What is Semantic Search and why is it important for SEO?

When Google first started, they used "Keyword Search" to index and rank pages. This had many limitations but overall gave a reasonable result as long as you knew the right way to search and chose the correct keyword phrase to search for. However, over the last few years Google has started moving to "Semantic Search" and this impacts the best way to build a website and do search engine optmisation.

What is Keyword Search?

Keyword search is where you put a specific keyword phrase into Google and they show a list of website pages that include that keyword. It really is quite simplistic and in the early days lead to dodgy SEO tactics such as adding the same keyword phrase multiple times behind images etc. The idea was to have the keyword phrase mentioned more times than your competitor site and still have the page make sense.

The limitations of Keyword Search

The main limitation to keyword search was it is too simplistic. You need to think about the keyword phrase you are going to use, otherwise your results may not be ideal. A good example of this is if you searched for "leaking taps", Google would only show you websites that have the words "leaking taps" in them. They wouldn't show you general plumber pages if you didn't mention "leaking taps".

Of course, the other limitation was that it was easy to manipulate the search ranking. Just make sure that you had the keyword mentioned more times than your competitor's website.

How Semantic Search works

As Google has developed, they have moved more into machine learning and as a result natural language processing (NLP). This has allowed their search algorithm to look at searcher intent, query context and the relationships between words. Let's look at these 3 areas in more detail.

Searcher Intent

One of the difficult parts of keyword research is trying to decide what the searchers intent is. What do we mean by this? Let's use a common example we continually grapple with. If someone searches for "blocked drains", what are their intentions? Are they wanting to find a business that can help them unblock their drains or are they looking for a website that explains how they can unblock the drains themselves.

What about another example; someone searching for "electrician" is most likely looking for an electrician business whereas someone searching for "electricians" may be looking for a directory of electricians.

Query Context

This is where Google looks at the context around your search. Once again, let's use an example. A simple example would be if you searched for "plumber" when you lived in Brisbane; Google would know you are in Brisbane, assume you are looking for a plumber and show you a list of plumbers in Brisbane. With keyword search they would show you a list of plumbers from anywhere.

Similarly, if you don't know how to spell a word correctly, Google would attempt to work out what you really mean and show these results. So "electrican" would show results for "electrician".

Relationships between words

In this case, Google looks at words that are common alongside the main keyword. So Google understands that a Plumber handles hot water, blocked drains and leaking taps. So even though you may not have this on a page, Google will make the assumption that you do these things. In other words, Google is attempting to understand language in the same way that humans do. Previously, if you searched for "italian food", you may not have got any straight up pizza cafes showing; as Google didn't realise that pizza is a form of italian food. Now they will show pizza cafes alongside italian restaurants.

Why Semantic Search?

There are many reasons why Google has decided to move to semantic search. Some of this has been driven by changing technologies (eg. machine learning etc) and others from how we have changed the way we search. Some of the main reasons are:

  • Semantic search gives a much better search result then the old keyword search. It better understands what you are searching for.
  • With the growth in voice search, we now use longer and more natural search requests. For example, instead of searching for "plumber brisbane" we say "find me a plumber to fix my tap".
  • It's harder to manipulate the search results, using keyword stuffing and backlinks for specific keywords.

What does this mean for SEO?

In the beginning we did keyword research for what people searched for and then created a page for each of the top keyword searches. When we created this page, we made sure that we mentioned the keyword phrase several times throughout the page.

Now we look for keyword topics. That is, we group similar keyword phrases with the same intent and similar meanings together and create a page for these. We then write this page in a more general sense, ensuring that we fully cover all areas of the topic. We don't need to get too hung up on keywords themselves, but more the relevant words we would expect to see.

Overall, semantic search has benefited onsite SEO, where you create good content for your website. Google has stressed for years that good SEO meant writing great content for humans, not bots. Now with semantic search, they are better rewarding those doing SEO this way.