At the 2011 NewMR Virtual Festival, Alistair Gordon of Gordon & McCallum discussed NewMR and facial imaging and how it could be revolutionary for new market research. As with my NewMR presentation, Alistair was a defender of surveys. “Survey are simple (relatively) to do and create. They are cheap ‘per kilo of data’! Clients understand them or think they understand them. Respondents are mostly accurate (they seldom lie). Surveys provide specific, targeted information.” Finally, and especially during the ongoing economic crisis, surveys are simple, scalable and fairly easy to implement.
Unfortunately, Alistair said, when it comes to the “soft stuff” – emotions and reactions – surveys have come under increasing criticism from others for a lack of accuracy and discrimination when assessing emotions. Surveys lead to respondent’s rationalizing answers and require a complexity of dials, sliders and gadgets to even attempt to measure emotion.
Measuring the emotion in people’s faces is an easier answer. It is based on the research Paul Eckman has done into the universality of the six basic emotions that can be recognized in faces: happiness, surprise, fear, anger, disgust and sadness. You can do facial recognition using manual coding if you understand the types of faces. Don’t want to learn how to code faces? The Swiss software firm nViso (Alistair’s client) maps 143 points in photographs of faces to code emotion.
The nViso system works with almost any webcam – “Measuring emotions over a webcam is really sexy!” – providing objective and culturally unbiased measurement across quite different kinds of faces and cultures. It works with faces in real life – whether or not people are wearing glasses, even if lighting in a room is quite low.
You can add facial recognition to any online survey, passively recording reactions without the use of dials or questions. The facial imaging provides “7 data points a second, per stimulus – we once did 7 bank ads in one test.” The data is directly linked to what people saw and can be time-sequenced to the commercials respondents are watching. “The emotions instantly reflect the story in microsecond intervals.”
“Combining facial imaging with survey data adds value to both.” You can compare results by demographic segments, for instance, or the data can be integrated with the respondents’ other survey answers.
The real NewMR is taking “soft research” and making it repeatable. New MR is not just moving from Likert scales to second-by-second direct observation, not just moving from banks of emotion sliders to passive watching, and it is not just removing respondent burden: NewMR “is in fact a genuine revolution offering scale and simplicity.”
In a world with rich marketing imagery, what is truly driving emotion? Which images have the greatest impact? Facial recognition “puts emotion at the center of creative discourse + better diagnostic feedback”, which is important because “emotion drives interest and purchase.”
Author Notes:
Jeffrey Henning
