randyseoppc
Junior Member
- Feb 24, 2016
- 174
- 61
How do you optimize your website content for voice search?
hey man!!There's some great academic papers around this!
Have a search through Google Scholar to pick them out.
As a starting point, you need to know how the AI is going to interpret your query. Siri and Google Assistant work slightly differently in context, and remove certain words from queries.
You used to also need to know what search engine the voice search is using (The good news is Siri and Google Assistant both use Google, Alexa obvs uses Amazon.)
Next, you'll probably notice that your query changes from when it's typed. People tend to ask longer search queries with more filler words in voice. On average, people will say 4.2 words in a search query, but they'll only type 2.2. Plus 35% of people using voice search will say 5 words or more!
They might also describe something in different words - so they use more conversational or slang words when speaking.
They also tend to ask more who/what/why/when/where queries. And more queries with verbs in them.
Finally, they have very different search intent. When people are using voice search, then tend to have more of an informational intent and a lower commercial intent. So they use voice search to look for information relating to songs, movies, weather, directions, distance, recipes, and other things that they naturally glance at on their phone. The problem with those is that they usually trigger specific behaviours inside the phone itself that don't go through to Google, or are rarely clicked.
The consequence of all the above is that it means voice search skews heavily towards websites like Wikipedia, IMDB, and other big content portals.
If you had to optimise a website to attract as much traffic as possible from voice searches, you'd probably create a site with a lot of trivia style articles like "What year did the Titanic sink?" or "How do you make gravy?" or "What's the best way to get from Times Square to Newark Airport?"
It might also be good for things like "The best seafood restaurant [near me]" in platforms like Google/Apple Maps.
Otherwise you'd just optimise for keywords that make sense for non-voice searches and you'd likely get to the top of rankings on the rare occasion that someone searched for info on that.
P.S. The academics also found an oddly high % of voice searches were for porn. Surprising!