Watch as Sentient’s Chief Behavioral Scientist Dr. Aaron Reid shows us what market researchers can learn about the measuring the non-conscious impact of advertising from everyone’s favorite illicit drug manufacturer:
Walter White, Heisenberg.
In this talk, Dr. Reid reveals how measuring the emotional impact of an ad, in combination with the consumer implicit connection with the brand, more accurately predicts whether a Super Bowl ad will “go viral” than the USA Today Admeter alone.
Thank you for being here. I want to start off with some numbers that you may be familiar with. This is from the GRIT report, produced by GreenBook. And this is the current adoption of “neuromarketing” techniques (I use that term kind of loosely) within the industry overall.
This is the amount of people—the percentage of companies—who say they are currently “in use” of neuromarketing techniques. You can see we’ve kind of plateaued, we haven’t quite crossed the chasm, from early adoption to the early majority.
But we’re almost there. If we cross 16% we’ll be in the early majority and, from a new product adoption perspective, we should really take off then. But, we’ve been at that stage for quite some time. And when I present these kinds of talks, and Lenny is listening, he usually asks this kind of question:
“When are we going to see broad scale adoption of non-conscious impact techniques? And what is it going to take to get there?”
At Sentient, we think there are three keys to crossing the chasm with non conscious measurement techniques.
The first one is sound science. We have to practice science correctly now, so that we’re laying the right foundation for the industry.
That includes doing non conscious measurement correctly, but it also includes integrating conscious measures correctly. We have to actually stop disparaging conscious measures, because conscious measures are critical for understanding human behavior.
A better position for us, as in industry, within the market research industry, is about appropriate integration of non conscious and conscious measures together.
The second thing that we need to do is market validation. We need to publish validation studies. And there’s at least two kinds of validation studies that are particularly useful.
One is providing evidence that we’ve giving deeper insight in the “whys” behind behavior—so adding to explanations, trying to understand consumer behavior.
And the second one is evidence of how adding non conscious measures increases our predictive accuracy in forecasting actual consumer behavior.
If we can publish more studies that accomplish those two things, we’ll achieve broader industry adoption.
Increase Awareness of Applications
And the third one is, we have to increase awareness of applications.
Most people, when they think of consumer neuroscience, they think of ad testing. Copy testing. Which is a great application. But the applications for non conscious techniques are really vast.
And we need to educate the industry on the ways in which we can actually use these techniques. Hopefully you’ll see all three of those here in today’s presentation.
So, if that’s our goal, what better place for us to start, than in the middle of the desert.
Consumer Behavior Insights from Breaking Bad
[Clip from Breaking Bad plays]
Anybody know what that is? Everybody know what that is? That’s the current trajectory of traditional market research.
No, [laughs] I said don’t disparage conscious measures.
That’s the opening scene from Breaking Bad, right? How many people are currently watching Breaking Bad? I’ll try not to do any spoilers here. Anybody who has watched Breaking Bad?
Alright! Perfect example.
That story, about the transformation of Walter White to Heisenberg—chemistry teacher to meth cook and drug lord—carried five seasons. 100 hours of story. It won over 100 awards. It had 10 million+ views of its final episode.
Why was this story so compelling?
It’s really about the behavioral transformation and the power of the sense of self to transform behavior in extreme circumstances.
You remember Walter in the beginning? (Again, no spoilers.) But you see his gradual transformation into Heisenberg. And his expression of elements of his core sense of self, that had been left unexpressed for 50 years of his life.
This one has a little profanity in it, so you might want to close your ears.
[Clip Breaking Bad clip plays]
You feel that conviction?
That’s his sense of self. It’s so interesting to study the extremes of behavior. And we do this in psychology because studying the extremes of behavior gives us insight on what’s driving behavior within the norms.
And that sense of self is a behavioral transformation mechanism. It regulates our behavior.
Robert McKee, who’s a very well known figure within the story/screenwriting space, requires everyone who comes to his TV seminar, to watch all 62 episodes of Breaking Bad before they come.
Is this just entertainment? Is it just for our, kind of, fun?
We’re going to argue that the core sense of self is more fundamental to human and consumer behavior than you may have previously thought.
