keywords-nlp

Voice Search Engine Optimization and the New Rules of Discovery – Part 2

Voice search is growing rapidly among consumers and many of the old rules of search engine optimization do not translate well into voice-first interactions. Yes, there is the standard voice search with a text-based response that mirrors text input. This is particularly common on mobile. However, Part 1 of this article revealed that voice assistants are starting to respond to more queries with vocalized responses. This means that there is a new prioritization of responses for voice-based queries and the results list is limited to 1-3 items. Voice SEO will be significantly different than the desktop and mobile SEO we have all learned over the past decade.

These changes are important for brand marketers. They are also important for anyone building a voice application that hopes to be discovered. Everyone needs to know how to ensure their content or voice app will rank highly for voice search results. Last week, I discussed how featured snippets and deep-link voice application discovery are important interim strategies. There are four more factors that marketers and voice app developers also need to understand about the new discipline of Voice SEO.

  1. Bing will grow in importance
  2. Voice assistants move us past keywords
  3. Voice assistants are the new search boxes
  4. Search is about choice, voice assistants are about agency

Bing vs. Google

How is your ranking on Bing? Many marketers know that Google commands about 63% of search on desktop and over 79% of all search queries when mobile is included. Bing is second, but a significant distance behind. Few marketers have sufficient budget or attention to run SEO programs for both search engines. As a result, most people focus their traditional SEO efforts on Google and hope for the best with Bing. Oftentimes, a good Google result also translates into a good Bing result. But, not always. This is going to be more important as voice assistants become the new primary input for search.

Why? Well, all of Google’s rivals have lived with the company’s de facto search monopoly for a decade. Not wanting to give Google even more advantage, the search engine of choice for voice assistants is Bing. If you want to be discoverable as the top voice search response, get to know Bing.

Source: Activate Presentation, WSJD Live

Voice Assistants Move Us Past Keywords

Not only is the frequency of voice search heading toward 50% and will Bing grow in importance, the search interface is also changing. More precisely, what is behind the voice search is evolving. Everyone is familiar with keyword based search. The state-of-the-art for voice search was a speech-to-text (STT) conversion based on automated speech recognition (ASR). That was then followed by standard keyword based search. The process was the same as text-based search with the added step of ASR and STT up front. This was not natural language processing (NLP) nor did it include natural language understanding (NLU) with an attempt to determine context and meaning. Siri search and clicking the microphone button on Google originally followed this process ASR and STT process.

What has changed? The voice assistant as a search agent goes beyond ASR and applies an NLP model. The seeds of this shift can be seen in Google’s change to Chrome. Results were influenced by your search history and other activities that Google tracked across its properties. The NLP was rudimentary and usually applied to text, but it started the process of an intelligent search agent that considered more than the textual input of the moment.

Google Assistant, Cortana, Alexa, Hound and even Siri now go further in attempting to divine user intent based on context. That context might be what you have done in the past, what you were doing recently, what is on your calendar next and so forth. And, your context will be persistent. The voice assistant will maintain awareness of past interactions.

This is about more than keywords. It means that search results will be tailored to what is said combined with context. We see some of this taking place now, but will see it more often as the voice assistants learn about users and the NLU gets stronger. What does this mean for marketers and voice app discovery? More work. You will need to go beyond keywords and consider the context and scenarios that are common to users you want to reach. You will also need to build up digital relationships with customers that can serve as signals to put your results first when related requests arise.

Voice Assistants Are the New Search Box

The adoption of voice assistants over traditional voice search using text input is subtle but occurring right in front of us. The fastest way to search is not to touch a smartphone and open an app. It is far more convenient to say, “Hey Siri,” “Ok Google,” “Cortana,” or “Alexa,” than to thumb-type our queries. Mobile devices are displacing basic voice search with voice assistants as the primary interface. Microsoft even introduces users to its new tablets by essentially forcing them to set-up their devices using Cortana.

Image Credit: Android Central

Smart speakers that are rapidly populating our homes and offices default to voice assistants. Amazon Echo only provides access to a smart assistant. The same applies to Google Home. There is close to zero cognitive load or physical requirement to ask for information by voice. When we do that, we are increasingly forced into using a voice assistant. That means the future of search will be heavily influenced by user context and the biases of voice assistants.  This is particularly important now that we know voice assistants will often return only one or two results to a vocalized query. But, wait! There’s more.

Search is About Choice, Voice Assistants Are About Agency

The emerging importance of Bing is worth noting while the voice assistant as a new information intermediary should materially change our expectations about how search works. The remaining impact is even more profound. Voice assistants will have agency. They will do things for us. Search does not have agency. It is a response to a request. You then must decide which of the responses to delve into further to answer your question or fulfill your intent. Even if it only provides you with a single result, the search output is limiting your information but isn’t really making a decision for you.

By contrast, voice assistants can decide for you what response it best based not on a knowledge graph alone, but instead based on context. The assistants can also take that information and complete tasks for you. John Done is a simple example of this for telephone activities. A more complex example is x.ai for meeting scheduling. Google Assistant, Alexa, Cortana and Siri will go well beyond these use cases very soon. They will suggest information to you, respond to questions from other people, schedule activities and make purchases on your behalf. How do you help your brand or voice app become discoverable when search may be conducted by voice assistant agents? This is yet another scenario to prepare to address.

Voice Search Engine Optimization is a New Beginning

It is easy to think about voice as simply a new user interface. If that were true, the only material changes to consider would be the heightened importance of being the top search result, building voice apps with deep link discovery phrases and developing an appreciation of Bing’s growing influence. However, the new voice interface is obscuring an even more profound shift in computing. Artificial intelligence manifested through voice assistants is changing the way information is gathered for and presented to consumers. That means Voice SEO will increasingly require understanding intent of both consumers and the voice assistant agents that work for them.

Voice SEO and the New Rules of Discovery

42 Percent of US Smartphone Owners Use AI Personal Assistant Monthly

Introducing The Voicebot Podcast Episode 1 with SoundHound’s Katie McMahon