CMS resources
Marketing insights
Back to Marketing insights

SEO trends: Understanding natural language search 

image of Google Home voice-activated device

What is natural language search?

Search engines like Google and Bing are working to deliver search results that respond more accurately to questions phrased in conversational language. The search language characterized by robotic, keyword-heavy phrasing (“Empire State height”) may have satisfied our needs until now, but meeting complex requests phrased in our natural language will be key to maintaining optimal search performance in the future.

Miles De Feyter
By Miles De Feyter
March 12, 2021
In order to succeed at SEO in 2021, enterprise organizations need to pay attention to the tactics and trends taking shape to stay ahead of the curve. Here are three trends to keep a close eye as you build out your SEO strategy for the year ahead.
5 Min Read

Why? Searcher intent hasn’t changed, but expectations have. With the rise of virtual assistants like Siri and Cortana plus voice command functionality for nearly everything, search engines are growing and adapting to more human-centric communication. And search engines like Google have been very explicit in articulating their goal to provide the best, most intuitive results for searchers—no matter the complexity of the query.

Why natural language?

Search is trending in the natural language direction due to a number of key shifts:

  1. Search engine evolution means users expect perfect results on their first attempt
  2. Search technology has advanced enough to be able to handle complex reasoning
  3. The rise of voice activated technology 

In other words, the trend towards natural language underlying future search parameters means we’re teaching computers how to think like us. In turn, publishers stand to benefit from the advancing work in this area to make their content more effective.

What are the potential questions your customer is going to ask after the first query? Google really begins to reward these content clusters or chains. What are the questions in natural language that your customers are asking? What is their follow up question? Build content for both.
image of Corbin Haresnape, Director of Digital Marketing, BodyLogicMD
Corbin Haresnape, Director of Digital Marketing, BodyLogicMD

Creating content with natural language search in mind

As the future of search trends in this direction, the content production paradigm requires a shift. Fortunately, there are tools that can help publishers better deliver the right content for their audiences. What was previously in the realm of creative know-how or intuitive insight can be tested against things like Google’s Natural Language API.

Google’s Natural Language API helps content creators "derive insights from unstructured text using Google machine learning." However, the API as it exists today is far from perfect—it doesn’t allow for much tampering, still struggles to derive perfectly accurate context and seems to work best only for English. That said, publishers stand to benefit from learning with the tool as it grows, as this technology will only become more enmeshed with search functionality.

As an example of how publishers can start to incorporate this thinking into content creation, we looked at our most recent post—Headless CMS explained in 5 minutes—through the lens of Google’s Natural Language API demo.

screenshot of Google Natural Language Search api result

The API allows you to run several layers of analysis on a piece of content in the following categories:

  • Syntax 
  • Sentiment
  • Text classification or categories
  • Entity 

And here are some insights we gleaned from a natural language analysis for this specific piece of content, all of which can be used to ensure you are crafting the right content to reach the right audience in search—and beyond.

1. Syntactical analysis
Syntax breaks down text into individualized tokens to find their relationship to derive meaning. The results don’t return much insight other than showing how Google classifies text, but viewing this is a further reminder to ensure your writing is clear and grammatically correct.
screenshot of Google Natural Language Search api demo
2. Sentiment analysis
Sentiment analysis determines the positive or negative emotional affect of a piece of text. The analysis shows a paragraph by paragraph breakdown of the numerical score (between -1.0 for negative and 1.0 for positive) and magnitude or intensity of the emotion.

Example analysis findings
In our effort to explain what a headless CMS system was, we see that the article delivered a positive, but closer to neutral assessment in the 0.1 sentiment score, with a magnitude of 17—highly positive. The difference between the two? Magnitude is not normalized. Every expression of emotion (both positive and negative) contributes a numerical value to magnitude.

Publisher benefits
Sentiment analysis can help content creators determine if their content - whether it’s an article, video, or series of product reviews—is hitting the right chords with their audience in terms of their intent and persuasiveness. Publishers can look to Google’s Natural Language API prior to a piece’s publication to optimize for performance, or after the fact in the case of product reviews to assess customer satisfaction.

Is your content meant to be strictly informational? You’d be aiming for a neutral score in terms of both scale and magnitude. Are you trying to win audiences over to your way of approaching certain technology? You may opt for a subtle approach, using the API to make sure that you give equal attention to pros and cons in order to gain trust. Is your new product winning customers over as intended? Analyze your reviews in their entirety with the API to get an overall assessment.
screenshot of Google Natural Language Search api demo
3. Categories or content classification
Content classification seeks to find the category of your content. What “bucket” of organized interests does your content fall into in terms of its general topic and any subcategories?

Example analysis findings
No surprises here—our piece on headless CMS falls into the Internet Software category with a 90% confidence score.

Publisher benefits
Our own piece had a very distinct audience and subject matter, but these results become interesting when we look at a different subject matter. For example, say you’re selling specialty items such as digital cameras or running shoes for marathon runners—secondary and even subsequent categories of content become important to review. A digital camera meant for a professional may return “Hobbies and Leisure” as a category that needs to be tuned. Are those shoes more suited to someone searching in the Fitness category or Sports & Entertainment? Evaluating your content through this lens may reveal areas for improvement for your entire content strategy. Checkout the complete list of categories here.
screenshot of Google Natural Language Search api demo
4. Entities
Entity analysis looks for weightier, “known” entities such as proper nouns, public figures or landmarks in addition to common nouns such as points of interest (restaurants, parks, etc.). While we don’t know to what extent, we have a sense that entities have the largest impact on SEO. The entities themselves and their salience—or the importance of that entity to the overall context of a piece of content—are classified as a way to determine what questions your content may answer.

Example analysis findings
“CMS” and “content management system” are the terms most salient to the overall content piece—easy enough. We see that we’re starting to expose some of the less intuitive reasoning in the API, which categorizes "Headless CMS" as an organization.

Publisher benefits
Our article analysis showed a simple rendition of SEO-enabled content, but what if you’re trying to bridge multiple topics under a new idea? While there’s no one solution to optimize content for entity salience, this score can become a barometer for your success in hitting on the true essence or subject you intended to reach audiences with.
screenshot of Google Natural Language Search api demo

Conclusion

Natural language search will be a key piece of each search giant's engine, so publishers need to optimize content accordingly. The shift marks one of the biggest moves from a strictly keyword-tuned approach to one that takes into account the entire search intent and experience of an audience. Embracing this new paradigm and using tools such as Google’s Natural Language API can help content creators tailor content to match prospects with exactly what they’re looking for.

Alistair Wearmouth
By Alistair Wearmouth
February 28, 2023
Amazon Comprehend powers Brightspot's "Suggested Tags" feature. When publishing a piece of content, users have the ability to request a list of suggested tags that are relevant for the article they are editing. Comprehend's key phrases/entities detection matches keywords relevant to the story against existing tags in the Brightspot instance, then offers these as suggestions increasing publishing efficiency for editors and reporters.
1 Min Read

Share

Related resources

Let us give you a demo
Hear how Brightspot can turn your digital strategy goals into a reality and see how the lives of your content creators and developers will be changed using our platform.


Request Demo