One of the core streams of research at the Sentient Consumer Science Research Lab, is understanding the nature of the sense of self, and how that actually applies to consumer behavior.
And the psychological literature and behavioral science literature on the sense of self is really rich. And we tap into that literature to find new insights about how these principles might apply to consumers.
That last one there, is “bad is stronger than good,” by Roy Balmeister. One of the best authors on self regulation.
If we understand the sense of self as it relates to humans, is this relevant to us as marketers?
At Sentient, we find that people who self identify with your brand exhibit behaviors that are very meaningful for your business.
For example, people who highly self-identify with a beauty brand are willing to pay 50% more for that beauty brand versus competing beauty brands. We find that people that highly self-identify with a healthcare brand are wiling to drive 20 minutes across town to get that brand. And people are three times more likely to visit a restaurant brand if they highly self-identify with that brand on an implicit level.
Those are behavioral outcomes that are very meaningful to us as businesses. So one way in which we can illustrate the power of the sense of self and its importance in driving consumer behavior is to do some research on research.
And that’s one of the things we like to do at Sentient.
Analysis of Super Bowl 2015 Ads
We’re going to share with you today a study that we did on the Super Bowl where we measured the implicit impact of all 61 national ads on 50 different brands from this past year.
The brands that we measured were brands that you know: Microsoft, Always, esurance, Pepsi, Coke, Disney, the NFL. There were over 3,000 consumers in this study, and we captured nearly 600,000 measures of system one processing.
The key to this is that we didn’t measure implicit reaction to the ad itself. We measured the change in implicit attitude to the brand or product that was featured in the ad.
And that’s key to applying implicit association techniques properly to ad testing.
Weight Watchers is one that you’re going to see in a second. You might all recognize this particular ad from the Super Bowl.
Always #LikeAGirl Ad
[Always ad plays]
That’s a powerful ad. It’s got an emotional peak. There’s a change. You can feel a change in the music. It takes you on a different emotional journey in the second half.
But it was controversial. It actually didn’t resonate with everyone. Here is the data, the behavioral data, following the Super Bowl, on online behavior toward this ad.
This data is from iSpot.tv, which collects behavioral data relative to ads:
- It had 1.5 million online plays
- It had 630,000 social actions (like shares)
- It captured 11% of what iSpot.tv calls “digital share of voice”
So it was very successful from an online behavior perspective.
But it wasn’t driven by everybody who saw the ad. It was driven by people who had an implicit sense of self connection with the Always brand.
We measured the implicit emotional association with the Always brand, following exposure to that ad, and the Sentient Prime Implicit Index Score was 104.
Now the mean on that scale is 100. So among everybody, this just has a little bit of emotion resonance. But when you measure it among people who self identify with the brand, you get a significant boost in the emotional implicit impact of that ad: the Sentient Prime Implicit Index score goes up to 113.
What if that score was predictive of the online behavior? What if we measured the emotional associations of people who connected at an implicit level with our brand to their sense of self? If we measured that, could we forecast the potential success of an ad?
We’re going to show you that you can.
Here’s another ad from the Super Bowl. There’s a lot going on in the subtext of this particular ad. See if you can tell who the narrator is, of this ad.
Weight Watchers Ad
[Weight Watchers ad plays]
How many people know who that narrator is? Yeah? It’s Aaron Paul. Jesse, from Breaking Bad.
And what’s he doing there? He’s talking about food addiction right? It’s powerful.
The Sentient Prime Implicit Index Score: 107. For everybody. You think, “hmm, that’s not that great.”
Look at what the emotional association was with the Weight Watchers brand following exposure to that ad among people who self identify with Weight Watchers: 128.
This ad connected, at a core level, with the people whom it was trying to connect with: those people who self identify with the Weight Watchers brand.
So once we’ve collected all that data, and we look at the impact of exposure to an ad, compared to a control (no exposure to an ad) we’re able to then, across all of the ads, correlate the implicit lift to the online social behaviors to see how well these measures actually predict some behavior of interest.
This is a best practice in validation.
You don’t want to validate implicit measures by comparing them to conscious measures.
What does that tell you? Do you expect them to correlate? Do you not expect them to correlate? It could go either way.
