Balancing creativity and brand safety can be difficult. Today’s advertising must stand out; bland does not sell. Marketers must push the boundaries and create content that earns attention, while navigating cultural nuance and potential consumer backlash. However, it can be difficult to chart a safe course because it is hard to predict exactly what will give offence and to whom. Fortunately, today’s automated research solutions allow marketers to get fast and actionable feedback on how people respond to their advertising.
Vocal minorities can trigger unexpected negativity
In the Czech Republic, Vodafone wanted to portray a real family in their advertising and found the Budimans, a single mum bringing up her two sons who are half Czech and half Indonesian. But when the content launched, internet trolls targeted the sons with intolerant commentary. Vodafone stood firm against the hate, believing that the negativity came from a vocal minority, unrepresentative of broader society.
Vodafone’s decision was motivated by what they believed was right. But for those marketers who don’t want to worry that they may have misjudged popular opinion, timely consumer feedback can provide helpful guidance.
Pre-testing ad campaigns anticipates people’s emotional response
Pre-testing content before launch can provide insight into how the wider majority is likely to respond, but it is important to understand people’s instinctive emotional response, not just what they say. An important tool in understanding that emotional response is facial coding, which classifies peoples’ emotions by tracking their facial expressions in real-time as they view a video. Kantar partners with Affectiva to provide advanced facial coding technology as part of the LINK ad testing suite. The combined solution provides specific guidance on optimising creative by opening a window onto consumers’ unfiltered, moment-by-moment, emotional reactions to video creative.
Make informed advertising decisions
Automated market research platforms like Kantar Marketplace allow brands to quickly test ads before launch to identify how people will respond, and to refine and improve the execution. And solutions predictive of sales like LINK offer useful feedback when advertising triggers unexpected negativity, giving insight into how the silent majority feels, not just the haters.
But whether testing before or after launch, it is critical that the test accurately reflects the breadth and depth of people’s emotional response.
Go wide: Ensure you capture the diversity of response
In today’s diverse world, it is critical to ensure you obtain responses from a good cross-section of people. But representation is not just a matter of sampling but measurement too. To ensure that expressions can be recognised irrespective of gender or ethnicity, our facial coding partner Affectiva has re-trained their A.I. to better identify the emotional response to advertising using Deep Learning for greater accuracy and consistency across demographic groups. Based on 5 billion facial frames, 90 countries, and 11 million faces the coding system can now recognise a smile with an accuracy of 95% or higher for each of five groups: Africans, Caucasians, East Asians, South Asians, and Latins. Better representation means less chance of missing a minority response.
Go deep: Make sure you get the depth of insight you need
Facial coding classifiers need to capture a wide range of emotions if you are to properly understand how people are responding to your content. To that end, improvements in facial coding now better discriminate between different emotional states.
Being able to identify sentimentality – that happy-sad feeling – and distinguish it from a regular smile is important. The Starbucks “I am” ad from Brazil, winner of the Glass Grand Prix at Cannes, invites trans people to have their names legally changed at a place where they are always welcome (their local Starbucks, of course). When tested using LINK on Kantar Marketplace, the response to the ad was highly sentimental, with that emotion peaking as the video showed people leaving the Starbucks with their new legal papers in their preferred name.
Two other emotional states that have been difficult to differentiate between are confusion and concentration, making it tricky to identify whether someone is trying to figure things out or is genuinely discontented or confused. In the case of Nike’s Dream Crazy, featuring Colin Kaepernick, many people frowned when watching the video, but we can now tell that the brow furrow observed implied concentration for most, not confusion or negativity.
Go long: Go further when a need arises
Comparing and contrasting responses by different audience segments is just as important to facial coding as any other survey question. This is why Kantar Marketplace is launching the capability to segment facial coding traces, at the time of testing or later, as the need arises. Whether it is to better understand polarisation of emotional response or diagnosing poor brand association (a critical component of ad effectiveness), it can be helpful to compare traces to identify key transition points within the flow of an ad.
For example, one ad tested suffered from weak branding. Facial coding identified an increase in frowning at key points in the video among people who did not remember the advertised brand, indicating increased mental effort as they tried to figure out what was going on. Subsequently, two edits were made which much improved the video’s brand association.
Navigate cultural conflict with confidence
The need to balance compelling creative ideas with the risk of consumer backlash is getting ever greater. But just because the audience responds based on gut instinct does not mean you need to make decisions the same way. Fast, agile pre-testing with enhanced facial coding can help you navigate cultural conflict with confidence.
Book a demo to see how the enhanced facial coding on Kantar Marketplace can help you develop more culturally relevant ads.
