Today, there are more opportunities than ever to engage customers with voice apps—but that doesn’t mean it’s always easy to know where to begin. To help put things in focus, Alexa Chief Evangelist Jeff Blankenburg shared 5 common characteristics of successful voice apps in a recent Tech Talk.
From tips on streamlining your interface to standing out in the crowd, these insights draw from Jeff’s experience working with hundreds of developers. Read on to find strategies designed to help you engage customers and deliver excellent voice experiences.
When you’re designing for voice, the possibilities are endless—and that itself presents a key challenge to building effective skills. In voice design, there’s no visual hierarchy to help users along like there is in screen-based devices. That means there’s a greater chance of someone getting overwhelmed.
So how can you reduce the cognitive load? Start by trying to accomplish one thing as best as possible rather than trying to do it all.
One example is Jeff’s own skill covering baseball analytics, GamesBack. In baseball, there’s a huge variety of metrics you could cover. Yet users are more likely to gravitate to a skill if they know exactly what they’re going to get out of it. That’s why GamesBack centers on only one thing: Seeing how many games back your favorite MLB team is from first place.
By building skills that put a spotlight on the data users care most about, it ultimately becomes easier to incorporate them into routines—and that pays dividends. Routine users have 40% higher retention rates compared to non-Routine users.
What’s in a name? When it comes to effective Alexa skills, it’s more than meets the eye.
If you think about your favorite app or website, an icon likely pops into your head. Voice skills don’t have the luxury of that type of shortcut, meaning users need to remember the name itself. Don’t get discouraged, though; this opens new opportunities to get creative and develop a brand that sticks.
One skill that Blankenburg highlighted is The Magic Door. It’s an adventure story that employs its name to set the stage with a simple command: “open the Magic Door.” After that, users hear the creak of a big door and are welcomed to an immersive audio world. All of this provides a sensory memory that reinforces the name and even melds directly into the skill experience.
You can also try to use other ‘launch’ words within your skill’s title. If your skill name matches the action word—for instance, ”run My Race Trainer” or “play TKO Trivia”—the experience is seamless and the name is easy to remember (or, in the best case scenario, there’s no need to remember it at all).
One final note: Developers sometimes want to come up with especially clever names, but this can be a mistake. Instead, keep it simple and avoid complex spellings. When in doubt, think about what goal the user really accomplishes, and consider basing the name on that.
When you consider how a user interacts with your skill, it’s tempting to start out by listing the ways people will ask for specific content. For a recipe skill, an anticipated command could be something like “tell me about the right way to Julienne a potato.” In reality, though, this isn’t usually how an interaction will work—and, furthermore, it assumes new users are
already familiar with all available content in a skill.
A more likely scenario is that they’ll say, “teach me something new” or “play a speechcon for me.” These are both examples of intents.
Try to account for the fact that someone might ask for something broad without knowing the specific name or detail, then let them add knowledge organically. That way, new users won’t get discouraged from exploring your skill due to a lack of background information.
When people talk to a voice assistant, they’re much more literal than they would otherwise be. That means developers sometimes mistake what an open-ended question really is. Although a question like “Is there anything else I can help you with?” may seem open-ended, most people interacting with Alexa give a yes/no response.
Your goal should be to promote active listening. The best way to achieve this is by presenting questions that are easy to process, giving adequate context, and ensuring the conversation flows. Here are Jeff’s strategies to elicit useful responses:
Websites, smartphone apps, and other visual interfaces thrive on familiarity and even encourage users to ‘skim’ through content. Successful Alexa skills, however, should take the opposite approach.
As Jeff describes: “We’re trying to hack humans. The more they pay attention, the better experience they have… If a voice experience is predictable, it ends up being a hindrance for users because they stop listening.”
So how can you introduce just the right amount of unpredictability? Start by keeping things fresh through variation. For every bit of speech you use, write five to seven versions of it—even a little goes a long way! simply shifting word order helps retain attention, in turn promoting higher engagement.
You can also change the order in which responses are presented. For instance, a travel agency skill would require a few inputs (dates, locations, number of passengers), but there aren’t rules about when you need to gather that information.
If users tune out, they give wrong answers—and if they give wrong answers, they’ll have a bad experience.
Tune into the full Tech Talk on-demand to get even more details.