TV commercial effectiveness predicted by functional MRI What makes for a good TV commercial? The question itself can already be understood in very different ways. What is good? A commercial that wins prizes for creativity or originality? Or does a good commercial simply increase sales or brand value? Obviously it is best when the two go hand in hand. However, while advertising agencies often know very well what makes for an original or creative design, there is less agreement on what the ingredients are of a TV commercial that pays off in the more literal sense. Here, we use brain scanning techniques to get a grip on that more elusive feature of TV commercial making: how to get consumers off their couches and run towards shops. We have looked at the neural signature of effective ads, using functional MRI, a technique that allows for the recording of deep brain structures involved in (partly unconscious) emotional valuation and decision making. We find a distinctive pattern of neural activity that is exclusive for effective TV commercials. Moreover, using this pattern as a template, we are able to predict whether a TV commercial will be effective or not. Using this technique, we even predicted one of this year s Effie awards. The elusive ingredient of effectiveness In the Netherlands, prizes are awarded for TV commercials on the basis of effectiveness, popularity, and annoyance. To win an effectiveness award (the Effie ) advertising agencies send in their TV commercials together with evidence substantiating the claim that the commercial led to an increase in sales, recognition, or valuation of the brand or product the commercial was for. A professional jury looks at each case and decides who wins. Popularity awards (the Gouden Loekie ) are voted for by the general public, and winners are announced in a festive program on national TV. Awards for the most annoying TV commercial (the Loden Leeuw ) are also voted for by the general public, and are announced in a consumer interest TV program. Although each of these awards are based on a variety of aspects, they also form distinct categories. Gouden Loekies are typically won by TV commercials with high GRP that are funny, cute or otherwise likeable. Loden Leeuwen typically are high GRP ads that annoy, for all sorts of reasons. The Effie winners are a much more difficult to describe category. Their distinctive feature obviously is their capability of increasing sales or brand loyalty, but it is not so easy to get a grip on how this is achieved. They are often not particularly funny, sometimes they are aesthetically pleasing or technically well made, but Effie awards have been won by very non- outspoken TV commercials. Effective TV commercials have a unique neural signature Over the course of our neuromarketing research, we have looked at the response of the consumer s brain to 114 TV commercials. Eight of these were Effie winners, many of which we have recorded multiple times, resulting in a total of 17 Effie recordings in more than 200 subjects in total. Similarly, we have 20 Gouden Loekie recordings (8 unique ads), and 20 Loden Leeuwen recordings (9 unique
ads). We first conducted a whole brain analysis on what makes an effective TV commercial different from other award winning commercials. The results (figure 1) show that Effies stand out on the basis of their ability to activate brain regions such as the orbitofrontal cortex and the striatum, known to be involved on the valuation of reward, expectation and desire. Figure 1: The neural signatures of effective (blue), popular (red), and annoying (green) TV commercials. Note how effective commercials (blue) activate prefrontal and deeper subcortical regions associated with emotional valuations such as expectation, reward, and desire. Popular commercials (red) activate visual association areas and parietal regions involved in attention. Annoying commercials (green) are characterized by a strong activation of early visual and auditory areas, indicating they are not just metaphorically loud (example with 8 TVC s of each class, 16 subjects, recorded on a 3T Achieva TX, 32 channel headcoil, Philips). To quantify the differences between Effies and other award winning commercials, we analyzed them using our standardized Brain Rating procedure (fig. 2). In this analysis, the activation of neural networks associated with 13 different emotions and valuations is given a standardized value (a number between 60 and 140, comparable to IQ). Effies stand out from the other types of award winning TV commercials because they evoke higher levels of positive emotions like desire, lust, expectation, and trust, while at the same time evoking lower levels of negative emotions like fear, anger, disgust and danger. Another striking feature is that they do not attract a lot of attention and do not evoke much activity in neural networks that mediate personal involvement (which appears to be the defining feature of popular commercials like Gouden Loekies).
What do we learn from this? Most remarkable is that effective commercials do not need to draw a lot of attention or personal interest. That explains why they are often quite inconspicuous. Negative emotions are typically avoided. Eliciting desire and expectation seems most important, together with evoking a sense of trust. In our sample, we have seen these emotions for example being brought about by portraying craftsmen or people in a family setting, and by using easy listening type music. Filming is generally of high quality and the cutting is of medium to slow pace, all contributing to a composed rather than exhilarating quality. Lust Expectation 120 Trust Desire 110 100 Value Attention 90 80 Involvement Novelty Familiarity Fear Danger Anger Disgust Figure 2: Brain Rating of effective (blue), popular (red), and annoying (green) TV commercials (weighted average of 17-20 TVC recordings of each class in ~200 subjects. For a description of the Brain Rating methodology and the emotions characterized see www.neurensics.com). Can effectiveness be predicted? Now that the neural and emotional signatures of effective TV commercials are known, can this be used to predict effectiveness in advance? Testing commercials prior to airing is fairly common- practice, which makes sense because the costs of airing usually outweigh the costs of production by a factor of 10 or so.
We analyzed whether pre- testing for effectiveness is possible. We constructed templates from the full set of effective, popular and annoying commercials, constantly leaving out one of them. The pattern of neural activity of the left out commercial was then correlated with that of the three templates (fig. 3). Correlation values were fed into a classification algorithm to categorize each commercial as either effective, popular, or annoying. Categorization was particularly successful for the Effie commercials: 15 of the 17 Effie commercials were indeed classified as being effective (88% hit rate), while 35 of the 40 non- Effie commercials were classified as not effective (87% correct rejections). This included correlations between tests of the same commercial but recorded in different pools of subjects (usually ~25 average consumers). Such test re- test classifications are typically identical (92%). But even if these commercials were left out, 6 of the 7 Effies (86%) and 10 of the 13 non- Effies (77%) were still correctly classified. Classification of popularity or annoyance, however, was more or less at chance level in that case, indicating that Gouden Loekies and Loden Leeuwen form a more heterogeneous type of commercials than Effies.
Figure 3: Leave- one- out correlations for Effies and non- Effies show that effective commercials (blue circles) stand out as a distinct category from non- effective commercials (red and green). Effies (blue) have high internal correlation and are negatively correlated with the other two types of award winning commercials (red and green). Note that the graph depicts both test re- test correlations of the same commercials tested in different groups of subjects, as well as correlations between different commercials of a particular category, tested in either the same or different groups of subjects. Pushing the right buttons In sum, we show that effective TV commercials evoke a specific neural signature and stand out as a distinct category when rated via an fmri based neural analysis. This allows for the successful pre- testing of the effectiveness of a TV commercial. Two conclusions can be drawn that are of particular importance to the neuromarketing field. The 92% correct repeat classification tells us that recording from ~25 subjects yields sufficiently stable results to reliably classify a TV commercial. That classification of different Effie commercials is almost as good (82%) tells us that from a neural perspective effective commercials form a very homogenous category, all pushing the same buttons in our brains. We now know which buttons those are. A final proof of this pudding came recently when the 2012 Effie awards were announced. One of the awarded commercials (Eneco Dutch wind ) was tested by us about 6 months earlier, where we classified it as effective. It won bronze medal (gold and silver medals were not tested by us). Similarly, this year s Gouden Loekie (Telfort Calling as long as you like ) was indeed previously classified by us as a popular TV commercial. Prof. dr. Victor A.F. Lamme (head of Science, Neurensics) Dr. H. Steven Scholte (head of Research and Development, Neurensics) Neurensics, Westerdoksdijk 40, Amsterdam, The Netherlands www.neurensics.com 18 dec 2012