Emotional marketing, neuroscience and the evolving art of sentiment analysis
In this guest post, strategy consultant for Carat, Catalina Burge, claims that emotional marketing is here and measurable, we’re just waiting for the most effective way to measure it.
I read with great interest Alex Vishney’s recent piece on the increasing emphasis on emotional marketing and the subsequent demotion of rationality as a driver of consumer behaviour.
In it, among the strong case he makes for there still being a time and place for rational comms – a contention I definitely agree with – he raises a very interesting point: that the way we measure and use emotions within product categories may be a little ‘off the mark’.
He’s absolutely right, and that should be of concern for any marketer, especially those about to dive headfirst into their next 60-second tear-jerking TVC. Don’t get out the Kleenex just yet.
See, the challenge is not in the creating the emotion itself, it’s really in knowing whether or not you are creating the right emotion to drive your desired outcome. As Alex so rightly contends, the same emotion that is associated with ownership is not always the one most adept at driving purchase.
The prize for triggering the right emotions is certainly tempting – hefty sales, improved brand metrics and brand loyalty – but despite this, effective measurement has remained a persistent challenge.
Emotions, like the humans that have them, are unique and very difficult to predict with absolute certainty. If we as marketers are going to put so many of our eggs in the ‘emotional marketing’ basket, we could do with being a little more rational about it.
How are we measuring emotions currently?
Market researchers have come a long way in their quest for the Holy Grail of emotional measurement.
These days, we are increasingly analysing behaviour rather than stated reactions, and use techniques such as implicit research to gain deeper consumer insights without relying solely on verbalised responses.
Most recently, neuroscience techniques have emerged as the most effective way to avoid the impact of ‘claimed behaviour’, giving way to the concept of neuromarketing.
Neuromarketing sounded promising with its advanced technology and a promise to go inside the brain, so both the global research powerhouses and a swathe of niche start-ups have been working in recent years on the application of neuroscience to marketing.
As Alex touched on, electroencephalogram tests (EEG) capture the levels of electrical activity in different parts of the brain to help determine whether an ad captures attention, which aspects are most effective and, in particular, whether the ad is emotionally engaging, and in what way.
The same approach has been applied to packaging and other marketing communications and mediums.
Take everything with a pinch of salt.
So has neuromarketing made progress? Yes. As we now understand more about how the brain reacts, its levels of engagement and which areas are involved in processing advertising and marketing. Has it gained as much traction as some initially anticipated? Not really, because even this advanced technique remains imprecise and ambiguous when trying to measure emotions.
The main reason, as neuroscientist Molly Crockett points out in her TED Talk ‘Beware neuro-bunk’, is that we need to be rigorous and careful with the way we measure brain activity.
A study can conclude that we love a brand because we recorded activation of the Insular Cortex (a part of the brain linked to love and compassion) but activation of this part of the brain can also trigger a wide range of other emotions like anger or even disgust or pain. Our brain is anything but simple, so we shouldn’t expect simple answers.
Going one step further
While researchers have been busy trying to apply neuroscience to marketing, social media has exploded, providing a new and different source of insight to aid our understanding of how consumers engage with brands, in the form of social listening (sentiment analysis).
The difficulty with sentiment analysis, however, is that it only measures positive vsersus negative sentiment and hasn’t yet explored deeper, more complex emotions such as love, hate or desire.
At least, not until now.
A new approach to sentiment analysis recently came to light with the announcement that Disney Finds Emotion the Next Metric for Advertising. Walt Disney, a company with decades of experience in marketing emotions, is searching for a new metric to measure emotional response via their innovation incubator, Disney Accelerator.
This new approach suggests that emotions are now “communicated through clicks on search ads, movie titles queried, colours in images and videos uploaded to Facebook and Instagram, music downloaded from iTunes and products bought, but clicks don’t measure engagement in an increasingly social world filled with content”.
Disney-funded start-up, Decisive, has gone a step further and begun exploring the realms of artificial intelligence (AI) to analyse real-time data in images and video in order to predict emotional responses on social platforms with the intention of launching the technology in the next few months.
