Fresh from designing and deploying a new Alexa Skill, I am struck by three things:
- Designing for voice interaction requires a new set of skills and tools. It’s not as simple as “audio content.”
- Voice interaction is a relatively new behavior for which there are few real experts or templates to follow.
- The pressure to develop skills that people will actually use will make us better marketers.
To say that voice interaction will become big may be like saying the computer mouse would be big back in 1968. The popularization of Amazon Alexa and Google Home have done more in the last three years to make verbalizing commands almost normal than anything before (including Siri commercials from Apple).
“U.S. Smart Speaker Ownership Rises 40% in 2018 to 66.4 Million”
Homeowners Ask for Help
Check out Travelers Home Central Skill. We designed it for homeowners who want a little extra guidance on what home maintenance tasks to tackle and when. We mined search data to understand the questions people have. We dug into the performance data of the hundreds of content stories we have on our Prepare & Prevent resource. We paid attention to the scarce data available on current voice skills. And we dug into the homeowner inside each of us to craft a utility that helped people like me who need to be told to tackle certain chores at certain times to protect the value of their home and the safety of those who live in it. We will learn how to grow its value to homeowners over time.
Designing for Voice Makes Us Better Marketers
Here are 4 ways designing for voice interaction actually makes us better marketers overall:
1. Identify trigger moments of need. The biggest uses of Alexa Skills include playing music and getting your news (“Flash Briefing/News"). Voice search is on the rise as people translate their smart phone behavior to voice – “Alexa, what was the last Jim Jarmusch movie?” In general, we are all just learning what we can do via a voice assistant. For brands, that means a healthy appetite for experimentation. It also means thinking in terms of trigger moments – what would cause a user to ask Alexa a question?
For homeowners, it might be prepping your home for a big, forecasted storm or simply that amorphous moment, called “winterizing your home,” when we want to take care of overdue maintenance chores.
Finding or creating a true moment of need that would be remedied by a brief voice interaction is not as easy as it may sound. Since talking to your fill-in-the-blank device is a new thing, it's not as easy as asking people. With voice interaction, identifying a need that can best be met with a voice interaction is tricky. We need to constantly be trying things and seeing how people respond.
2. Design to use-cases or journeys. Like trigger moments, we need to think in terms of what a person is trying to get done. Some skills – like “Flash Briefings”, now “News” – deliver information with no specific next action. People want a snippet of information to chew on. When someone asks for a home maintenance tip, they may be ready to tackle that job immediately. More likely they may want to schedule a reminder for the weekend when they are more prepared to pull out the ladder and tackle the gutters.
3. Align to what people say and intend. Many a good prayer includes the phrase, “seek rather to understand than be understood.” Figuring out all of the ways someone might ask a question and what type of answer they are after become the core skills for “voice interaction designers.” This builds on the skills of great search engine optimization specialists who seek out the myriad ways people may type a question into Google. Conversational English is different than the typed word. When people engage with voice assistants, they often seek to be as brief in their question as possible. Thankfully, for now anyway, people understand that the tech is early days, and they will go out of their way to try and frame an answerable question.
4. Iterate, expand the “corpus” and figure out how to apply artificial intelligence to increase the value. More than any new channel or format, voice assistants need constant iteration and improvement. Not only is the format young and hard to get “right,” it begs for iteration as the system learns from the questions posed to it. Adding new answers or anticipating a variety of utterances takes ongoing work. Natural Language Processing (NLP), one of the tools in the AI and machine learning toolbox, makes learning while doing more possible.
Designing for voice demands we think about the journey - what people are actually trying to get done - and the trigger moment for seeking a solution. Because its voice, we literally have to listen to our customer and lean into understanding what they need. We have to be great at data and technology from the start with voice in order for its value to the customer to grow. Automation via AI and machine learning cannot really be after thoughts. Designing for voice pushes us to be better marketers.
Useful Resources:
How Amazon Alexa Works: Your Guide to Natural Language Processing
Exploring VUI Design: The Basics and Getting Started
The future of voice search: 2020 and beyond
How Voice Technology will Influence Marketing in the Future?
The Top 20 Brand Innovators in Voice
Comments