Finding new ways to engage with voice and chat
This article was published in Research World.
Market research methods have, by and large, done an excellent job of reflecting societal norms in terms of communication and interaction. Face-to-face intercepts gradually giving way to CATI, which gradually gave way to online surveys; group discussions gradually giving way to online qualitative methods. The ‘old way’ doesn’t go away altogether, or at least it doesn’t have to, but through advancements in technology, and changing consumer preferences and norms, there has always been a willingness amongst innovators in research to try something new.
Yes, there is a risk – failure, sample bias, technical mishaps. But there are also rewards: greater engagement and reach, reduced fieldwork cost and the ability to get closer to the true insight and System 1 thinking.
Over the past 12 months we have conducted several experiments to understand how people respond to surveys conducted by voice and via chat compared to online surveys. This includes the paper ‘Giving Voice to the Voiceless’, which we collaborated on with Telstra, Australia and which will be available on ESOMAR ANA library after the Asia Pacific Insights Festival 2020, broadcasted 2-3 November 2020.
Voice and chat in 2020 are to the research industry what online research was in 2001. Increasingly normal, familiar easy ways to communicate, and in the case of voice, a new way to engage with technology without the limitation of having to use a screen. There is some overlap in how these technologies operate, but in practice we use them differently.
A quick look through my Google Assistant history (shared by five people in my household) shows that these were the last five requests, all in the past 24 hours:
Play ABC Radio Melbourne
Turn on the fairy lights
Turn off the fairy lights
How can I play Deadpool?
What streaming service plays Deadpool?
Play Folklore by Taylor Swift
If I was to copy and paste the last five ‘chats’ I had with people on the various chat services I used they would be vast and varied: short exchanges with colleagues on Teams, as well as banter and memes. Lengthy debate/discussion/dialogue in an ongoing WhatsApp chat that covers parenting, lockdown, the second wave, Trump, plans for Christmas, the weather, and some really helpful and pleasant interactions through Facebook Messenger with people from my local community groups, running groups, businesses I follow and, of course, friends.
When looking at how we might be able to utilise chat or voice technology for insight gathering, it is important to understand the context first and to idiosyncrasies of the audience and the medium.
There really is little merit in taking an effective online survey and converting it to chat or voice and sending it out to a panel.
Instead, the opportunity is to think creatively about ways to incorporate these new research methods into studies in a way which address business and research objectives and in so doing, solve problems that can’t otherwise be solved. Engage with people in the ways that they are used to engaging. When it comes to voice, focus on using it for opt-in respondent-driven feedback in the moment. This might be a tool which sits alongside say an online diary or forum. Allow the respondent to quickly share a thought, or experience by designing a simple voice bot with one or two interactions and a memorable code word.
For example:
Bot: Hi, [first name]. Have you washed your hair today?
User: Yes
Bot: Why did you wash your hair today?
User: I wash it every Monday
OR, I went for a run and it was sweaty
OR, I had an important meeting and I wanted to look my best
The interaction is short, which aids compliance and reduces drop out, but the insight is powerful because it’s timely. And it is a bot which can be used over and over again with different audiences in different markets and can leverage machine learning to ask smart follow-ups over time. Wash, rinse and repeat!
When it comes to chat, focus on connecting with audiences that are just not interested in engaging through other methods like online surveys or forums or who are under-represented in panels. Find out the channels they use and design an ongoing chat which can work seamlessly with that platform and which pushes notifications from that platform (like Facebook Messenger). Or, recruit them from social media using a bot, and send a series of small chat-based surveys via SMS. Because chat is a visual medium, it is easy to use a combination of open-ended and prompted follow-ups so it can be longer than voice and can include stimulus. The ‘conversation’ (survey/discussion guide) itself is only limited by the imagination of the researcher. And, like our voice example, if it’s built on the right foundations, it can be used over and again with different audiences once set up.
Chat and voice aren’t for all audiences and certainly aren’t for all researchers but, when well designed and executed, are a fun and popular way to engage audiences in research and uncovering truth which lies in the moment.
