In 2014, the use of voice search was statistically zero. Today it accounts for about 12%, and by 2020 that’s predicted to be around 30%.
What is voice search?
Voice search was originally little more than an alternative way of performing an internet search. Today, the game-changing nature of the technology is really starting to be understood. It utilizes AI and speech recognition software to allow users to enter search queries audibly instead of typing.
Voice search will fundamentally change search
People use voice search because their hands are occupied, because it’s quicker to speak than to type, or because it doesn’t require a keyboard or screen. But it’s about more than just convenience. People speak differently to the way they type, and they expect different responses. It’s critical that search engines have the capacity to understand the context and meaning of queries, beyond simple keywords.
Increasingly, people’s primary interaction with voice search is through a digital assistant, such as Amazon’s Alexa, which has the ability to take action as well as return information. This means the distinction between query and action becomes blurred, as it’s much quicker to ask Alexa to order a pizza than to ask her to find a takeaway, read aloud the menu, and phone it in.
This changes the dynamic of shopping itself. The majority of consumer purchases are “low-involvement”, that is, the specific brand of product we receive doesn’t actually matter too much. If we ask Alexa to simply “buy more toilet paper”, she may well will be making decisions about retailer, brand and product for us.
How results are generated
Generating results based on context, intent and history is clearly more complicated than traditional search. The engines need to understand if the query is a simple one that requires a simple answer (is it raining?), a complex question that needs detailed content (why does it rain?) or a request for action (order a raincoat). However, some queries will still be best answered with visuals.
At present there is no single approach to the issue of voice query/response. Microsoft’s Cortana uses Bing as its default search engine, while Google Assistant uses Google. Amazon is creating a different process which references “skills” – third party apps and websites – to provide the data it needs.
What does this mean for SEO?
The industry is yet to see any clear best practices for marketing to voice search. However, much of the existing guidance for text-based search can be extended to apply to voice.
Search engine specialists have been working on long-tail keywords and phrases for some time now, and this approach remains a priority. In addition, marketers must consider how a question or instruction might be worded differently when spoken, the potentially different context for the search, and how to phrase the response so that it’s clear when read aloud.
Content needs to be optimized for specific queries and query types. Websites must be fast, responsive to multiple screen types, and structured in a way that makes it easy for the search engines to understand the context around it, as well as what’s on the page. Schema markup will help search engines understand the meaning behind different pieces of information on your pages.
Voice activation in the workplace
In the workplace, voice search is likely to be tied to digital assistants. 19% of organizations are already using digital assistants, and 46% intend to implement them within the next few years. However, there are additional security and privacy issues for businesses.
Adapting security measures for voice-activated digital assistants
Digital assistants utilize machine learning to tailor and improve their responses. This means storing large amounts of data about user location, preferences, previous instructions and even passwords. Many voice-activated programs are switched on by use of a “hotword” such as “OK, Google”, which means they must be always listening, and there have been concerns around how much passive noise they record. Data may end up being stored at several locations – the local device, the digital assistant, and the search engine.
Some threats are posed by external forces with malicious intent, like hackers. However there is also the threat of a digital assistant malfunctioning or misunderstanding an instruction. This might simply mean a hotel being booked in the wrong city, but in more sophisticated setups it could mean orders being placed incorrectly or information being shared with the wrong individuals.
To manage the risks, digital assistants need to be monitored just like employees. Their security privileges need to be defined in advance based on what they are required to do and what the risks are. Corporate data should be segmented depending on who needs access and how important it is, with the most valuable data ring-fenced separately. The location of microphones should be mapped, and microphone-free areas set aside for when sensitive information needs to be shared verbally.
How should businesses tackle voice search marketing?
Voice search is clearly going to be big, but there are still questions to be answered before businesses can develop clear strategies. For example, how will it be monetized? How much will it still rely on visual devices? When will cross-platform standards emerge?
Fortunately, a lot of the advice offered so far is simply good practice. Targeting long-tail conversational queries was already a requirement for search marketing, and ensuring company data is well-structured is beneficial for many reasons. By making sure this basic groundwork is underway, companies will be in a good position to adapt as further details are released.