RAIN, a New York-based digital consultancy, has worked with Amazon for over two years building skills for the company’s Echo, Tap, and Fire ranges. Co-founder and chief digital officer, Andrew Howlett, spoke to WARC’s Sam Peña-Taylor about building new behaviours and linking voice into an overall brand strategy.

How did RAIN’s relationship with Amazon, and specifically its voice technology division begin? 

We’ve always prided ourselves in staying ahead of new technology, so Voice was not new for us in terms of looking at what was emerging as a way to engage consumers. But we really used a lot of our same methodology, strategy, planning, and testing. Through the natural course of business and strategy with our Campbell’s client, we saw Voice technology as a central part of a broader ecosystem that we were building for them. That took Amazon by surprise -- that a team would build such a high quality, high integration skill before they had even really done anything like that themselves. From there, we became a trusted agency partner to Amazon.

How do you build a strategy around a new consumer behaviour? What do you think made [voice] so successful?

First and foremost, "voice" itself is the original user interface. None of us really have to be taught how to do that behavior. We all know it, because that was one of our first learned behaviors as humans. So today, we’re really just coming back to something that is a very natural, normal way for us to communicate and interact.

Technology has finally caught up with our intentions and our natural interactions. That said, any strategy around a consumer behavior has to be around removing barriers. You remove friction, and consumers will want to adopt it. So really, it starts with building the right tool, the right utility, and something that is useful or helpful to their lives. That will influence them or bring about a change in behavior.

What were the barriers to promoting the Echo?

One barrier is privacy. Some people are resistant to an always-listening device, so we're helping to educate them on what that really means, and what that doesn’t mean. Just like any new technology, Voice also had more limited distribution. Brands don’t necessarily want to jump onto a new platform that only a few of their customers have. Clearly in the last two years, we’ve seen a dramatic adoption of smart assistant speakers as well as other devices that have voice assistants on them. So now, it’s not a question of “when” -- more about “how” clients want to build on the platform.

What questions do brands need to ask when considering a voice strategy?

How does this fit into a broader ecosystem, or a broader strategy of consumer engagement? Where will this solve a problem? Voice should not just be ‘tech for tech’s sake’ -- for it to be truly useful to a brand, Voice needs to fit into a broader strategy.

Which case study in your own portfolio best illustrates what voice can do?

Voice tech isn’t always about “talking” -- sometimes it can be about listening. We built an Alexa skill with Tide that offers stain removal tips -- so if a user asks about a specific stain and fabric combination and it’s not one that Tide has encountered before, maybe a new product comes out of that. Voice, in this case, is like a focus group of one.

Who, outside of your own work, has done voice well?

Dominos has always been a leader in emerging technologies. Most importantly, I think they’ve done a good job in weaving Voice into part of their broader system. You can seamlessly track your order via Voice just as you can on your Apple Watch or in your mobile app.

The in-house teams at Amazon and Google have also built some great experiences -- one that has been very successful is the Jeopardy skill. It has very high use, and follows a lot of the central game mechanics like “Come back the next day for a new set of questions” as well as a leaderboard. They’ve built something that people like to use, and is engaging. They’ve also successfully leveraged a recognizable brand: Alex Trebek.

What are the most common challenges brands and agencies face when designing for voice?

A common challenge is identifying where a Skill or Action fits within the brand’s digital ecosystem, and what should be built to best serve customers. A lot of brands and a lot of agencies love to show off with the buzzy and splashy, but that doesn’t necessarily help the company or the client address business drivers.

The next piece would be voice experience planning and design. Some teams underestimate the voice experience design process, which is actually very complicated. It’s important that the agencies and brands really think through that conversational strategy and conversational flow, because there’s no screen to take you back to home if you get lost.

For many brands, Alexa skills have struggled to attract repeat usage, making this one of the key challenges to widespread brand adoption – how else might a brand design a voice strategy outside of creating a skill?

There needs to be reason for people to want to come back -- a repeatable content or utility. Whatever a brand’s voice experience is, they should be talking about it in the other places where they’ve got connections to their customers. So whether that’s social, whether that’s on packaging, if it’s on some of their earned media, or their own media -- they should be talking about how to get into their skill, and why it’s useful. You see brands all the time reminding people to use their mobile apps for a certain functionality or utility. And how long have mobile apps been around? But brands are still having to remind people that that’s there.

Outside of creating a skill, brands are really going to have to start thinking about voice and conversational search. We know that 50% of all search is going to be by voice by 2020. Brands have to consider that our search via typing is different from our search via voice. I would never ask you, “Sushi, Salt Lake City, medium price.” But that’s how I would search it in a Google search bar. I would say, “Hey, do you know of any great sushi places, kind of medium price, in the east bench of Salt Lake City?” If content on a brand’s site is not written in a way for that to be scrubbed or to match up with that, they’re going to see a decrease in search traffic whether that’s organic search traffic or paid for search traffic.

Often we hear about the need to develop a ‘brand voice’ – what does that mean to you?

In this case, we’re talking about a personality, which is absolutely something that should be considered in voice-driven experiences. Personalities allow for variation and a more conversational flow. A robotic experience won’t feel personal to a listener -- if a brand is sort of cheeky or sarcastic in the way that they interact with their fans, that should be how their voice experiences interact with consumers, too. The brand voice and tone should show through.

Brands can also record their own content. We launched a skill for the movie Dunkirk for example, where a user only hears Alexa’s voice a few times in that whole skill. The rest of it is all pre-recorded, voice actors playing the different parts in a “choose your own adventure” style experience that we built.

In cases where an assistant’s voice is used, brands need to work a little bit harder to make sure their brand personality matches up with that voice. Think of it as a shift from brand voice to a "brand character" moving forward, where the output is personality and tone.

How do you measure the effectiveness of a skill in an overall brand strategy?

We track and measure against objectives -- it’s not any different from mobile or web in terms of how you determine effectiveness. If the objective is transactions and dollars, you need to track that. But the objective could also easily be market research. You can build that tracking into the analytics, and tag things you want to look for -- and it’s going to be a lot of trial and error. But there’s no question brands need to stay committed to the platform, because the platform is expanding exponentially and adoption is happening faster than many thought was possible.