Kraft taps into emotions to take sensory research up a gear

By Elaine Watson

- Last updated on GMT

Testing testing ..  But how does my product make you feel?
Testing testing .. But how does my product make you feel?
Food manufacturers need to move beyond traditional sensory testing tools if they want to gain a better understanding of what really drives food preference, Kraft has argued.

Speaking at the IFT annual meeting and expo in New Orleans, Kraft Foods’ principal scientist, consumer scientific research, Melissa Knorr, said measuring consumers’ emotional responses to products could explain why two identical-looking products could achieve the same score in acceptability tests but perform wildly differently in the marketplace.

She added: “We were doing some testing on a new formula for an established product and the results of consumer acceptability/liking tests conducted at a central location suggested we had a parity product, so we were feeling good about it.

“But we wanted to get insights beyond liking, so we did some qualitative home tests and emotional profiling over four days and looked at the consumption experience, and we realized we might have a problem.”

How do you measure and describe emotions?

After ascribing emotional attributes to the products, key differences emerged despite the fact that both the new and the old product performed well in liking tests (8 and 8.1 on a 9-point scale), she said.

“Emotional profiling gave us critical direction. Traditional tools were not enough​. We used emotional research to define unique points of difference and create a new hierarchy of attributes that go beyond ‘liking’. The failure of consumers to make an emotional connection​ [to the reformulated product] was driven by changes to sensory attributes we hadn’t measured before.

“In the test product, ​[positive] emotional attributes appeared early but were weak and faded fast, leading to a disappointing experience, so we reformulated again to get closer to the original sensory profile.”

Kraft has been using emotional profiling for three years as part of its sensory and consumer testing work, Knorr told But it was still a relatively new area of sensory science and the jury was still out as to what constituted best practice and which tools to use, she said.

“I want to go quantitative with emotional research, but I would always start with qualitative research. What we do know is that you’ll miss everything if you sit in a backroom typing while consumers are experiencing your product. Traditional research tools may no longer be enough to answer all of our research questions."

Is 6 out of 10 for 'energetic' good?

Silvia King, distinguished scientist at McCormick & Co, agreed that emotional evaluation added a new dimension to the traditional sensory evaluation process.

Citing the example of four fried food products achieving very similar scores in acceptability testing but eliciting quite different emotional responses from guilt to happiness, King said understanding the sensory attributes that correlated with emotional attributes could help firms “drill deeper and improve your marketing”.

However, tools and techniques for collecting, measuring and interpreting ‘emotional’ data still varied considerably, she said, and manufacturers were still feeling their way. “Is six out of 10 for ‘energetic’ good?”

She added: “Emotion testing is still a relatively new approach in sensory and consumer science, but I think we may get to a point where we start identifying best practices.”

Fellow presenter Shane Skillen, chief executive of consumer insight firm Hotspex, added: “Consider emotion when doing your R&D. Without such a consideration, you’re literally missing half the equation.”

Related topics R&D

Related news

1 comment

Basis of Emotional Reactions

Posted by Marian Schwartz,

Experienced emotions are based on beliefs which, in turn, are based on perceived information in conjunction with values. These are "measurable" entities and can be evaluated in an organized manner by currently-available procedures. The task is to identify the many contributing information factors (color, sound, satiety, resemblance to familiar product, etc.) and correlate them with the same individual's values (e.g., low fat is good, sweet is bad for me, familiar is trustworthy).

Report abuse

Follow us


View more