At the SGInnovate event in Singapore, advertising, behavioural science and AI experts gathered to discuss the role machine learning and big data will play in helping brands access consumer insights and overcome bias.
According to Christopher Graves, founder of the Ogilvy Centre for Behavioural Science, our current communications methods are akin to “doing brain surgery with a chisel”. There’s no point in sending out ads and trying to change behaviour unless we understand what drives and influences a person’s existing behaviour.
“The tools to do that are here,” he says. “We’re able to decode human beings by personality and worldview, which is hugely important if you’re trying to change behaviour.” The benefits of this can be applied to all spheres: health and wellbeing, crisis management, employee engagement, and of course, advertising.
Our data is our DNA
A testament to the sophistication of these AI tools is that the single best predictor of heart disease isn’t smoking, nutrition or sedentary behaviour, but language used on Twitter. Natural language processors are able to parse combinations of words that parlay to certain behaviours, cross-reference with CDC data, and find correlation.
What began as simplistic segmentation by age, gender and income is now forking off into a spider’s web of personas, preferences and predictors. “I think we’re going to see the end of billion dollar brands, and instead have hundreds of million dollar brands, and that’s because consumers are looking for more micro-personalisation,” says Som Choudhuri, co-founder of AI Palette.
That doesn’t necessarily mean brands or agencies will need to exhaust their resources making 7 billion different versions of the same content; for the most part, consumer personas will still fall into a finite number of clusters. As Graves puts it: “A huge falsehood is that everyone thinks they’re an individual.”
Consumer research can still be inaccurate or untrustworthy
While AI can yield deep consumer insights, we have to be sure that consumer research isn’t skewed by confirmation bias. “I believe almost nothing that comes out of focus groups or consumer research,” says Graves. When we see part of the brain becoming active at the same moment that the subject is presented with external stimulus, it’s common to assume causation, when it might just be correlation. That must always be a caveat to be kept in mind, especially in consumer research where there is often a goal or agenda.
And then, of course, there is the unconscious bias of the consumer, who is an unreliable narrator. The amygdala, located deep in the brain’s medial temporal lobe, feeds fight-or-flight responses to the ventromedial prefrontal cortex, which is more evolved and able to make sense of that raw data, but the consumer is still largely ruled by the former. “Emotion is ultimately the boss for all decision-making, and reason is the press secretary,” says Graves. “You might buy a car because it makes you feel ten years younger, then the press secretary comes out to justify that by explaining the great deal you got.”
Graves sees the future of consumer research as a learning curve punctuated with errors and overreaches. “I think we’ll go too far, then have to rethink,” he says. “It may be overturned by newer science, and you’ve got to have the stomach for that. With the science that will work, instead of thinking in an arrogant way that we’re going to cook up the best ad ever, or we’ll win a Gold Lion for a campaign that never sold anything, I believe we’ll think more like an empathetic doctor. If I want to enact this behaviour change and shift mind-sets, I need to understand what the real biases and barriers are.”
Behavioural science can drive omni-channel excellence
As these consumer insights increase in complexity and detail, they will enhance consumer interactions both online and offline, for traditional media companies and digital native brands, for ad-driven businesses and subscription platforms.
Rob Gilby, founder of Blue Hat Ventures, cites the example of Disney, who have incorporated a greater degree of personalisation into their theme park passes to deliver a better guest experience. Elsewhere in their business, Disney are working on facial recognition in cinemas, analysing which parts of a movie make audiences laugh or cry. “While not a commercial product yet, this capability will help them predict audience reactions and create better movies,” he says.
We’re seeing this across every medium, from intelligent script analysis and AI-informed casting in China’s movie business, to text created by machines that’s practically indistinguishable from human authors, and even visual content, with machines designing hi-res humans. Sound is another arena with great potential for brands to be of service, says Graves, predicting that soon Spotify will be able to tell you when you’re five days away from your next depressive episode, due to your choices in music right now. “As the data gets richer and richer, the quality of those very human nuances will get better and better,” says Gilby.
to News Alerts