By Laia Cardona, on 6 November 2023
NLP (Natural Language Processing) is a change in Google’s algorithm that affects SEO. NLP focuses on understanding the context of Google searches, not just selected keywords.
If you’re serious about kick starting or maintaining your page rankings, you should begin implementing NLP SEO tactics as soon as you can.
In this post we’ll breakdown what NLP is and how it applies to SEO, as well as techniques you can use to help your rankings.
Natural Language Processing Definition
NLP (Natural Language Processing) is a linguistics subfield for understanding communications between humans and computers. Using computing and artificial intelligence, NLP attempts to find the context of these communications and the natural patterns that occur. Perhaps the easiest way to get a grasp of how NLP applies to SEO is to first understand BERT.
Understanding Google’s BERT and SEO
“With Google weighing in on NLP to gain a deeper understanding of user’s queries, it implies that content-makers who get more specific, relevant and descriptive with their content and information (including links) in their pages tend to rank higher.” - Jaya Kumar Data Scientist, Deep Learning and NLP Specialist
Google introduced BERT (Bidirectional Encoder Representations from Transformers) in 2019. This is Google’s latest NLP algorithm which helps to better understand search queries in the way that humans would naturally. BERT is neural-network based and open sourced, and considered one of the biggest leaps in understanding search intentions.
One main thing that sets BERT apart is its bi-directionality which means it can understand the context of phrases based on the words or sentences before and after them. In other words, basing search results not just on key words used but in what conversational context. And as with understanding anything in the real world, context is key. For example the difference between just ‘horse’ and ‘rocking horse’. Additionally, BERT can make more sense of prepositions like to, in and for—which can greatly complicate search queries.
BERT’s impact on SEO can be summarized like this: it's less about keywords and more about the content around those keywords: whole sentences, paragraphs and overall sentiment.
How Does Sentiment Impact SEO?
What do we mean by sentiment? Sentiment refers to the undertone or feeling of the search query. There’s positive, negative and neutral sentiment and it is defined based on the types of words used in the search. Words like ‘great’ obviously being positive and ‘unsatisfied’ being negative. Sentiment is specifically helpful for things like reviews and voice analysis. Google has a sentiment scoring method which looks like this.
Google’s sentiment score: Green is positive, yellow neutral and red negative.
If most ranked pages for a given topic (or entity, which we’ll get to) have an overwhelmingly positive sentiment and yours has a negative sentiment, there’s a good chance NLP will reduce your rank as you won’t be seen as relevant.
Entity, Category, and Salience
Other BERT metrics that you need to understand in relation to NLP and SEO are entity, category and salience.
- Entity: Word or phrase that represents a tangible object such as people, places or things. BERT identifies and evaluates these entities.
- Category: As with any SEO work, category is always important and are the keywords people search for most frequently. NLP will separate text into subject categories which you can explore here.
- Salience: How significant the entity is within the given text. These entities are given a salience score based on how relevant they are.
Natural Language Processing Techniques
Hopefully you now have a firmer grasp on what NLP is and how Google is applying it to search using BERT. Now we can introduce some NLP techniques for SEO that can help you improve or maintain your rankings.
Overall you want to create high-quality, well written content. Ask yourself: Is this text the best it can be as a readable and logical source of information on this topic?
Google API Demo
Thankfully Google has provided us with it’s NLP API Demo. You can plug in any text you’d like to analyze and see where you can improve it for SEO. From here you can compare your text to the text on pages that are dominating SERPs and make necessary changes.
Internal and External Links for NLP
Link structure and placement on a page have become more important than in the past. So besides authority and relevance, links need to fit into the greater context of the article. Where are your links inserted? Their placement needs to make sense in terms of the flow of the text. The anchor text is also an important factor.
Keyword targeting has typically focused heavily on keyword placement and related keywords. Now indicator keywords, or words that often lead or build up to keywords, are considered important for NLP SEO. So using keyword tools that basically give you variants of your target keywords isn’t quite as useful as it once was. Using these indicator keywords as a way to generate context or relatedness can give your salience a boost. Aside from Google’s API tool, IBM also has a tool which can help you identify areas for salience improvement.
