Virtual assistants and smart gadgets have generated a voice search. Nowadays, this trend is developing at a rapid pace. Already today more than 60% of the US population looks for example, for custom truck parts, with the help of voice search with Google Home and Amazon Echo.It’s astonishing that the fact that voice search is the least studied area in SEO optimization of sites. To date, more than 50% of requests to search engines come from mobile devices. Side by side, it is 20% of these requests that make up speech recognition. From this, we can conclude that at least 10% of all queries to the Search Engines are just speech requests.

Based on the rapidly developing trend in the development of mobile platforms in three years, more than half of the requests are asked precisely by voice. It is necessary to strive to understand what factors influence the outcomes of voice scan and how to dominate the results of speech recognition.Speech recognition is vital for several reasons. Apart from the direct conversion and increased profits, issuance produces a single result, including a voice-activated device generator, is an excellent way to increase brand awareness. That is, the dominance in the outcomes of speech recognition is a significant contribution to the reputation component of the site.Optimizing the site for voice research is an undoubted trend in 2018. Leading global optimizers are currently conducting research, trying to identify the most important ranking factors in the speech search.Let’s look at the factors that influence the ranking in the voice search.Consequently, on a sample of 10,000 inquiries, the researchers identified such regularities correlating with each other:1. The site should be well optimized for working with mobile devices2. The speed of the site affects the ranking. The average download speed page of the results of speech recognition is -4,6s., Which is for 52% more than average results on the Internet.3. The HTTPS protocol is also relevant, about 70% of websites from the results of speech recognition used this protocol.4. Schema showed a weak correlation with the outcomes of speech recognition. Only 36% of the results of the voice request are used. On the other hand, on average on the Internet as a whole, the average utilization of Schema is 31.3%. Thus, there is a mutual influence, but it is insignificant.5. The authority of the domain and the result of issuing speech recognition correlate very well. Specialists indicate that in the sample of 10,000 results, the average DR from Ahrefs (Domain Rating) is 76.8. This is a direct analog of PR, despite the fact that its stupidity one can understand that Google still loves a large number of links.6. There is a small correlation with SMM promotion. On average, the content from which the result of voice search is produced has1200 a Facebook ball and 44 tweets. Don’t forget that the sample conducted among 10 thousand results of speech recognition.7. Content is written in a light, precise language. On average, by the American method, the level (class) of content used to produce the result of the research is 9.8. There is no correlation of the query in the content with the optimization of the Title header. The exact query was present in the Title is disappearing a small number of times.9. The average volume of content – 2312 words (English). We can say that the results of research are generated from long posts, generously endowed with LSI keys.10. Content well ranked in research organic on desktop platforms is likely to be displayed in the results of voice search. Here the correlation is about 75%.11. Content used in the Featured Snippet (occupies the “zero” place in the issuance) is likely to be used to produce voice research results. Here the correlation is 41%.There is a significant correlation between the speeds of access to the content of the site, the issuance of voice search. Thus, between the samples of 10,000 results of the requests, the usual download speed of the 1st byte of TTFB is 0.54 seconds. At the same time, the average TTFB on the Internet is 2.1 seconds.TTFB is a relatively new SEO metric (2013), of paramount importance in SEO and, especially, in mobile search. At once, the metric characterizes a delay parameter in data transfer between a browser and the server, an indicator of a degree of loading of the server and an indicator of the speed of content generation by a site engine.It is much easier for the scanning robot to measure the loading time of the 1st byte than the entire page. Given that different browsers process incoming code at different speeds, there is an opinion that the point of view of Google the download time of the 1st byte is a more reliable indicator of the speed of the site than the full page load.Let’s return to voice search. The average download speed of the entire page as a result of the experiment was 4.6 seconds, while the average download speed of the page on the Internet as a whole is 8.8 seconds. Thus, it is quite apparent that fast websites predominate now.The simpler the content is, the more likely it is to appear in speech recognition results. Confirmed by the fact that:· In some cases, Google needs to read the answer in a voice· It needs to understand the response without any references, visual information· Try to look for some compound words or even better phrases – the search result is dramatically different from simple queries.The more straightforward, more precise the answer to the query on your page is, the more chance it is going to appear in the outcome of the voice search. Publishing simple easily understood content could help in SEO.To ensure that your site remains relevant for voice search, it is necessary to anticipate the intentions of users by optimizing the site for possible voice calls, which in their mass will be lower frequencies.