250 Billion and Counting: The New Rules of Search by Voice

It’s increasingly clear that AI-voice will soon be a primary interface to the internet.

 We’re on that path now.  Through voice assistants, we talk to the internet, and it talks back.  In time, the talk will evolve to the give-and-take of dialogue and dialogue will evolve to human-like conversation. 

 In time.

 Voice is now in its early days.  Similar to the early, browser-war days of the internet.   Access is limited.  Interoperability, nil.  Usage is cramped and constrained in walled gardens.  

 And some very big questions wait for answers.   

 One of the most important questions for enterprise users is that of voice-based search.

 For guidance, let’s turn to voicebot.ai’s superb recent study “Voice Assistant SEO Report for Brands,” which you can find here:  https://voicebot.ai/2019/07/09/new-data-on-voice-assistant-seo-is-a-wake-up-call-for-brands/.  

 (A strongly-held opinion:  if you’re working on voice for brands, or in marketing, this report is essential reading.   And if you’re working in voice, a subscription to the voicebot.ai work of Bret Kinsella and Ava Mutchler is a must-have.)

 Citing the voicebot.ai research:

How important is voice search?  Very.  Google itself reported that voice was the interface for 20 percent of its 2018 mobile search.  Serious calculations suggest that voice was close to 13 percent of Google’s 2018 total search volume.    One in eight, trending toward one in seven internet searches. 

 In total, some 250 billion voice-based searches. 

 Will the voice search number go up or go down?  Well, it’s more than likely that the trend line will continue to rise, given that there are now more than two billion devices worldwide that provide voice assistant access – and that we can expect some eight billion by 2023.

Who’s using it?   A January 2019 survey by Voicebot found that nearly 60 percent of US adults have used voice search, and 47 percent expect to increase usage this year. 

Which devices are used? Smartphones, for sure.  Used by more than 90 percent of all voice searchers.  And, smart speakers more and more – used by roughly 45 percent of all voice searchers.  Keep in mind that smart speakers will soon be available to one-third of all US adults.

 What are they asking?  Commerce-centric questions.

What are they asking?   The most common voice search topic is a fact-based question (asked by more than 60 percent of users), followed by a request for directions.  More than a third of users have asked about restaurants; nearly a quarter (23.5 percent) have researched a product or pricing before purchase; another 23 percent ask for recipes or cooking instructions. 

Does the search process work the same as for web-based search?  Well, not really.  First, keep in mind that internet search is increasingly balkanized.  It’s not just Google – especially in commerce industries.  Nearly half of product searches in the US now start on Amazon – it’s by far the top search engine for shopping.  YouTube (owned by Google), is the top search engine for video, and the number two worldwide for all searches.  Facebook users conduct more than two billion searches a day. 

But once you choose what platform to search, aren’t the results the same as web-based search?  Well, not really.  Screen- and voice-based search can be radically different.  On the screen, a user may see 10 or more search results.  In voice, however, “a single audible result” may be the  only decent user experience.  That’s right:  one.   

Winner take all?

 There are some options for additional information beyond the first results (depending upon platform), and smartphone responses can be more extensive than smart speaker (again, depending upon platform.)

 And what about the beneath-the-hood search technologies? Aren’t they the same? Well, not really. When it comes to search, the interface – with apologies to Marshall McLuhan – is the message.  Search experience and results can differ based on the interface.  According to voicebot.ai’s research, many voice assistants consult multiple data sources before tapping into Google’s or Bing’s knowledge graphs.  These include proprietary databases for “reserve terms” through which the voice assistant provider can control all messaging.   Ask Alexa about products, and you’ll most likely get information from Amazon.com.  Ask Siri about the “best tablet,” and you’ll learn about the iPad.   

 Voicebot’s summary:  “as voice assistant optimized content increases, the traditional web search knowledge graphs will be consulted less frequently.”

 Hmmm.

 What did we learn?  Three big things.  1) Voice search is big, and getting bigger.  2) It’s being used for a queries that directly impact commerce industries – from getting directions to researching products to getting recipes.   And: 3) proprietary voice platforms may have their own rules for search – ones a bit different from those on the touch-and-tap internet. 

 It’s time for the Open Voice Network.

 This is but one of the issues being explored by the Open Voice Network, a new industry association dedicated to artificial intelligence-enabled voice that is open: standards-based, interoperable, accessible (to all and through all), and secure. 

 For more information, visit www.openvoicenetwork.org, or find us on LinkedIn and Facebook.