Analysis

Do device swipes deliver a perfect match for innovation testing?

Analysis

Do device swipes deliver a perfect match for innovation testing?



What started as a swipe to find your perfect match on a smartphone can now be used to find your perfect job, your perfect pet and your perfect pair of shoes. So, why not use swiping to build the perfect new product? Dating app style swiping is a hot topic in consumer research. The potential insight gained from using swiping in consumer research is being presented as everything from being a way to generate “System 1” measurement to being capable of offering a more intuitive respondent experience.

For those unfamiliar, the concept of “System 1” and “System 2” thinking was introduced in Daniel Kahneman’s book “Thinking, Fast and Slow,” which describes two different, yet parallel approaches the brain uses to make decisions. System 1 is the implicit, non-conscious approach and System 2 is the deliberate, conscious approach. Survey methods designed to evoke a non-conscious System 1 response have recently come to the forefront of consumer research to help uncover motivations that consumers are unable or unwilling to express. Survey research, by design, evokes a System 2 response from consumers by asking them to consider a question and provide an answer. However, because swiping is fast, there is a common perception that it’s a new way to tap into System 1 non-conscious thinking.

At NielsenIQ BASES, the prospect of deriving new insights from consumers by using different interfaces is an exciting idea. However, predictive analytics is at the core of what we do, so adopting new methodologies is not taken lightly. To consider a new research method, we need to fully vet that any new approach will pay off in at least one of three areas: improved accuracy, increased efficiency or new insights.

We took on swiping as our newest research into methodological innovation. So what did we find?

Swiping data compared with multiple-choice surveys gives us directionally similar results, though it lacks the same precision. The right-swipes or “yes” responses correspond with the respondents who would make a selection on the positive half of a multi-choice response list. In other words, using swiping just converts what we would normally ask as a five-point scale down to a two-point scale. That’s exciting because we can replicate some of the same data we normally collect with swiping. However, to make accurate predictions, we need more specificity from consumers to parse out the “yes, definitelys” from the “yea, maybes.” Without this level of detail, we could judge an above-average and below-average innovation in our database as having the same potential, even though one would generate 3x as many dedicated buyers.

Many research companies are promising that swiping can access System 1 measurement, which traditionally could only be derived from studies using methods that track our responses to stimuli like electroencephalogram (EEG) or implicit response tests (IRT). Unlike the non-conscious measurement captured by EEG and IRT, we found conclusively that swiping is a replication of other conscious, or System 2, techniques. Specifically, we measured the time it took for consumers to swipe to test the notion that a fast swipe indicates a System 1 response and a slow swipe indicates a System 2 response. What we found instead was that a fast swipe indicates the conviction of a “yes” or “no,” essentially replicating what is derived from a multi-choice question, which is clearly a System 2 measurement technique. And while the time to swipe may help us better represent the diversity of response from “YES!” to “Um, yea?,” it will still be subject to uncontrollable distractions, such as being called from the waiting room for your appointment or a child asking for a snack. Ultimately, while swiping may give researchers a way to tap into some of the insights we collect today, it lacks further analytical benefits, as it is purely a System 2 measurement tool.

Though not a System 1 technique, there are some potential benefits to measurement using swipes. In terms of respondent experience, it’s about 4 times faster to swipe than to answer a multi-choice question, which could be used to efficiently evaluate many products repetitively once the swiping action has been practiced. However, while multi-choice questions have been prevalent in everyone’s lives since their first standardized test, swiping is a new behavior that is more embedded with younger generations. BASES research found that only 20% of consumers over the age of 55 said they have used a swiping app in the past, and they are an essential segment of the consumer population. These high-spending consumers are critical to predicting the success of innovations given they roughly account for more than half of all U.S. spending. Further, only half of consumers were able to accurately complete the full practice exercise, calling into question the claims of swiping intuitiveness.

So, what’s the bottom line? Though swiping delivers interview time-savings, BASES didn’t find any evidence that swiping offers fundamentally better results than traditional survey research methods. More importantly, it decidedly does not tap into consumers’ non-conscious processing to deliver insights on System 1 responses. And while it may offer an intuitive interface for younger, dating-app-savvy consumers, it comes at the expense of older consumers. At the end of the day, great research needs to use the right tool for the job, and there are too many variables still at play to consider swiping a reliable path for making innovation predictions.