Up

    Google is highly focused on quality content and has incorporated maching Learning into their algorithm to make optimize a user's search results to fit their needs. There is a significant connection between contextual content, long-form articles, and topic modeling. Make sure to keep all of this in mind when you’re developing a content strategy to expand your rankings out to grab new users at the top of the funnel.

    New Call-to-action

    Subscribe here:

    Recent Posts