Has Voice Search Stalled?
July 9, 2019 | Contributed by: Greg Sterling
Amazon Prime Day is next week — it’s actually already underway — and the company will be pushing its various Alexa devices hard. Aggressive discounts and bundling will drive hundreds of thousands of sales of smart speakers, displays and smart home gadgets. Google will counter with discounts on Nest/Home products too.
Then will come the post-Prime Day press release that says the Echo Dot was the #1 product sold on Prime Day. Another round of virtual assistant/voice search hype will kick in. Some of that hype is justified but there’s also increasing evidence that smart speakers and displays are not the dynamic new “search” channel that everyone expected — at least not for the immediate future.
First, there’s an important distinction that most “voice search” discussions don’t make between smartphone and smart speaker usage. Voice search on smartphones is regularly used and growing. However, the range of smart speaker usage is more limited. Surveys often blur the two categories and behaviors.
The chart above, based on data collected at the end of 2018, reflects the most common smart speaker use cases: ask a question (a version of search), music streaming, weather, alarm, timer and radio. Shopping and search behaviors are further down the list. Use a favorite skill is in the top half. However, I’m extremely skeptical about claims that people are regularly using Alexa Skills or Google Actions, although many people have tried them.
Amazon recently announced general availability for “Skill Connections,” which enable one Skill to use another to accomplish more complex tasks. There are more than 80,000 Alexa Skills (it’s probably over 100K now) but the overwhelming majority are mediocre and there’s a massive discovery problem. I have yet to find anyone using multiple Skills in earnest, even one regularly.
To fully realize the potential of these devices, voice apps/Skills need to be more easily discoverable and more useful without the awkwardness of specific invocation phrases and “magic language.” They also need to be more elegantly integrated with smartphones for the completion of tasks that are difficult to accomplish with voice alone (e.g., booking, shopping).
Another problem is privacy. Negative publicity surrounding surreptitious recording and storing of conversations is having an impact on demand for smart speakers, according to at least one survey from NPR and Edison Research.
As the graphic indicates, privacy and security concerns are now a bigger purchase inhibitor than they were in 2017. The top-three reasons why people aren’t interested in owning a smart speaker/display are now:
- Hacking risks
- Concerns that “smart speakers are always listening”
- Worry about government eavesdropping
All this points to slower adoption and usage than the industry predicted or would like to see, though there are more than 100 million of these devices in US homes. (And there are still very bullish projections out there.) Again, voice search on smartphones and smart speaker usage are two different things. One is growing the other has arguably stalled.
There’s still great “voice search” ‘potential for smart speakers (in all their forms), but it’s going to take a number of developments for that to be realized:
- Better privacy controls to rebuild consumer confidence
- Simplification of Skills and improvement of overall functionality and quality
- Simpler and better integration between smart speakers and smartphones
- Richer data and more conversational capabilities around that data