In 2018, voice technology will go mainstream. According to comScore, in 2017, half of all smartphone owners used voice technology on their phones, with one in three using voice technology daily. Further, voice-first devices—which include smart speakers such as Amazon Echo and Google Home—are expected to cross a “critical adoption threshold” in 2018, and their growth is likely to accelerate in the coming years. As the market for voice-assistant applications and smart speakers continues to expand, brands must incorporate voice technology to stay relevant and competitive. However, brands have only a finite window for owning the voice user experience.
According to our research, 70 percent of smart-speaker owners have experienced problems or frustrations using their devices. Plus, 25 percent felt that these smart speakers were not designed with their needs in mind. The numbers were also troubling for third-party voice applications. The research found that 63 percent of those surveyed encountered some type of problem or frustration with these applications. In fact, 29 percent said they would delete an application with which they had a negative experience.
Voice applications are yet another user touchpoint for brands. Failing to deliver a satisfying experience at each touchpoint can damage a brand’s reputation across every touchpoint. So how can brands develop voice technology that will help them satisfy their customers?
To build great voice experiences and convince your stakeholders to use them, we recommend following these seven steps.
1. Gather your stats.
It’s no secret that the smart-speaker market is growing exponentially, but risk-averse stakeholders may not recognize the opportunity. Are they aware that 47 million Americans currently own a smart speaker or that lead analysts at Gartner expect the global market to grow to $3.5 billion by 2021? Gathering statistics and sharing them with your stakeholders is the best way to showcase the market opportunity and get buy-in.
2. Know your user.
These days, most organizations have a wealth of data about their users’ behavior when using their brand’s products or digital experiences. But how well do they really know their users? Let’s be honest: your users aren’t thinking about your product. They’re thinking about what your product affords them. What are their motivations, aspirations, and challenges? How does your product or service fit into their lives?
Ethnographies and contextual research are the best way to gain these insights because rather than focusing just on users’ reactions to a given solution, the goal is to gain an understanding of how people live—and, subsequently, what you could do to make their lives better. In addition to seeking product-market fit, great companies look for product-human fit. Does your voice application truly address users’ wants and needs? Ethnographies and contextual research ensure that you truly understand your users. We recommend that every company developing a voice app begin by first researching their customers’ usage of voice assistants in their natural environments to discover the opportunity space. What would truly solve a customer painpoint?
3. Determine the highest value task.
Armed with this rich understanding of your users’ motives, goals, and challenges, start to percolate some potential solutions. You’ll likely have a number of ideas for ways in which you could improve the lives of your customers, but you’ll need to determine which idea to tackle first. We recommend surveying your users to establish a quantitative priority for each of those ideas. Then, look for opportunities to score a quick win that will allow you to demonstrate success with this platform and secure additional funding to expand your offering.
4. Conduct concept testing on low-fidelity interactions.
Once you have a clear idea of your priorities, it can be tempting to build a solution and test it. However, an iterative approach to product development is best. Before you invest in building anything, we recommend that you do concept testing with a low-fidelity manifestation your ideas. In UX research for voice, this is often called Wizard of Oz testing because instead of code, there’s a person behind the curtain.
First, map out your conversational interactions, creating rough decision trees and dialogues. Then, do usability testing of the designs to see what you haven’t anticipated. You can do this simply, with the participant in one room, talking to a speakerphone they pretend is a smart speaker, and the person acting as the wizard in another room, responding to the participant by reading from the script.
You can also go to a slightly higher fidelity approach by taking advantage of Amazon Polly, which is text-to-speech software that speaks in Alexa’s voice. In this case, the wizard cuts and pastes written responses into Polly, so the participant hears the replies as if they were coming from Alexa herself. This approach can result in greater delays in response time, but it also reduces the bias that can result from the wizard’s changes in tone or emphasis that would occur when not using Alexa’s voice.
5. Build a business case.
As we previously mentioned, Gartner projects that smart speakers alone will constitute a $3.5 billion market in just three years. Among those who already own smart speakers, 83 percent say they often use their devices as part of their daily routine, and 42 percent now try new voice applications more often than they did during their first month of owning a device, indicating a desire to do even more with their devices. Consumers want to interact with brands in ways that are relevant to them and leverage the unique opportunities of smart-speaker platforms. For a business to succeed in today’s market, they must begin harnessing the power of voice technology.
So where does your product fit into that market? Having gained an understanding of product-human fit and defined a roadmap for solutions or features, you’ll also need to understand the product-market fit. In such an untested market, it may be difficult to develop these projections, but stakeholders want to understand your voice application’s break-even and return on investment (ROI) before allocating budget to your project.
Once you’ve mapped out your intended voice experience and tested it as a concept, it’s time to build it. But before committing it to code, we strongly recommend a design-test-iterate approach to development. Thankfully, there are a variety of ways to test prototypes before voice applications go live. As you know, conducting user research is essential because it can prevent the need to make costly fixes post production. Further, given the high risk that users who have a negative experience with your voice application will abandon it, the importance of this phase of development should not be underestimated.
Sayspring is a free, Web-based, prototyping-software service that lets you easily design for both Amazon Alexa and Google Home. Similar to the wireframes and visual-design mockups you create for testing application designs, you can use Sayspring to create high-fidelity voice prototypes that let you explore and document the user experience before you start to code. Once you’ve designed your prototype, conducting usability-testing sessions will provide critical insights into what is working well and what isn’t.
Once you’ve built your voice application, you can beta test it using the tools that Amazon and Google provide. Using the available tools, you can test your design changes and invite feedback from friends, family, and your larger network. You should also consider Pulse Labs’s Web-based beta-testing platform, which provides richer insights and lets you upload your own list of test participants or identify your desired demographics and let Pulse Labs find appropriate participants outside your network. While prototyping tools help you to refine your designs before coding begins, beta testing helps you to test your implementation before it goes live.
7. Create an ongoing feedback loop.
Once you’ve launched your voice experience, take advantage of voice-analytics products from Chatbase, Dashbot, or Adobe to understand the actual use of your product. Much like the Web or app analytics we’ve all come to rely on, these products let you track users’ behavior and understand key factors such as active users, engagement, top intents, interaction funnels—even when errors occur—and let you analyze users’ sentiment while engaged with your voice application. This provides important feedback, allowing you to optimize the voice experience—just as you would a Web site or app.
One could argue that every company today is a digital company. A business’s many customers represent an equal or greater number of users on a variety of platforms, including Web, mobile, and tablet. In the coming years, we’ll see a revolution in voice technology that is comparable to the revolutions we’ve already experienced around the Web and mobile devices.
For a brand experience to truly resonate with customers, it must connect with them on several levels. With voice-activated technology, brands have the opportunity to speak directly to their consumers and guide them through a variety of applications. Speech can be highly nuanced, but it also has the power to deliver personal, meaningful experiences. Tapping into voice technologies offers a key differentiator for brands in 2018, and by taking the seven steps we’ve outlined in this article, brands can ensure they’re developing smarter, more efficient, more effective voice technologies.
Amy founded AnswerLab over a decade ago to help the world’s leading brands build better digital products. Under her watch, AnswerLab has grown to become a trusted UX-insights partner to companies such as Google, Amazon, and FedEx. Read More
As a Senior UX Researcher at AnswerLab, Chris helps Fortune-500 clients create exceptional experiences for their users. Chris has led studies to understand the role of emerging technologies in the future of banking as well as the behavioral effectiveness of workspaces, but spends most of his time researching conversational interfaces. Read More