But if you’re correlating them to some behavior of interest, now you know you’ve got real validation. In this case, we’re using online sharing behavior as our behavior of interest. But you might use sales, you might use retention rates, depending on what your business question is.
If you want more information on this, go to bit.ly/implicitadtesting and you’ll see a review and critique of the application of neurophysiological measure to ad testing on our blog.
Super Bowl 2015 Ad Results
We’ve got all of the behavioral data from the ads, from iSpot.tv, and we want to compare the implicit measures to conscious measures in an additive way. So we also have a really nice conscious measure, that’s collected game day, which is USA Today “ad meter.”
Everybody know that? “Who won the Super Bowl” ad contest? So if we do the correlation of these online behaviors (digital share of voice) with how much people “like” the ad, using the USA Today ad meter, encouragingly, for USA Today, the correlation is a .35. That’s significant!
There’s actually a seven point correlation on that Likert scale in how much people like an ad, and its social sharing of that ad.
In this study we also included a response latency technique, which is called a fast explicit technique. It’s a conscious measures, but its actually a very good conscious measure. And we added that, on top of the Likert scale from USA Today, to see: does it add any predictive utility? It does.
And that’s equally encouraging for those people who are actually using fast explicit techniques as measures of conscious associations. It’s additive to a LIkert scale. That’s good.
But what happens if we add a true implicit emotional lift to this model? The correlation goes up even higher. It’s a .52. That’s additive above and beyond the USA Today score and the fast explicit score.
Now we have a correlation of .52 with real in market behavior.
But here’s the punchline. If you measure the emotional lift among those people who self identify implicitly with your brand, the correlation goes up to a .63. That’s additive above the USA Today Likert scale and the fast explicit measure.
If we can capture the emotional reaction to an ad among those who people self identify with our brands, we’re in a much better position of predicting the success of our ads.
Just in case you think this only applies to ad testing, I talked about awareness of applications, we see the same kinds of results when we’re doing new product sales forecasts.
if you just look at the correlation of overall implicit combined with conjoint sales you get a correlation of .7. If you do it among people who implicitly self identify with the brand that’s launching the new product, the correlation goes up to a .96. That’s forecasting actual sales behavior of products.
We also see it in new concept testing—understanding in your brand extension, resonates with the core brand attributes among those people who self identify with your brand.
This is an example from Doritos, on the back end of Doritos Locos Taco. This was Doritos Loaded. And it had 3x the predictive run rate, because we understood what the lovers of that brand really wanted in the brand extension.
And its also true for brand tracking studies. Incorporating implicit self identification into your brand tracking, measuring emotional association with your brand, better correlates with actual brand performance and sales in the market place.
I’ll mention the fourth key, very briefly. I’d mentioned the first three to start.
Obviously, Sentient Prime implicit association testing is globally scaled.
There is a map, that’s located at whereis.sentientprime.com where you can see everywhere in the world where we’re currently measuring someone’s non conscious associations with a brand, a product, a package, an ad.
It’s a live map. Every time somebody comes into the software, a little green button lights up on the map. (This is a screenshot from that map.) You can come visit us there.
It’s easily integrated with online eye tracking and facial coding, as you just heard Jack from Sticky talk about. An we can incorporate those two into a seamless experience for what your testing needs are.
And its mobile. Thirty precent of all data collected on Sentient Prime is on mobile devices, and we’re seeing that number increase. This is critical for the future of market research.
So where do we end?
Essentially, consumers want you to “say my name.”
And this is more than just about capitalizing on fundamental human nature. Really, what this is, and what our jobs are, are providing meaningful identity markets, for consumers who want to express about their core values, in an increasingly conspicuous consumption world that we live in.
We have a responsibility to produce brands that are meaningful. That allow people to express their self-identity in unique ways.
If that’s our responsibility, we have to make sure that our methods are the best methods for understanding and delivering that sense of self.
I’ll finish with one more video, which I think you’ll probably find amusing.
[Esurandce ad plays]
Fun. Overall it did OK. Among people who self identify with that brand: 109. Not bad.
We’re teaching everybody how to do this. Any researcher in the world can have access to Sentient Prime to incorporate it into your own implicit studies: client side/supplier side.