Les Binet and Sarah Carter get a little bit angry about some of the nonsense they hear around them… like the idea that details don't much matter.

We attended two fascinating research debriefs recently… for two different versions of the same TV ad. The brand was a health product, and the advertising idea was a simple one. Each ad showed real people who suffered from a particular health problem, and in each ad we heard how they had tackled the problem successfully using our brand. The ad was designed to inspire viewers by showing them someone they could relate to who had used our brand to solve the problem. The desired response was: 'If they can do it, so can I.'

The ads that went into testing were both finished films. They were identical in every way but one: same length, same script, same music, same narrative structure, same product sequence and so on. They differed only in the choice of the lead character.

The first ad featured an attractive middle-aged woman. This version failed miserably in quantitative testing. People understood the ad well, took out the product's key health claim, and logged the brand at high levels. But it totally failed to engage and inspire people, and so they showed no particular evidence of wanting to try the product. Further analysis of their responses suggested that there was a slight problem with the female 'hero' – she seemed just a little too cool and perfect. People found her hard to identify with, and found it hard to believe she really suffered from the health complaint we were talking about.

The second debrief featured a man as the hero, and the response was very different. Same understanding, same recall of the health claim, same levels of branding. But this time people found the ad inspiring and engaging, and felt that 'If he can do that, I can too.'

Close analysis revealed that this wasn't a gender issue – the health issue in question tends to affect men and women equally. Rather, it was a subtle difference in personality that swung it. This guy was just a bit more down to earth and warm, and viewers of both sexes felt they could identify with him more.

This example illustrates brilliantly how in the real world, small, seemingly irrelevant details can have a big effect on how people respond to an ad. The traditional model of advertising assumes that people process advertising messages in a fairly rational way, and it is the persuasive power of those messages that determines whether or not they buy the product. But behavioural economics paints a very different picture.

People cruise along on autopilot, making most judgements and decisions at a semi-conscious level based on heuristics. They react to characters in advertising on an intuitive, emotional level, and these gut reactions colour how they process everything else. Psychologists call this 'the halo effect', and it means that who says the words can matter much more than what they say.

And this is just one example of a broader 'affect heuristic', whereby positive emotional responses to individual executional details get transferred to the advertised brand. So small changes in lighting, photography, editing, etc. can be more important than the advertising message. A good example is the choice of music, which usually has little to do with the rational content of advertisement, but can improve the RoI by 20%—30%.

This has huge implications for research. If advertising is primarily about communicating ideas and messages, then animatics should be just as good for testing as finished films, and this is what we are routinely told. But if advertising effectiveness lies in the details of the execution, then animatics may not be good enough. It also raises some interesting questions for us all about strategy and briefs. How do we plan for effective advertising when so much of the RoI power lies in the execution, not the message?

It seems that in advertising, as in so much of life, the devil is in the details.