This AI-based approach involves employing predictive analytics to discover what content your fans will be sharing and why – a promising breakthrough immediately applicable to current marketing and media management.
Does this mean that media and social media agencies will soon be incorporating AI and predictive analytics as their next significant data analytics tools? Yes, and probably sooner than we expect.
The ability to optimise social media content based on individual emotions, and drive content development based on emotional effectiveness could be a major step forward.
Marketing and communications management, media buying and planning would all finally begin to measure and manage emotional responses with the speed and efficiency necessary to enable widespread adoption.
Watch this space!
Putting the entire article in the sidebar. What a bold move.
User ID not verified.
I love where this conversation is going on mumbrella.
I would suggest that neuroscience is still at the point of measuring and demonstrating ‘internal response’ as opposed to ‘consequential external behaviour’, albeit more accurately than ever before. While it is possible to track behaviour from prompt through to online purchase in a controlled but relatively artificial EEG testing environment, I’m still not convinced it is necessarily applicable to real world location-based purchasing occasions – whether at home or in a shop.
With regards sentiment analysis and its confinement (at this stage) to measuring valence. What we could possibly surmise from Kahneman and Tversky’s Prospect Theory is that loss aversion (i.e a negatively-valenced framing: ‘Drink this milk or your bones will be brittle’) may be more motivating than its positive counterpart (‘Drink this milk for strong bones’). So any assumption (which is not implied by the author of this piece) that the optimum emotion-type is a positive one might not necessarily point to the most efficacious prompt to purchase.
But, most crucially, this doesn’t account for how well the communication might have been represented; the ‘right emotion’ conveyed poorly would likely be trumped by the ‘not as right but still relevant emotion’ conveyed masterfully.
I would argue that the dominant factor in the purchasing decision outside of habituation resides in the interplay between state and trait at the moment of action. Can a dominant ‘state’ be induced by the right prompt? Maybe. Is this prompt emotional or rational? Probably both, but again I stress that it is just as dependent upon the quality of the communication.
Furthermore, what the emotional component might be comes down to the individual; one person’s cute kitten is another person’s allergy trigger. But then we’ll have individual-centric predictives to account for that soon enough. Unless we already do.
Thank you Catalina for this informative and thought-provoking article.
User ID not verified.
Good morning. If you click on the link the story opens on to a full page. The sidebar is merely our opinion section of the website, offering a home page glimpse of the piece to attract viewers; the link opens on to a full page just like the rest of the site. Hope you enjoy the piece; it’s certainly a very interesting topic.
One hopes that not too many snakes were harmed getting the oil to lubricate this theory.
User ID not verified.
I’m more interested in measuring how people who spend their lives intentionally manipulating the emotional states of people with the sole aim of generating corporate profits – to the direct and explicit detriment of our planet’s ecological health and the mental health of its inhabitants (especially its children) – manage to sleep at night?
User ID not verified.
I feel the main point has been missed: it is not just a matter of assessing emotional involvement and the quality of the emotions (positive, negative), but the marketer needs to know if these emotions are aligned with the brand positioning and, most importantly, if the emotions are linked to the brand or the context within which the brand is being presented.
Take a well known example from the past: Pepsi Refresh offered to spend some $20 million to support good causes and invited the public to list – and vote for – causes. More than 80 million votes were registered and the social media universe was so rapidly expanding with positive comments that it nearly burst. Yet, Pepsi’s market share declined. The emotions were linked to the initiative, not the brand. An assessment of the concept and the ad would have suggested high emotional involvement – but how useful is this?
I think it would have been more useful to at least monitor progress by using a reaction time test that can tell us about the connection between the brand and relevant qualities – and if these connections in the consumer’s mind have been strengthened by the campaign or not…
Anyway, I believe the real issue is not to measure emotional involvement but to understand the specifics of this involvement and the impact of this involvement on the brand. I am not convinced sentiment analysis or AI will address this challenge as the focus remains on the emotions we can deduct from what people say and share, rather than how their brand memory has been impacted by the exposure to the marketing initiative. But I am very happy to be proven wrong! Peter
User ID not verified.
