The respondent of the future will be a digital native. According to Felix Rios, of ESOMAR, “These shoppers of tomorrow arrive fully wired. They pinch to zoom instinctively. They learned to communicate their ideas in 140 characters and share their thoughts and opinions without being asked to do it. They grew up watching videos that are 15 seconds long. They text to communicate rather than talk. The participants of the future are growing up with the concept of instant gratification deeply coded in their DNA.”
Many researchers believe that marketing information will be dominated by passively collected big data, with the survey used to support behavioral analytics. Other researchers have a more optimistic view of the future of the survey. But in any case, one thing is clear: the old-fashioned survey will have to evolve to keep these younger consumers engaged and responding with high-quality data. And if that survey is to avoid frustrating the digital native, it will have to rely on artificial intelligence and machine learning to do so.
Artificial intelligence (AI) is the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. Machine learning gives computers the ability to learn without being explicitly programmed. Together, they could create a simplified, faster and more natural online survey interface. The survey of the future will give the impression of instant reaction, minimal wait time, and that the questionnaire is “learning” your preferences and predicting your responses.
Think about the way you interact with Facebook. Each time you visit it is different, depending on your device, what you have posted, what others have posted, and what advertisers are targeting you. While the jury is still out about those “stalker ads” that seem to follow you from app to app, that sort of experience is the norm for the digital native. Rios explains: “While the participant is answering each question, in the background the survey engine has to be loading the rest of the survey, prioritizing the next question in line. Artificial intelligence should help us predict what is the most likely path this participant will take and prioritize those in the loading queue. Pretty much in the same way that Google starts to suggest search terms as we start typing a query in the search bar. These terms are highly relevant, and it’s not by chance. The system can’t wait until the participant clicks next to trigger the next steps. By the moment the participant is halfway through a 5×5 grid, the survey platform should have predicted what will happen next. This needs to happen in real time as he is interacting with the questions. Machine learning will make sure that the system becomes more and more accurate as it successfully repeats the prediction process.”
As we wrote in our recent blog on paging versus scrolling in surveys, online surveys today are modeled after paper surveys, a design that has gone by the wayside for younger respondents. Scrolling is the new preferred technique. As long as we continue to evolve the online survey to replicate the digital experience for respondents, the survey as a data collection method will remain viable. And AI and machine learning will contribute to that.