How data can put the emotion back into advertising
A view from Leila Seith Hassan

How data can put the emotion back into advertising

Rather than being polar opposites, data and emotion are now working together to help us understand our audiences more effectively.

In the past 80 years, we’ve seen emotion in advertising shift from the highly scientific understanding of a few (in, let’s face it, laboratory conditions) to the highly rational evaluation of the many. But this year, we will finally see the merging of the two. 

Data and emotion, or analysts and creatives as I’ve been told, are often cited as polar opposites in advertising. But, thanks to recent advances in technology that help us to better understand language and context in people’s communication, we’re arriving at the era where brand activity is informed by data that builds emotional appeal.

A large proportion of recent research has focused on advertising working so well because it inspires emotion in people, and emphasising the corresponding decline in creative effectiveness for brands that ignore this. 

Until very recently the industry has attempted to understand which emotions to target and those that work for a brand through very traditional approaches: focus groups (tried and tested since the days of Mad Men); ethnography studies; behavioural science/behavioural economics; psychology and desk research. It was scientific.  

It’s clear then, that emotion and brands have a long history.

As an analyst, however, the process hasn’t been very "data driven". And that’s always been OK because no one ever really thought that data could adequately understand, capture and quantify emotion. Data was often seen as rigid and inflexible. Not based on feeling. Not based on hearts. Not emotionally messy. 

However, it’s true that in the shape of the ‘data-driven marketing’ that began exploding some time in the 2000s, we’ve long been using statistically robust research techniques to identify groups of similar consumers and their needs, motivations, wants (a.k.a segmentation). The maths had been around for a while, and was by no means new in the new millennia. But we now had the data and ability to do something faster, easier and cheaper. 

This changed direct mail, outbound telemarketing, email, then digital advertising, social and CRM, because we could create strategy and creative based on segments. But I remember the backlash working with creative teams – the feedback being that the data was so rigid that it was caging their ideas. Data was for validating and testing, not liberating or inspiring.

But, in 2019, things changed. Rapid advances in technology allowed us to understand emotions in a very detailed, data-driven way. Advances in cloud computing, providing the ability to access state-of-the-art computing resources to build, train and validate our models, has meant that we can work with large and complex datasets in a cost-effective manner.

Google’s AI Hub lets us transform our models into services so that they can be used at scale. And, finally, recent breakthroughs in Natural Language Processing (such as the BERT transformer) have increased our ability to accurately detect different types of emotion and reduce the time it takes to model unseen data. 

This has several important implications for brands because we can use algorithms to understand what people feel when they talk and write. In doing this, we can remove the biases and preconceived ideas/journeys about our consumers by allowing (albeit complex) maths to identify the topics, ideas and pain points experienced by our audiences. Then we can deploy AI to align our consumers’ needs and desires to a full spectrum emotion, steering us into accurate emotional trigger points.

Beyond this, brands can monitor their emotional impact and that of their competitors, and we can also steer the right content to where consumers are emotionally versus in a point in time in their journey.

The biggest impact we’ve seen so far is for any advertiser with a call centre, in terms of reducing the amount of resource they need to allocate – they’re able to create experiences, content and journeys that prevent consumers needing to call in. For those who do call, teams can respond more efficiently, with better customer care.

Brands in the banking and travel industry are also using a data-driven approach to emotion to shape both their communications and product offer – by understanding and predicting when their customers are most likely to experience emotions such as joy and anger. 

And, in terms of the creative process, creatives are much more onboard than six or seven years ago. Due to the strength of the data and analysis they know it’s there to confirm what they feel in their gut, or to give them a new perspective on what people really feel beyond the agency bubble, rather than to cage their creativity. 

With all these positives in mind, it’s clear to me that 2020 will see a significant rise in data-driven, emotionally focused creativity.

Leila Seith Hassan is head of data and analytics at Digitas UK

Topics