Why Amazon will lose the battle for Voice Enabled Services and Devices
The Voice enabled devices is upon us, even though the interactions are still very clunky,my son’s first reaction to it after a few tries was to ask: “Alexa, tell me about Pokemon”. However the key word “tell me about” was not a trigger word, so essentially nothing happened.
It is important to distinguish between the Echo devices and the AVS (Alexa Voice Service). Essentially Amazon has opened up 2 market fronts: Offering Echo devices in various flavours, and offering the AVS to Software developers, enabling it to work with Apps, Websites and integrate into devices.
Since then Amazon are making general improvements every week to their talk interface, for instance adding new basic commands and native skills delivered from Alexa.
There are a lot of Amazon skills added to the system everyday. In Q2 2017 there were 15.000 skills added.
The Amazon Echo devices are nicely designed, and with the new products recently launched like the Echo Dot, Echo (2017), Echo Spot, Echo Connect and so on the possibilities are endless.
However the Alexa Voice Service might soon fall behind in the market, giving way for Google Assistant, the reason is simple: User Experience, Architecture/Data and Organization.
The perceived User Experience when asking trivia questions are very low compared to Siri, Cortana and Google Assistant. This hints at some general problems when expanding the vocabulary of the AVS service.
The Architecture of AVS is build using simple tables for the NLU translated text, in which specific textual intent translates to actions. This model seems to be inadequate as new speech patterns needs to be recognized and acted upon. Also it gives some specific issues when it comes to AVS Skills installed in the system: It is inherently a bad user experience for the user to say “Alexa, ask XYZ Skill to ..” every time the user needs to execute an action for a skill. Instead new skills installed in the systems should be an inherent part of the basic voice enabled user experience. Also the amount of user data available to Amazon are limited, so there is not much synergy effects with the rest of the Amazon Ecosystem.
The Organization of Amazon are build so that users in US belongs to the US based stores, European users to the specific countries based stores etc. This is a huge problem since all countries that have English as second or third language cannot effectively use the AVS or Echo. This effectively cuts out roughly 500.000.000 potential users, as well as countries where english is further down the list like China. Also this linking of users to specific sites opens up issues when users use AVS across several countries. This tendency is seen for the Sonos One product which is launched in many different countries, but in fact does not work for a very long list of countries.
At launch time Sonos One started to flag Google Assistant support to alleviate this problem, hence another market lost for Amazon.
Finally the 2 market fronts (Echo Devices + Alexa Service), means that there is also internal competition within, examining the licensing agreements for the Alexa Service have several indications that the Echo devices does not use the standard libraries that a competing device build using Alexa service have. This in fact gives an unfair market advantage to Echo devices build by Amazon, as opposed to devices build to compete with Echo.