View profile

Hearing Voices - Issue #13, or "Alexa Skills Ecosystem"

"There is a feeling, when you listen to radio, that it's one person, and they're talking to you, and

Hearing Voices

February 17 · Issue #13 · View online
This week's news about voice computing apps, delivered directly to your inbox.

“There is a feeling, when you listen to radio, that it’s one person, and they’re talking to you, and you really feel their presence as one person”
- Ira Glass
“There is a feeling, when this newsletter is forwarded to you, that you should subscribe. Do that here:
- Me

A bit of a different format this week. We did an analysis of the Alexa Skills ecosystem, which I’ll share. But there was also a ton of news so including that in shorter form here:
To Try:
Mylestone - access your personal memories through Alexa
Snowroll - a new photo sharing app that incorporates voice and photo
Overdrive - found this through my library of all places. content is free because, library.
Starbucks Reorder skill seems awesome. Thanks @brianroemmeleRandom Conversations - Start conversations with random people from Google Home. thanks @cemkozinoglu
> Going to SXSW? Anchor is hosting an Audio lounge on 3/13. Password is AUDIOFIRST thanks @mayafish for sharing

And now for the Alexa Skills ecosystem analysis…
So you want to build a voice app. Will anyone be able to find it?
One year after Amazon opened up Alexa to third-party skills, there are 1/10th the number of skills compared to iPhone apps built the first year of the App Store, with close to the same install base.
When people first got their iPhones, their natural question was “what apps should I get?” Nine years after Apple launched the App Store, people are asking the reverse – what apps can I delete?
If you’re considering building a voice app, before you write a single line of code, one question worth answering is: will enough people be able to find out about it to make it worthwhile to build?
Two major factors that impact discoverability are Channels – the number of places where the app can be discovered, and Competition – the number of other apps that are also trying to get attention.
Right now for those building for the Amazon Echo, Google Home, and other voice-first platforms, there is some surprising data that shows it might be a great time to start building.
The App Store on the iPhone sat right on your home screen, native to the device where you used the apps themselves. Amazon Alexa has a website which features apps, a dedicated email which sometimes features skills (but often focuses on holiday and other time-related easter eggs). However the only native skill discovery, “Alexa what are new skills” responds with three features skills to try..
It’s worth comparing that discovery experience with data from the app store in 2014, a mid-point or perhaps even late in terms of all saturation. In 2014, nearly 70% of iPhone apps were discovered directly within the App Store, either through search (47%), topping a category chart (10%), being selected for the featured section by Apple (9%), or being featured in a specific category (2%). It’s hard to see the Amazon Skills Store getting as much traffic, as it’s not native to the Echo (i.e., it lives on the web). The alternative channels to discover iPhone apps in 2014 were: Friends/family telling you (15% for apps), reading about it on the web (11%), searching the web (2%), and unpaid discovery through social media (1%).
So on one hand, skills are currently limited to about ⅓ of the volume of distribution as compared to iPhone apps. This makes things difficult for developers.
What’s so interesting about new voice products is how large the install base has become relative to the number of apps (“skills” in Amazon parlance and “actions” on Google Home).
When the iPhone app store launched in Q2 2008, there were ~5.5MM iPhones in market. A year later, there were ~14.5MM iPhones in market and ~65,000 apps.
Similarly to the iPhone launch, Amazon’s Echo launch included no third party skills. A bit less than one year in, the skill store opened and at that point there were ~2.4MM echos in market. One year from the skill store launch, there were ~6,000 skills and an ~11MM install base of Echos.
A note about our methodology on calculation of echo units, since this is not data Amazon releases: The most accurate data found comes from Business Insider’s analysis (here) which shows annual sales for 2015 & 2016, excluding the 2016 holiday season. Voicelabs data (here) estimates Q4 2016 holiday shopping resulting in approximately 4x increase in Echo sales over Q1 of the same year (approximately 40% of the year’s sales). We applied this ratio to the 2016 data, which has been reported to be 5.2m units sold in Q1-Q3 of 2016.
One way to think about this is the relationship between the number of devices in market and how crowded the market is for third party skills. The greater the install base, the easier it is to get developers to build, but as we’ve now found with the app store, at some point there is an oversaturation of skills or apps, and it becomes extremely difficult for consumers to discover your skill.
So what is the current state of affairs in the Amazon Alexa ecosystem versus when the iPhone and app store launched? The install base of the Echo is much larger relative to the number of skills developed than was the install base of the iPhone to its App Store at the same time.
One year into the launch of the App Store, there were 222 iPhones for every app built for the platform, meaning that there were lots of users looking for relatively few apps. This is good for app developers and generally for people who want to find new apps to use. The Amazon Echo ecosystem is even more appealing for developers. One year in – the same time period we measured for the iPhone – there are 1,824 echo devices in market for every 1 skill. This means less “app fatigue” for Echo users, who are still eager to find new skills to install, and means that it may be easier for developers to get their skills discovered.
This is true in absolute terms as well. One year in, there were 2x more iPhones in market than there are Echo devices during a one year period. However, there were already 10x more apps in the app store.
While discovery of new voice products is still nascent in terms of the number of levers (channels) that a market has open to builders, presence of relatively few skills offers an opportunity for new voice developers to build for a meaningful install base, but still cut through the noise (or lack thereof).

Thanks for reading through this week’s newsletter and analysis.  If you liked it, please share it. Here’s a sample tweet:
Check out @matthartman’s analysis of the Alexa Skills ecosystem in his weekly newsletter
Have a great weekend

if this email was forwarded to you, you can subscribe at
you can also follow me on twitter @matthartman 
Did you enjoy this issue?
If you don't want these updates anymore, please unsubscribe here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue