If you start searching around the web for those "What to look forward to in SEO this year" articles, you’ll invariably come across advice that says “optimize your site for semantic search”.

But what does that mean? Is it hard? Can you do it yourself?

A couple of different things. Not really. Yes, you can!

What is Semantic Search?

Search engines and search marketing is changing. In fact, they’ve already changed quite a bit. Things like Hummingbird and RankBrain have changed the ways that Google processes both queries and page content. These two algorithms impact the way Google decides what a user is trying to accomplish and how relevant your pages are to that goal.

That’s what semantic search is: determining a user’s search intent and finding pages that achieve that intent.

There are three basic types types of search intent:

  • To go somewhere: These queries are basically people looking for directions around the web. They know where they want to go, but aren’t sure exactly how to get there. Think people who search on Google for "facebook" instead of going to facebook.com.

  • To find something: These queries represent people who want to find an answer to whatever question has popped into their heads of find more information on a subject. Informational queries range from specific questions ("how old is the queen") to more general terms (“queen elizabeth ii”). These are the vast majority of queries processed by Google.

  • To do something: People use these queries when they have a specific action in mind, usually involving some sort of commercial intent. They might include words like "buy", “order” or product names, but not always. Transactional searchers could also be looking to download some content or conduct a vertical search (like for flight prices).

Semantic search isn’t limited to query processing. It also includes how Google ranks and displays the results of searches performed by users. This change is reflected in the way SERPs look now.

If you want a blast from the past (or two), check out the changes to Google’s results page over time here.

How is Semantic Search Different?

The biggest difference between semantic search and the old keyword-based system is summed up in one word: precision. You could also go with accuracy too I guess.

A quick practical example that shows what we’re talking about here.

Let’s say I do search for "leaky faucet".

Using the old system, any page that talks a lot about faucets or leaks has a chance of ranking highly for that query. Which means a lot of pages about the different faucet models, ecommerce sites that sell faucets or maybe some plumbing supplies companies.

But what do you think I want when I search for "leaky faucet"?

But almost anyone searching for "leaky faucet" is probably looking for a way to fix their leaky faucet.

A plumbing supplies company isn’t much use to me in this situation.

However, thanks to advances in semantic search, Google now correctly interprets that I’m looking for solutions to my leaky faucet problem:

Leaky faucet semantic serp

As a user who doesn’t want to call a plumber, these results are much more helpful and valuable to me.

How to Get Your Site Ready

Semantic search is happening, so you need to get your website ready for it.

But how?

Well, we’ve got three guiding principles for you to follow in order get your website ready for Google’s semantic search algorithms.

Optimize for Topic and Intent

Since intent is a big part of what makes semantic search work, it makes sense that optimizing your content for intent is a big part of making your semantic SEO work.

So how do you build your content around search intent and topical relevance?

Intent Optimization

Once you’ve decided on a good topic, think about what the situations in which someone would be searching for that topic. If you’ve created marketing personas for your business, this shouldn’t be too hard. Just look at the goals and/or challenges sections.

You could also turn to your customer service team to collect questions asked by customers/leads.

If you have a keyword research tool, there’s a good chance it will give you suggestions for questions people ask about the topic.

If you don’t have a keyword tool, you can always turn to Google itself. Do a search for your topic and scroll all the way to the bottom of the SERP. Here, you’ll find the questions people are asking Google.

Related searches at the bottom of Google SERP

Once you’ve come up with your questions to answer, there are a couple of tricks to help Google understand:

  • Put the question in your title tags and <h1> content. For how-to guides, include each step in a subhead.

  • When answering a question, put the answer in the body content directly after the header tag. Resist your urge to write a long introduction.

Google’s Answer Box is limited to 50 or 60 words, so keep all of the important text under that.

Relevance Optimization

Optimizing your content for topical relevance sounds pretty straightforward. Write about what you're writing about.

But how can you tell if you're actually being successful?

In the old days, you could simply measure where and how many times you used your keyword and call it a day. Thanks to the semantic web, we've moved past that.

The better measure of your content's relevance will be to measure how (or if) your users interact with it. Some metrics you should track to measure if you're relevant to your audience include:

  • Time on site

  • Bounce rate

  • Events (watching a video, enlarging an image, clicking a button, etc.)

  • Scroll rate

Be a Knowledge Graph

A vital part of using semantics in your SEO is making sure Google will understand your page content. Google is smart, but it can’t quite understand human language quite yet.

Luckily, there’s a way to make Google understand your context: structure data.

Structuring your website’s data helps Google make connections between the text on your page and the data stored in a machine-readable way. It’s not required, but Google has a slight preference for JSON-LD since it can be managed using Google Tags Manager.

Adding schemas (taxonomies that explain to search engines how pieces of data relate to each other) by linking between entities will make your website look authoritative and relevant on a subject.

If you aren’t sure how semantic markup can help your content, or how to go about adding it to your site, check out WordLift. WordLift is a great resource on linked data and pretty nifty tool when it comes to adding it to your articles.