How voice search and natural language change SEO. And what to do about it

Disclaimer: this text was originally published on our blog

As time changes and new technologies evolve, user experience also becomes different. This includes the way people search the web for information, make reservations, orders, purchases, etc. Smartphones have changed user behavior. The search focus has shifted to local and user expectations to getting faster (even immediate) results. Now, technology’s moving further offering people voice and digital help. Alexa, Siri, Cortana, Google Assistant, and others — voice assistants are bringing user experience to a new level.

A new approach to search — voice, screenless, conversational

According to the voice report by Microsoft, 72% of respondents perform search with the help of a voice digital assistant. Smartphones and smart devices (aka smart speakers) are changing users’ approach to search. From typing to tapping to, actually, speaking. This, consequently, results in a significant change in how people formulate queries and receive search results.

Natural language. Today, semantic search outpowers good old keyword-based search, allowing for better user experience. Search algorithms become smarter, and search engines already guess the search intent behind queries to immediately answer their needs. Moreover, today search engines recognize natural language way better than some 5–10 years ago. Which is, together with improved speech recognition, actually, works in voice search favor. People like asking questions and receive answers. This is our natural comfortable form of interaction with each other. And this is right the way we like to get information. That’s why the majority of people use natural language to form their queries to voice assistants, like in a real-life conversation.

Making machines understand and communicate with users looks like some kind of magic. But there are several scientific subjects that work on making all this possible. There are parts of computational linguistics, computer science, artificial intelligence, etc.

Speech recognition is what makes machines understand spoken commands. Natural language recognition and processing lie behind the ability of machines to understand what people ask using natural language. Conversational AI — the voice assistance component — allows digital assistants to interact with users in the form of a dialogue, answering questions, asking questions, giving directions, recommendations, etc.

Assembled, they do the magic of machine-to-human interaction, changing the way users perform search (and user behavior in general). Simply put, as soon as you say, “Hey, Siri”, “Ok, Google”, etc. and make a command, some intelligent technology translates it to the machine, making it handle this command. In case of searching, the assistant searches the web through a corresponding search engine and comes up with a result.

No more 10 blue links. About 35% of voice searches are made through smart speakers that don’t have screens at all. Which means no more 10 blue links as search results. That’s bad news. A voice assistant won’t read aloud the whole SERP. It will choose the most relevant and appropriate result. Like search engines, voice digital assistants strive for maximal value for users. Thus user behavior, preferences, location, search history, etc. will be taken into account.

This makes voice search hyper-competitive and challenging. But at the same time — and this is good news, — voice search opens a large area for experiments, being at the beginning of the way. Consequently, there’s no precise recipe or some kind of basic how-tos. SEOs will have to observe, analyze, and react based on the analysis.

It’s hard to predict if voice will dominate in search in the upcoming years. But it is quite probable. Taking into account the active adoption (75% of households are to get at least one smart speaker by 2020) and use of voice digital assistants, as well as an increasing interest in smart speakers, voice search is here to stay. So it’s time to start taking care.

How to optimize for voice search

All these changes are, definitely, affecting SEO, and require several specific changes to be applied to an SEO strategy. Does this mean, you have to give up on the classic SEO? Definitely, no. You don’t need to put all your effort into voice search, not for now, at least. Voice search is still search, so old rules are applicable. Moreover, content that ranks well in classic SERPs is more likely to appear in voice search. About 75% of voice search results rank in the top 3 for that query. What you probably need now, is to revise several tactics in order to increase your chances for appearing in search voice results. There are some optimization ideas:

1. Answer questions.
2. Be readable.
3. Provide for featured snippets.
4. Implement local SEO.
5. Use schema markup.
6. Improve page speed and security.

Answer questions. Given the majority of voice search queries are made using natural language and in the form of a question, it’s quite obvious you need to provide for this. The better you answer a user’s question, the more probably it will fit for the voice search results (there’s a short guideline by Google on what answers are considered to be meeting users’ needs).

This will require more specific keyword research, as you’ll need to build content around long-tails containing question words. In fact, it’s not a big deal, as there are plenty of possibilities to learn what people usually ask. You may explore search engines’ autocomplete functions. People Also Ask boxes can be also useful.

There’s also AnswerThePublic, a great service that gathers all kinds of popular questions around a keyword and visualizes it.

