Here are the Mary Meeker slides that announce the start of the Voice First revolution.

The Voice First Revolution Started And Few Noticed

Mary Meeker is a legend in Silicon Valley. She has such a well deserved designation as being accurate about the direction of future events I estimate billion of dollars per year move in the direction of her annual report and slideshow. It all started with The Internet Report in 1995 where she correctly foresaw the rise of the World Wide Web and Internet Commerce.

Many who have followed my work know that I have been working around voice since I produced the first Voice board for a Commodore 64 and later wrote a Voice Manifesto in 1989 that has become over 800 pages. I presented a very small sliver here:

On June 1st, 2016, Mary gave the 2016 Internet Trends Report and at 213 slides, it is comprehensive and candid about all elements of the many satellites orbiting the internet. The slide show captivated the annual conference hosted by Kara Swisher and Walt Mossberg now called Code Conference by Recode [1], [2].

The Mary Meeker Voice First Slide Presentation Breakdown

We are at the dawn of the Apple World Wide Developer Confrence on May, 13th, 2016. It seems certain we will see a new version of Siri I have been calling Siri2. Between Alexa, Viv and the new Siri2 we are about to embark on a very accelerated rate of change. As a researcher that has been studying and building around voice since the 1980s I wanted to look at a few of Mary’s slides that relate to what I am calling the Voice First Revolution and present some of my own insights. I recently spoke to the slides listed below in the Around the Coin Podcast I get to the slides about half way through after a bit of a digression:


Indeed from the very first voice systems from Bell Laboratories in the 1950s up till just about 5 years ago, speaker independent voice recognition was inaccurate and error prone.

Humans have been trying to adapt to the limitations of the machines we create. Even the most advanced iPhone requires levels of learned skills we tend to take for granted and is usually forgotten.

It has been my thesis since I began my work that evolution has spent a few million years optimizing human speech to become a very accurate high data exchange system.

This is a critically important and often overlooked truism on Voice, word recognition has jumped from about 100,000 words in 2000 to 1 million words in 2010 to over 10 million words by the end of 2016. Along the way accuracy has moved past the 95 percentile.

The unprecedented growth of the admittedly rather limited first generation of smartphone Voice Systems still has pushed usage to over 65% in 2015 and by the end of 2016 that number will be over 75%, even if nothing changes with the existing systems.

The pioneer mass scale Voice First device is the Amazon Echo. The primary citations in this study do not really take into account much of the Echo use cases and relate more to Voice as an appendage to the operating system like Siri and Google related services. The use settings are quite interesting. I have used an Echo in the car, kitchen, bedroom, bathroom and office. Voice First devices define a physical space and clearly it starts in the home and car. But interestingly less useful on the go. I have many studies exploring the reasons why.

Clearly one of the fundamental reasons Amazon created Echo was to aid commerce. I call this Voice Commerce and there is absolutely no doubt that the cognitive load and mechanical load on us is significantly lower when we use our voice. This single element will not only define the Alexa platform but will also define Siri2, Viv and other systems. Voice Commerce replaces advertising.

The arc of growth for the Alexa platform is astounding. With just about 5% owning a device, we can see that a network effect will take hold very soon. I predict this will take place by the end of 2016 when about 15% of Amazon customers will own a device that has Alexa built in. Along the way the system will become an order of magnitude more powerful.

Mary estimates about 6 million Echo devices shipped thus far. We can see the growth compared to the early years of the iPhone. I agree that on the aggregate in 10 years all Voice First devices growth combined will exceed the growth rates of smartphones.


I spoke to much of the underpinning to Mary’s slides earlier this year with th the Tech.pinions posting I mentioned above [3]. We can see in the slides the foundations of an inflection point is taking place for Voice First with Echo and Alexa. It is important to note that with this point there are two divergent systems here:

  1. Echo- Voice First hardware
  2. Alexa- Voice First platform

To fully understand these points it is important to see that Alexa is a developer platform. Amazon has presented a new platform allowing developers to build Voice solutions using Alexa tools. For most of 2015 I would privately say: Alexa is to Voice First as AWS is to the internet. Alexa may very well become the AWS for Voice.

This is how it looks when a massive change is upon us. This is how it looks when even people that work in AI and Voice don’t get it. This is where the Voice First revolution officially started.


Some are just not ready for the revoution, and if this does not make sense, here is a simple version: