What is Semantic Search?
Semantic search is essentially an intelligent search process that seeks to generate more accurate results by focusing on intent and the “contextual” meaning of keywords being searched for rather than simply looking for web pages that have been optimized for the keywords used in a search query. Google’s Hummingbird algorithm is powered by semantic search technology.
In the good old days, SEO revolved around the express manipulation of keywords and links. Effective SEO techniques consisted of creating low quality, keyword-rich articles and PBNs with thousands of artificial links pointing back to the target website using exact match anchor text.
Semantic search has shifted the focus from simply reading the keywords typed into the search query box to understanding the mindset of the searcher by analyzing the keywords used in search. With semantic search, Google’s main objective was to bring a full stop to the manipulation of its search algorithms. It wanted to stop pages ranking simply because certain keywords appeared so many times on a particular web page and in backlinks to that page.
That system was easily manipulated using the underhanded techniques described above, and those activities often led to inaccurate and unreliable results.
Focus on Searcher Intent
With semantic search, rather than attempting to match the individual keywords used in a search query, the search engine looks at the entire query as a whole and delivers search results that are based on the “searcher’s intent”, effectively moving from keyword-based searches to context-oriented searches. In simple terms, it considers the sentence in its entirety, understands the searcher’s intent and then looks for the most relevant results for the search query.
Semantic search has done away with the inherent ambiguity of conventional search. For example, pre-semantic search, a search like “where can I buy black suede shoes in New York?” would return web pages that match each keyword in the query – buy, black, suede, shoes, New, York. It would also look for pages that contained the query in its entirety. All things being equal, the most authoritative web page that included the exact phrase would have been ranked highest.
By contrast, semantic search considers the entire sentence as a whole and tries to understand the searcher’s actual intent, rather than simply parsing through each of the keywords used in the search query.
Rather than looking for web pages optimized for the keywords typed into the search query box, the search engine looks at web pages that match the user’s intent and surfaces pages that are competing for the keywords that are searched for. This means that optimizing a particular web page for specific keyword phrases has become less effective if the “theme” of the web page does not match the intent of the searcher.
In fact, for an increasing number of search queries, most of the top 10 search results on Google are not even optimized for the keywords used in the search query. This means Google’s algorithm is better able to find and return the “most relevant” web pages to a particular search query, even if a particular web page does not contain keywords used in the search query. Consequently, optimizing a web page for specific keywords are no longer effective in driving traffic if Google’s bot deems that those keywords do not match the searcher’s intent.
However, it is important to bear in mind that keywords and keyword optimization are still extremely important for SEO, and continue to play a central role in search. The fact is, Google’s semantic search index is still a work in progress and far from complete.
As a result, if semantic search is unable to adequately establish relevance to a search query, Google will revert back to a more conventional approach through the use of some of the ranking signals associated with “old SEO” practices. This is why it is still essential to reinforce the content you produce with relevant keywords, and keyword optimize the on-page elements of your web page accordingly.
However, the focus is now on returning pages that are deemed relevant based on the intent of the searcher rather than those that are best optimized for the keywords used in a search query.
Universal (Blended) Search
With the advent of universal (blended) search in 2007, the search engines integrated all of the different verticals directly into the search results page. Today, depending on the keywords you use to search, a typical page of search results for a given query may now include local results, travel information, maps, directions, video, images, a news feed from a leading news site, a relevant Tweet, a Google+ page or Facebook comment.
For example, a branded search for “new york hilton” returns a dedicated page featuring the following assets:
- the hotel’s paid ad
- YouTube videos
- indented page result links (the site’s most requested pages)
- latest tweets
- LinkedIn page
- Facebook page
- Knowledge Graph listing; and
- other search results related to the hotel.
Consequently, SEO has become less about getting a single site ranking at the top of the first page of Google, and more about acquiring multiple listings on a search results page to increase visibility and brand awareness.
Prior to semantic search, if your website was the best result for a search query, Google typically gave you real estate on the front page in the top two positions. That has now changed. Today, if you own several web properties, you may be able to dominate the entire front page with the following web assets:
- a Wiki page,
- Customer reviews
- Local results
- Knowledge Graph box,
- Latest news
- YouTube videos
- Google Map
- Google My Business page
- LinkedIn Company Page
- Social media profiles
- Yahoo answers
- Press releases
- Podcasts, etc.
SEO today is about applying SEO techniques to all of the digital assets and web properties you own so that you can really dominate the front page of Google for your chosen keywords.