The Rise of Voice Search and Conversational Language in 2018
When the Amazon Echo was first released to the American public in 2015, people were quick to judge the smart speaker and Alexa by extension, as a novelty. The view was that even though asking Alexa about the weather would make for a good party trick, it wasn’t something that could be relevant. Now though, with more companies (Google with their Assistant and Microsoft’s Cortana) piling resources on A.I. and natural language processing, SEO services and companies have finally taken notice of this rising trend.
The continued rise of voice search
The American analytics company comScore predicted that by 2020, 50% of all searches will be done in voice. A report from Voicebot earlier this year unearthed some illuminating results, with more than half of smart speaker owners in the United States interacting with their devices at least once per day with an average use of 2.79 per day. This is limited to the US, where a fifth of the total population is reported to own a smart speaker.
As companies from other countries adapt to this trend, or alternatively as Google and Amazon fit their devices with more language capabilities, you can expect voice assistants to become more widespread across the world. Japan and tech-crazed China is now firmly on in the trend and now that competition has heated up, it’s only a matter of time until companies from tries to break into the emerging markets.
The thing with voice assistants is that they are artificial intelligence. Just like how the human brain develops from experience, A.I. develops their algorithm by taking in more data, a method referred to as machine learning. The easiest way for companies to gain these data is by putting their devices on more customers and asking us to interact with them on a daily basis. It is one of the main reasons why voice search will be the focus as we head into the future, with companies developing their natural language processing technology on public usage.
The difference between voice search and typed search
Typing a question into Google is different than directly asking Google Assistant. A study from 2015 showed for example that written and spoken language can exist separately in the brain. If you’ve been wondering why you can impeccably flirt with your matches on Tinder only for your brain to turn into mush when you go on an actual date, this might explain why.
Still using Tinder as an allegory, the pairing of search queries and results could be explained in terms of Tinder users, except with Google acting as matchmaker. If we assume that the actual language being used for typed and voice search is different, wouldn’t it also naturally mean that Google would match those queries with different results? Taking this idea to its natural outcome, wouldn’t it also mean that optimizing for voice search involves different methods when compared to conventional search?
Conversational writing in voice search
Obviously, given that voice search is still relatively new, there’s no concrete answers to be had with regards to the above questions but you have to admit that the assumption is a reasonable one, which leads us to this point; conversational writing. Voice search is inherently informal. It involves language you would normally use over the course of your day when you’re ordering coffee or during general chit-chat with your colleague.
As such, it would be a good idea for you to optimize for keywords that are more natural. For example, if you’re looking for wireless earphones because your new phone has no headphone jack available, think of the words you’d use when you’re typing your question into Google and one you’d use when asking for opinion from a colleague.
In Google, you could simply type ‘good wireless earphones’ and the results would be in line with what you’re looking for. Now, technically, you can use those exact words with Google Assistant and be presented with decent results but saying those words out loud would be awkward. The language you’d use will eventually be more natural, along the lines of ‘what are the best wireless headphones available today?’
Now, one other point of focus is that in voice search, it’s not just the search query that’s being spoken. The result is actually being spoken as well. Both Google and Alexa tend to answer questions using the most concise answer possible by sourcing from other sources, like Wikipedia. If you ask these devices who Lleyton Hewitt is, they’d answer by using the first line from Hewitt’s entry in Wikipedia. Anything more specific and they’d try to sum up what they find in those sources.
Google has actually chimed in with their own take on this, saying that “when a displayed answer is too long, users can quickly scan it visually and locate the relevant information. For voice answers, that is not possible. It is much more important to ensure that we provide a helpful amount of information, hopefully not too much or too little.” A study from Google Home search results for example, found that the average search result is only 29 words in length.
One other point to consider when it comes to writing is that you have to appeal to the lowest common denominator. If you’re looking to put your Literature degree to good use in your writing, you need to take a rain check on that idea. From the same study linked above, the average Google voice search result is written at 9th grade level. Accessibility is key in optimizing for voice search.