Using chatbots to enhance audience understanding

Technology is helping bridge the gap between quantitative and qualitative research practices by offering more ways to connect with consumers.
12 June 2019
Young girl with robot
Wei
Guo

Business Development Director, Hong Kong

Get in touch

We’re getting creative with research. Classic quantitative and qualitative techniques no longer fit every research project, whether that be due to time, budget or resources. Thankfully, technology is helping bridge the gap between quantitative and qualitative research practices by offering more in-depth ways to connect with consumers.

One way Kantar has seen success is by utilizing chatbots to obtain benefits from both types of research. We’ve found that chatbots make information gathering more conversational between brand and the audience, and participation is more interactive.

Below we explore an example of how combining chatbot technology with a highly engaged respondent base, including Kantar’s WeChat respondents, provided great benefits to the overall research deliverables.

Understanding the research needs and barriers

Our client, a major global premium fashion brand, wanted a more detailed view into specific dressing habits and daily routines of their medium to high income audience across multiple markets. They recognised that this was aiming for a lot of detail from a lot of people, so traditional research methods wouldn’t be an ideal fit; traditional qualitative work wouldn’t provide the level of detail needed, and traditional quantitative methods would be expensive.

Why chatbots were an ideal solution and how we did it

In our experience, chatbots create a more conversational environment with respondents, offering in-depth feedback and greater engagement. By leveraging chatbots, our aim was to help the client receive detailed insight into who the audience was, the fashions they choose and for what occasions through dialog and image uploads.

We first developed a conversational, automated script that would work with responses provided in each respondent session. It was initially designed for the Facebook platform for some markets, and then converted to fit the local social platform in China, WeChat.

The benefits and research outcomes

Utilizing social platforms gave respondents the capability to upload multimedia content, including images and short video clips, in a familiar environment. The visuals provided richer information, giving specific detail on how respondents compiled outfits for different occasions. With the conversational style of questioning, respondents invested more time in providing considered answers and meaningful uploads. We also experienced the greatest overall engagement levels among Chinese respondents on the WeChat platform. The average sessions were twice as long as any other market and six times longer than the lowest performing market.

Overall, tailoring a research approach that blended qualitative and quantitative practices was a huge success. We completed the project with 80 respondents, each uploading three outfit images, per market as targeted. Utilizing the chatbot technology offered a more cost and time effective solution to traditional qualitative methods, while achieving more in-depth knowledge on the audience than typically collected in traditional quantitative research methods.

As this technology evolves, we expect to see continued opportunities for enhanced audience understanding. Building a conversation with respondents not only allows them to speak more openly, but to have a more engaging, rewarding experience overall. From creating personas to align with brand voice, to the use of familiar environments on social platforms, chatbots have the potential to change the way we realise Diary studies and bridge the gap between qualitative and quantitative research. Conversational AI is offering a new level of human interaction and human understanding and it’s only just begun!

Get in touch

Publications

Over 1500 scientific and peer-reviewed articles to further understanding in health research.
Search the database