Why type when you can talk?
Although voice recognition software can still be a bit hit-and-miss, they have improved massively and are now more widely used. The appeal is understandable. It allows people to perform searches much more quickly and it is closer to a human interaction which, as a species, we are wired to prefer.
Siri, Cortana, Alexa and other exotically named personal assistants are ushering in a new era, that of voice queries. But it begs the question: how are they different from typed queries and what is the impact on traditional keyword-based SEO?
The difference between text queries and voice queries
Google’s Hummingbird update paved the way for voice queries by introducing semantic context – i.e. the meaning of a word is connected to its context – which has, since then, become a key element in search advertising and content marketing.
Search queries can be divided into three main categories: navigational, informational and transactional.
A navigational query is about being taken to a specific website; an informational query is a request for information – e.g. ‘How old is the Earth?’ – and a transactional query is about completing specific actions such as a purchase, a booking, or checking your bank account for example.
A keyword-based search is relatively simple from a computer’s point of view: it just has to match specific words and phrases to identical content on the web. But when it comes to voice queries, context needs to be taken into account to understand the intent behind it. Think about the question ‘Do you think it is going to rain?’ A literal answer could be ‘yes’, ‘no’ or ‘I don’t know’. But the real question behind it could be ‘Will I need an umbrella?’ which would require a completely different answer. It may seem obvious to us, but when you are a machine, this is an incredible complex process which requires inference.
The three characteristics of voice search
• Successful voice search requires at least 3 words.
We generally advise optisiming a website with keywords made of at least two words as single words tend to drive traffic that is too generic. However, analysis of voice queries has revealed that, on average, the ones that were getting the greatest number of impressions and clicks counted at least 3 words.
This is perhaps not so surprising considering that natural speech is usually longer, and it also follows the general trend towards long tail keywords. However, it is important to remember that we have just started using voice queries and that how we use them may evolve.
• Voice search is dominated by question phrases.
Text queries have also seen, to some extent, a greater use of question queries, but data extracted from Google shows that 61% of voice queries are made in the form of a question, this, again, reflecting natural speech patterns where human beings are more likely to ask a question to get information.
• Emphasis on geo-location
A few years back, a query such as ‘Where is the nearest place of interest?’ wouldn’t have returned anything terribly useful. This has all changed. With the number of searches made from mobile devices catching up with desktop queries, Google has recognised that geographically based results are most relevant.
If you now ask for activities near where you are, Google will show you a map of your location, where you are and a list of attractions nearby. And should you need to contact them, you just have to ask your personal digital assistant which will dial the number for you.
From finding information to performing an action, voice query has considerably shortened the sequence and the time it would take to complete your intended goal. In our impatient societies, no wonder it is so successful!
Advantages and challenges of voices queries
Local businesses, especially retail ones, can greatly benefit from voice queries as those tend to be about finding information about business in the immediate vicinity. When you are optimising your website, think nearby landmarks and attractions. For example, if you are a bookseller near a cathedral, optimise around terms connected to the cathedral. Where you might have before targeted keywords such as ‘fiction bookshop’, you could now optimise instead for ‘fiction bookshop near cathedral X’ , which will tap into searches for that landmark too.
However, don’t throw the baby with the bath water and completely discard optimising for text queries. There are still technical challenges with voice recognition. For example, it is hard to predict how it handles mispronunciation of words. There are huge variations among native English speakers – think of our Commonwealth friends – and there can even be regional accents within the same nation! And if you are a restaurant in a touristic area, visitors’ digital assistants will have to contend with foreign pronunciations.
But even if potential customers pronounce your brand’s name just right, there are still chances that Siri, Cortana, etc,… won’t recognise it. Google, for example, has improved its voice recognition rate from 80% to 90%, but it still means that 10% of voice queries are unsuccessful. Likewise, if your business name is an acronym, contains foreign words or is a spin off a real word, Google will have trouble recognising it. Which means that if you haven’t optimised for text queries as much as voice queries, your business wouldn’t get picked up at all.
Last, there are searches that we are happy to type up but would be too embarrased to speak out loud, even alone in a room! So there is definitely a case for continuing to optimise for text queries.
As voice recognition technology is still quite young, there is no doubt that, as it improves, it will continue to affect how we live and search for information, but it is definitely one trend that is here to stay.
If you are looking for expert digital marketing consultants in Auckland, contact This Side up via our form or on 09 360 2299. We will be happy to discuss how we can improve your digital performance.