As voice-enabled computing moves to the mainstream and drives the next wave of digital innovation, voice analytics become mission-critical. (See our analysis of what is driving users to voice.)
“Alexa: tell me what my customers want.”
For years, “voice analytics” largely referred to how call centers analyze your conversations with their representatives. The frontier of personal voice-enabled assistants has introduced the opportunity to not only streamline communications with your customers/employees but to learn from those interactions in virtually real-time.
As apps and screen experiences become voice-enabled, it’s critical to start learning what works in a voice environment. The easiest playground is an Alexa Skill.
While there are still significant issues around discoverability and lack of voice-screen multimodal interaction, we think of Alexa not as an ROI driver, but instead as a very low-cost way to learn how voice can be deployed successfully for your business in a future voice-first world.
Even with small numbers of users, you can use Alexa’s analytics tools to begin understanding what users want and like.
Amazon recently revamped the UI for creating and managing skills for Alexa, with testing and measurement taking center stage. Some of the analytics now available:
- How many utterances users make in a typical session with your Skill
- What percentage of user interactions with your Skill result in a successful outcome
- Which of your Skills’ capabilities are being used the most and the least
- Basic cohort analysis, so you can see whether people acquired during a given month (say, after you launched that new feature) behave differently than those who joined in the month prior
How this plays out for your industry
The possibilities for analytics in voice interactions can translate into some valuable insights into customer behavior. Here are just a few industry-specific examples:
Fintech: Better understand which transactions people are most comfortable transacting over voice. Trust is a constantly moving target!
Retail: See where in the sales path customers are having to ask for help (note: transcripts aren’t available for voice assistants like Alexa at this point, but you can still identify friction points by volume of questions and the answers your skill is called on to give).
Media: Begin to understand how people want to interact with your content at different times and in different contexts (e.g. news briefings in the morning, in-depth coverage in the evening).
Enterprise: Identify key sticking points for field technicians based on frequency of certain voice queries, and build these FAQs into their training modules to minimize accidents and errors on service calls.
What’s next in voice analytics
As exciting as these new metrics might be, they still represent only the early days of the huge potential of voice analytics. To understand that potential, we need to look briefly at what makes voice distinct.
If someone leaves your website without buying anything, you can often only speculate as to what went wrong (or try to learn after the fact, via a survey). With voice, they’ve very literally told you what they’re looking for, even if your Skill (or Google Home Action, or Siri) wasn’t able to handle it.
To tap into this potential, what we most hope to see next in voice analytics solutions is some analysis of what your customers are asking for when your Skill can’t process the request.
That information, if presented properly, is a goldmine, a brand new avenue for understanding customer behavior. Admittedly, such a dashboard is more difficult to create than the one available to Alexa developers today. But we are talking about a company that can get coffee filters to your front door in 30 minutes via drone. It’s only a matter of time.
Successful companies understand what their customers are looking for. Voice, with its nuance, its immediacy, and its increasing ubiquity, represents a new level of opportunity for doing just that.