Reading Time: 3 minutes
How does Alexa knowing thousands of skills make Amazon Echo more valuable?
I fear that under the current infrastructure created by Amazon for ASK (Alexa Skill Kit) apps, the ballooning number of Skills will devalue the service. There are a number of fundamental reasons for this.
The Orphan Skills and Zombie Skills Quagmire
One of the largest issues is discovery, activation, quality and invocation. The model currently used by Amazon is based on the classic Apple iTunes App store model. This means one needs to:
- Search for Skills via an app or web store interface and read the basic descriptions.
- Anyone can make a Skill with little to no commitment. The entry barrier to create a Skill is about equal to the level of technical skill that was required to build a website in 2000.
- To activate the Skill, in most cases one only needs to perform a sort of install for the Skill to become available via a click in the app or website or a voice command.
- To use the new Skill one must remember the exact invocation phrase or phrases to successfully use the new Skill.
- There is no simple monetization for a Skill and therefore there is not an easy way to fund updates, upgrades and support in the future.
Quality Vs. Quantity
All of the points above suggest that the issues of quantity over quality will become almost insurmountable as the Skills reach past 25,000 on to 100,000. Data suggest that even with failed discovery and invocation just 31% of Skills have a consumer review.
The 7000+ Skills include many orphan Skills and zombie Skills that not only impact discovery and surfacing, the quality also weigh down users to become motivated to try new Skills.
As it stands today unless the entire edifice of how discovery, activation, quality, invocation and monetization is solved the number of Skills (or apps on other systems) pose a huge barrier.
I have researched and studied these Voice First issues since the 1980s. Part of my Voice First manifesto identified this quagmire and present dozens of solutions to solve every one of the issues. As time moves on it will become orders of magnitude more complex to solve this quagmire not only on the Alexa platform, but also on the Google, Cortana, Watson and Siri platforms.
For many reasons I will not go into detailed solutions here, but I can say that using an app or website to discover and activate a Voice First Skill or app is foolish and quite short sighted. Additionally not establishing a number of useful and adaptable monetization systems, developers will have little motivation to build the quality services and solutions over time.
We Can Fix This
Theses are all solvable issues and I will reiterate it presents a Google-class opportunity not to just one company but perhaps a multitude. Over the next 10 years as 50% of all computer interactions will be Voice mediated AI, these are the problems that startups are built upon. And it will likely come from a startup that no one knows of today to solve it, for many reasons, just as Google solved web search in a Yahoo world.
Thus, as it stands today, the number of Skills, voice apps or other programs can become an almost meaningless number that could signal some initial early developer interest that ultimately clog the system with Orphan Skills and Zombie Skills.
We can fix this—easily and I hope it is sooner rather than later.