Ninety-five per cent wrong

It has become a mantra that cognitive science 
is challenging the assumptions of our industry. 
It is, but not necessarily in the way people assume.

The typical story is that breakthroughs in behavioural economics, neuroscience and the study of heuristics have debunked the idea of the ‘rational’ consumer. These suggest that we think less than we believe we do and only engage our more effortful (slow) processes when necessary. An often-cited ‘fact’ is that ‘95 per cent of what the brain does is unconscious’, implying that 95 per cent of brand decision-making is unconscious.

There are two problems with this. Firstly, few 
in the research or marketing industry have ever believed in a dispassionate consumer. In addition, the idea that 95 per cent of brand decision-making is unconscious is every bit as wrong as the idea
 of an objective consumer.

For a start, the figure is a hypothetical construct. It is impossible to find scientific papers that evidence this number and while it is true that 
a huge amount of the brain’s energy consumption is spent on processes below our awareness, much of that is involved in biological regulation and low-level sensory processing – which is not the same as decision-making.

Furthermore, if the 95 per cent figure were 
true, then when we measure brands using research tools that tap automatic, non-conscious processes directly, we would see a seething mass of associations and responses. We now have 
such tools, and Millward Brown has deployed them on hundreds of projects. What they show
 is that many brands evoke modest emotional responses (typically neutral or mildly positive) and just a handful of ideas or responses, unless people engage in effortful thought. Similarly, when automatic responses to brands are modelled against behaviour, the degree of influence 
is way below the 95 per cent we’d expect if that proportion of decision-making were driven by unconscious processes. In the case of budget airlines, intuitive reactions to carriers were shown to have very little relationship with behaviour,
 as people were more willing to override their instinctive responses for the sake of a good deal. For a low-risk choice like social media usage, intuitive responses to brands such as Twitter did show a significant relationship with behaviour, but only to the same degree as explicit attitudes.

These findings are consistent with the cognitive science that sparked the whole discussion. The core of the science on ‘fast’ and ‘slow’ thinking is that both contribute to decision-making, with fast processes such as instinctive emotional reactions framing the slower ones. Where possible we will use our fast processes for decision-making, but our slower processes can and do override them when we are motivated to do so or when our instinctive responses are vague or muted.

This is the nub of the challenge for brands, because the evidence is that many brands lack the instant meaning that would allow consumers to make quick, effortless decisions that they are wired to favour. This makes them vulnerable to competitors that do create an instantly relevant and positive meaning, and a key objective for brand managers is to create not just rich brand equity, but equity that is immediately accessible.

The fact that many brands lack a deep meaning for consumers also has implications for research. There is a tacit assumption in the industry, evidenced by ideas such as the iceberg metaphor, that it’s there to dig beneath the surface to reveal the rich, subconscious world of brands. A better metaphor is that of construction. Most research methods encourage people to reflect on their faint brand associations and to see what they can build from them, there and then. Do those ideas mean the brand has a particular personality, or will work better on a particular dimension?

This perspective has important implications for research:

1. We need to remember that even if respondents are trying to focus on functional or factual information, they cannot help but be swayed
by their automatic reactions. Many research questions encourage more effortful thought than people will commit in reality when making brand decisions. This is still valuable – if people don’t register the intended brand meaning or advertising message when they think about it, then they certainly won’t in the real world. Similarly, brand values, personalities, archetypes and so on are valuable aspirations for a business to rally around in marketing a brand, so understanding if people can construct those ideas for the brand when they think about it has value in measuring progress towards those goals.

2. We need to focus more on gist. It is the gist of
a brand that counts for fast decisions. Research questions are likely to be much more nuanced than the associations consumers have in their minds, so looking across metrics at the gist of the answers will give a more meaningful insight than worrying about whether people differentiated between minutiae like ‘tastes delicious’ and ‘has a satisfying taste’. Millward Brown has introduced questions that focus on brand and advertising gist rather than detail for just this reason.

3. We need to complement established measures with approaches that isolate the instant meaning of brands and campaigns, so we can properly understand the nature and impact of brands’ automatic associations. Fortunately the industry is now adopting such measures extensively, and Millward Brown’s own work routinely includes:

• Measures of the instinctive emotional responses that will frame real-world decisions.
• Reaction time-based measures of intuitive 11 associations, which enable us to understand whether brands have the real-world meaning their owners intend.
• Automated facial coding of viewers’ responses to advertising (used on more than 2,500 ads by Millward Brown at the time of writing).

Integrating reflective thought and measures
of instant meaning is key. By including both, we can empirically determine the importance of each. In the Twitter example, measuring fast responses enabled us to improve our prediction of behaviour by 50 per cent. With facial coding, we can improve the ability of Link survey measures to predict ad sales effectiveness by up to 30 per cent.

Marketers should strive to give brands instant meaning, and as brand and ad researchers, we should be clearer about our established metrics, which often measure what brands might be if people think about them. By taking a more measured view of the science and integrating the tools to isolate fast thinking directly, we can not only understand the true influence of fast thinking, but provide more realistic brand guidance. That’s something to build on.