Focus on Readability
Focus on the reader. How many times have we heard that? By giving the reader what they want you’re essentially answering their search queries. One way to do this is to create content templates that make formatting automatic. Text that has a natural flow, hierarchy and transitions, and basically makes sense, is a simple way to get in line with NLP.
Reduce Distance Between Questions and Answers
If the query is, ‘How to build a treehouse’, avoid going off on tangents or full blown stories. Give the answer to the question as soon as possible. It doesn’t have to be immediately but allow the article to be skimmable and for people, and Google, to get the information they want without too much effort.
When Google crawls your pages using BERT it’s important that it can easily decide if your content is relevant to an entity. And, not only if it’s relevant but if your content answers the question being asked. If your sentences are too complicated and BERT needs too many steps to build relationships between words, you’re not going to rank well. In other words, how would Hemingway write this content? That’s right, lean and free of unnecessary adjectives and punctuation. Don’t beat around the bush, get to the point.
Provide Valid Information
It might seem like an obvious tip but make sure your content is up to date and gives correct information. Things change relatively quickly, especially concerning technology and statistics.
One Idea per Sentence
If you have run on sentences that are hard to follow, it's not good for readers or for Google’s algorithm. Avoid confusing pronoun usage and multiple clauses. Make each sentence a standalone idea that is straightforward.
Avoid Industry Specific Jargon
Don’t assume readers, or Google, can understand what jargon means let alone how it answers the question being asked. Use simple language that is understandable to the layman, to those who know nothing about the topic or entity. This comes back to being lean and getting to the point. The less distance between entities the better.
Use Correct Punctuation and Spelling
Yes, these things matter. Again you’re writing for the reader. Don’t be sloppy and careless with your content because Google will notice. With all the tools available there’s really no excuse for spelling and punctuation mistakes. If you want NLP to rank you as an authority on a topic, you need to write content that is grammatically correct and has correct punctuation and spelling.
Create a Structure and Dictate Your Text
Another way to be both strategic and natural in creating your content is to build a structure and then record yourself speaking about the topic. You do obviously need to have some expertise on the topic to speak freely and with impact. This can create a very natural sounding article once transcribed. Here are some tips for better structure to get you started.
- Use an inverted pyramid to convey importance. Where content appears on a page—top, middle, bottom—gives it meaning for NLP.
- Headings clearly define the content coming next.
- Relationships and proximity matter. Subheadings have parents and the closer two words are to each other the more they're related. The farther they are, the less related.
NLP and Google’s use of BERT is an expansive topic. And while you don’t need to understand everything there is to know about NLP, creating your content for SEO with it in mind is highly advisable. As mentioned, creating valuable, quality content that gets to the point is a huge step in the right direction. Using tools like Google’s API Demo, you can then begin to be more concise and relevant in the eyes of BERT, meanwhile achieving your SEO targets.
The Future of Natural Language Processing in SEO
As technology evolve, so too does the role of Natural Language Processing, especially when it comes to SEO. Anticipating advances in NLP can be tricky but it pays off. While current strategies focus on things like Google's BERT algorithm and the impact of sentiment on SEO, we can expect even more sophisticated NLP techniques to emerge. Search engines may gain a better understanding of context, user intent, and nuanced language structures. Future developments could involve a more personalized approach to search results, taking into account individual user preferences and habits.
In addition, the integration of NLP and voice search is likely to play a role in the future of SEO. As devices with this feature become more ubiquitous, optimizing content for natural language queries will become increasingly important. Marketers should keep an eye on how search engines handle spoken queries and what types of results they prioritize as responses.
Staying informed on how NLP is evolving allows marketers to proactively adjust their SEO strategies, ensuring that they position their brands well in the ever-evolving landscape of NLP and SEO. As the industry continues to innovate, embracing these trends will be key.