Q&A pages seem to be the most obvious option to create content in a conversational manner. Apart from helping you give users maximal information about your product or service, it helps doing it in the most appropriate manner for voice search. Then, if any other page contains a precise answer to a user’s question, it also may, possibly, be chosen as a voice search result. Here’s a small note. Backlinko, in their voice search study, observed that Google prioritized short and concise answers to users’ queries, no more than 30 words long.

Be readable. Actually, this correlates with the tip above and fits for any content you create. But it’s especially critical for voice search results. Easy-to-read content is more likely to be chosen by digital assistants. Because it is, literally, easy to read. You may check your content for readability using Hemingway Editor (or any other analog, there are plenty of them). This one is super easy to use. All you need is to paste your text into the editor and follow the app’s instructions until you gain a good readability score.

Provide for featured snippets. It’s believed (and is true for about 40% of voice search results ), ranking in position zero increases your chances to be read aloud. This position is called a featured snippet — a rich search result that contains a direct and concise answer to a user’s query. Thus, it’s quite obvious to work on optimizing for featured snippets. Which basically means you need to organize your content the way it (ta-da!) answers questions. Here are some options you may use (if applicable):

• Try to build your content in a conversational manner, asking and answering questions.
• Structurize your copy in short easy-to-read paragraphs with a question form subheading.
• Create a how-to part on your page.
• Use different lists with numbers or bullet points, tables, steps, etc.

Local SEO. Nearly half of voice search queries had strong local intent, according to this study. Which means voice search fits greatly to local small businesses. Thus, it’s critical to secure a place in the local search results. This brings us to the necessity of work thoroughly on local SEO.

Here, the first and most obvious step to take is to create and fine-tune a Google My Business listing. What people usually want to know: address, information about products or services, prices, working hours. Then goes a possibility to make an order, delivery options, sales, and perks, etc. In fact, here there are no unnecessary points. Every little thing matters.

As digital assistants tend to take information from search features, ensure to provide for appearing in the Knowledge Panel. Whether Google decides to show your Knowledge panel or not strongly depends on how much information your Google My Business account contains. So, seriously, you need to fill in every section of your listing. Plus, working on your GMB account is a constant process. Update it, add photos, videos — keep it alive!

There are some features that appear on the panel and may work well for voice search, as they can give direct answers to user’s questions:

• Questions and answers. An often overlooked feature that, however, creates a great user experience and plays in favor of brand reputation. It immediately answers people’s needs and provides them with the fullest information about your business and services right from the SERP. Bear in mind that anyone may answer questions on the Knowledge Panel, so please, please answer all the questions you receive yourself. Thus you’ll ensure your Knowledge Panel contains the right information.
• Google posts. Google posts are short posts you may write in order to acquaint people with your promotions, new products, services — to share news, to be short. They are kept alive for about a week. However, they are a powerful instrument of keeping people informed and they benefit your SEO.

Schema markup. Though there’s no strong correlation observed between voice search and structured data markup, it’s still better to use it than not to use. Here’s why.

Structured data (schema) markup is a great possibility to help machines understand what your text or page is about. This makes robots see what a certain set of letters and symbols actually means. Thus it’s easier for them to define the page’s relevance to a given query (which is good for SEO).

Google’s recently announced a new markup type called Speakable, designed specially to point out parts of the page that can be read by voice assistants. It’s still in beta now, however, you can test it and see how it works. There are some requirements you need to meet before you’ll be able to be eligible for using this very tag on your page.

You need to be included in Google News. You may apply for inclusion via Publisher Center. But first, you need to bring your content in line with Google News content policies.

Page speed and security. Last but not least, let’s talk about your site speed. It affects user experience, rankings and, consequently, it has a huge impact on voice search results. The average voice search result page loads in 4.6 seconds, which means you really need to put your effort into improving your site speed. PageSpeed Insights by Google can measure your speed and point to the bottlenecks you have.

Another interesting observation shows that the majority of voice search results are pages with HTTPS. Security is critical, so in fact, HTTPS migration is something that should have been done yesterday, no matter you strive for voice search presence or not.

Conclusion

I believe, voice search is not another buzzword to be eventually forgotten. It’s a trend that is going to develop and gain a bigger search market share. The future when we are able to speak to virtual assistants naturally, as we do it speaking to other people, has already come. Moreover, apart from getting information from them, we already can delegate certain activities to them. And there’s plenty of room for development in this area, as technology today proves to evolutionize fast. This definitely disrupts the existing concept of SEO, challenging marketers with a sophisticated task of complying with big changes in the field where lately everything’s seemed more or less stable.