Voice search is pretty exciting territory at the moment. Those up-to-date in the digital marketing sphere may have been discussing optimising for long-tail keywords for a while now, both in web content and PPC campaigns. You may have noticed more queries in your Search Terms report in AdWords beginning ‘Cortana’ or ‘Ok Google’. And yes, it’s exciting that you can now buy home assistants that can play your ‘rainy day playlist’ (what, you mean you don’t have one?!) – Zuckerberg’s Jarvis makes it feel like Kubrick’s Space Odyssey is finally here, albeit 16 years late. But these are just baby steps.
It’s when you start thinking about how far voice search could go that it begins to truly blow your mind, which we’ll get to shortly, but these are the basics of voice search:
Voice search is the ability to enter a query by speaking, then the search is conducted because the assistant is able to recognise the query with voice recognition. I’m sure you’ve heard of Alexa, Siri, Cortana and Google Now, which can help organise your life as they are integrated with other apps on your device. So, for instance, you could ask Cortana to remind you to email Distinction to help with a digital transformation tomorrow morning!
These voice search functions must find their knowledge from somewhere. Google Now is powered by Google’s Knowledge Graph, whereas Cortana and Siri are powered by Bing. Alexa is slightly different as it functions solely as a personal assistant service and is integrated with other services, such as AccuWeather in order to give you real-time weather information.
The exciting part
These little bots aren’t just a vocal connector between the user and the search engine to improve the user journey. Do you really think that the sweet nothings you whisper to your voice assistant are futile information, and don't go anywhere? These guys are smart; the reason that they're able to understand what you're asking them in the first place and provide you with the right answer is due to natural language processing, which essentially refers to a computer's ability to understand voice commands and is a component of artificial intelligence.
In addition to this, the machines are learning; every search query conducted is indexed and user behaviour from that is measured to detect whether or not the search engine is presenting the user with the results they wanted. Google Now, Cortana and Siri aren’t mindlessly answering our questions; they’re getting to know us.
More than that, if you continually use a voice assistant (which isn’t unlikely, as numbers are on the rise), it’ll get to know the kinds of things that you’re interested in based on your previous searches. For instance, based on my search history on desktop and mobile, if I begin typing in the word 'hummingbird' into Google, the top suggested search isn't 'hummingbird' or 'hummingbird cars' or 'hummingbird bakery', as it would be for someone who isn't a digital marketer; it's about a Google algorithm update and this is what it displays because it knows that this is more semantically related to my interests (which are enthralling, I know).
This is machine learning; in relation specifically to Google, that's RankBrain at work, and it knows a little bit more about me with every search query I conduct in order to present me with the most relevant search results based on my interests; as shown with this example of 'Pictures of Mario'. The same principle will apply as you use a voice assistant more often and the technology advances; as the assistant will be gathering more data on our behaviour, the interactions will become more human and tailored to you.
What this means for SEO
So how does this affect marketers? We've been told to optimise for long-tail keywords for a while now and this is good practice in general, as long-tail keywords are more specific and therefore less competitive. But now it's important to think about the differences between the language used for voice search and traditional search. Over 20% of voice searches are location-focused, so this is great news for local businesses. Ensure that you're using adjectives that people may use when searching for you; compare 'bar in Nottingham' to 'intimate cocktail bar in Nottingham city centre'. Also make sure that you're using keywords that are specific to the area to optimise for local SEO, such as 'near the castle'.
Also, use friendlier and more personable keywords, such as 'the best cocktail bar serving delicious food until late'. Searchers are more likely to use words like this when conducting voice searches, because they're more conversational and they don't have to make the effort to type. Finally, whereas many traditional searches may not be cohesive, such as 'cocktail bar notts', a voice search is more likely to be a full sentence, like 'Where is the best cocktail bar in Nottingham city centre?' - optimise for questions.
This is only the beginning. Voice search, personal assistants and AI present other exciting opportunities for marketers. One of these is chatbots.
So, you can ask Siri for the latest news on Trump and she can show you the top results for this according to Bing, but that’s not good enough! We want a detailed article right away, we don’t want to have to click on it – too much effort. What would be perfect for that? A chatbot that can deliver the information directly to you, of course, like this one from the Guardian, which works within Facebook Messenger. And they’re not just there to notify you of news; they can be your personal shopper, a financial consultant, or be there so that you can order your morning coffee for you to pick up on the way into work. Imagine if this kind of interaction were possible but verbally - not on a chat app.
Problems with chatbots
The above is pretty cool, but chatbots are new kids on the block and they're having a lot of teething problems. This example from the ASOS Facebook page shows what happens when you let a bot take over your customer service. It doesn’t seem as though this bot underwent much testing (unless letting it loose on the Facebook page was its testing), and it didn’t replace human interaction well enough – whereas voice assistants are getting there.
Many say that chatbots are supposed to be replacing apps, which makes sense if they’re integrated within another app such as Whatsapp or Facebook Messenger. Lots of big companies are jumping on the chatbot bandwagon, but most of the time the result is not much better than a gimmick (such as Starbucks' 'The Real PSL'). I see one big problem with everyone launching chatbots; if every big brand decides to do it, then more or less the exact same experience of moving between apps is simulated, just in a different platform. It doesn’t make our lives any easier.
The answer to the above is integration. Imagine a better version of Siri or Cortana, which has the capability to do what a chatbot does, and tell Starbucks to have your coffee ready to pick up at 8:45. It would take a digital assistant, which can book meetings for you and present you with search results, to the next level; it would fulfil tasks that used to take minutes out of your day.
This is something that has been discussed for a long time, and now all of the voice services have developer kits which mean that you can create apps to work with voice assistants. At the moment, however, this is very pick-and-mix and certain apps will work with certain voice assistants, such as Wunderlist integration with Google Now. In addition to this, in many cases, all that the voice assistant can do is launch a certain app for you, and the rest of it is up to you. But that's not what I want. I want to be able to control any app from my phone using my voice, without ever leaving the same platform in order to have a completely cohesive experience. Integration to the max, please.
A move towards closing the gap between chatbots and digital voice assistants is the future, to create a fully-integrated, fully-functional bot that is powered by voice and can do just about anything (…anything that an app could do, anyway). But wait, aren't I forgetting something? The small matter of AI. This bot will know you. Imagine you use it for more or less anything – online shopping, fitness tracking, texting your friends and more.
Now imagine that you’ve worked late one night. It knows you haven’t done a Tesco shop in a while due to geotracking and access to your browser history, and it knows you don’t have any dinner plans because you haven’t booked a restaurant or texted any of your friends. You’re walking home and your bot says, ‘The fridge is looking pretty empty. Let’s order a pizza. Same as last time?’. All you would have to do is reply in the affirmative and your dinner will be on the way.
The exciting future of digital assistants is that they will become predictive and proactive, doing things for you without being asked so that we no longer have to sweat the small stuff. Scary, right?