Guess common sense wasn’t so common after all.
Here’s another shiny toy for planners to justify their existence with. Anything to avoid having to come up with some actual solutions to client’s problems using your own mind.
Disney markets values, like all great brands. Emotions come as a result. For brands to try and elicit emotions directly is a huge mistake. People don’t shop for emotions. Our actions, including purchasing behaviour, are motivated by our beliefs and attitudes.
User ID not verified.
@Harmed Snake. Yes it is common sense and it’s something the best creatives and suits have probably understood intuitively anyway, but when this stuff becomes more measurable it makes the sell-in easier. The biggest danger is then relying upon the science too much, however, and using it to determine the solution.
I would also suggest emotion does play a role in our purchasing behaviour – for a random example; chocolate as comfort food for someone experiencing a relationship breakup.
User ID not verified.
I’m interested to know how these research technologies manage to cope with measuring the faint feeling of indifference in which most consumers hold most brands and most advertising.
How do you interpret the slight flicker on the EEG, or the faint glimmer on the MRI?
Must be very subtle gauges they use. Not so much “these go up to 11” as “these are highly sensitive between zero and 1”
Then again, the research business has been strapping consumers to electric appliances since the 1920s. Plus ca change …
User ID not verified.
@Anthony G
Ok, so let’s run with your cliche: chocolate is comfort food for breakups. The question is “so what?” How does that help a marketer trying to differentiate a brand, when:
It’s a category generic, and true of all chocolate?
It’s far from the only comfort food. What about ice cream? Or fast-food?
What’s the relevance to emotions? This is predominantly a cultural insight (even if it does have chemicals that produce dopamin or whatever)
And we don’t need advanced technical equipment to measure these things. If we couldn’t tell that chocolate is comfort food using common sense and experience, it wouldn’t be common enough to have any value.
User ID not verified.
@ Harmed Snake. That chocolate example was my response to the statement ‘People don’t shop for emotions’.
Let me try another with regards emotion and brand preference. Air New Zealand.
My exposure: Never flown them. Never been referred to them by a friend. Can’t remember any conventionally placed tvcs or ads. I’ve seen a couple of funny in-flight safety videos courtesy of articles within news.com.au or theage.com.au, and more recently the sweet ‘Santa Stop Here’ Xmas video courtesy of mumbrella.com.au.
By necessity the in-flight videos have to be rational, yet elicited enough joy to make a positive and long-lasting impression on me. The Santa piece made me sad, in a bittersweet sort of way if know what I mean, but the mumbrella piece also featured a commenter who had used the service and compared it very positively against their experience with other airlines.
As a result of my exposure to this accumulation of emotive (of both valences) and rational prompts, Air New Zealand is now on my shopping list.
We’ve recently had some really interesting articles around psychology and marketing; Alex Vishney on the place of rationality and Catalina on emotions in this context. I am not convinced either is exclusively correct, however I appreciate their viewpoints; cogently considered and well written.
I am already a convert to the idea that improved methodologies can be derived from a deeper understanding of psychology; whether through smart use of data or white hat application of the loosely-defined ‘behavioural economics’. I also understand psychology to be an inexact science, so I do not believe there are certifiably failsafe methodologies.
But in many ways the science is getting ahead of us:
http://wordplay.blogs.nytimes......t-go/?_r=0
https://automatedinsights.com/company/
https://mumbrella.com.au/making-a-murderer-and-good-design-344343
so it’s not a matter of whether we need the advanced technical equipment, it’s becoming more a matter of whether the advanced technical equipment will need us.
My primary concern for the future of marketing is that automation is becoming increasingly compelling to businesses and I still believe in the primacy of human intervention; a fantastic idea by a creative or a brilliant insight by a suit – perhaps derived from the data or perhaps from their own common sense. Or hopefully my understanding of the data as well as my own common sense.
Not sure if that answers all your questions, but doing my bit to progress the conversation.
User ID not verified.
“efficacious”?
Sorry dude, but what’s wrong with effective, a word we actually use?
User ID not verified.