Don’t say what you think, but do what you feel

David Ogilvy once said that the problem with market research is that people don’t think how they feel, they don’t say what they think and they don’t do what they say. Here, David Brennan tries to untangle the contradictions…

Anybody who has seen me present will know that my favourite work-related quote ever comes from the pen of advertising guru David Ogilvy who, in ‘Confessions of an Advertising Man’ opined that; “The problem with market research is that people don’t think how they feel, they don’t say what they think, and they don’t do what they say.”

As I was still very early into my media career at the time I first came across that quote, it held a particular resonance.

Part of the reason was that I had recently completed my degree in Psychology and there was an assumption that human beings were fascinating creatures that were far more complex than most of the prevailing theories could truly represent, and that clever ways of uncovering their mysteries (e.g. the Stanford ‘compliance’ experiment; the Prisoners & Guards study) had to be constantly devised.

So, I always felt quite uneasy with the simplistic way we would collect answers from fairly anodyne questions and assume they were a true reflection of the consumer’s world; the Ogilvy quote seemed to brilliantly encapsulate my unease, while also offering the clues as to how it could be eventually addressed.

The David Ogilvy quote has stayed with me ever since, as the more traditional Q&A methodologies continuously struggled to make sense of the data and new paradigms were introduced from the worlds of cognitive psychology, neuroscience, statistics, behavioural economics and social anthropology, to name but a few.

As new methodologies developed out of these academic spheres, suddenly we had a way of understanding consumers as something more than the rational, linear, mono-dimensional, self-interested beings we had inferred from the research (and much of the marketing theories around it). We began to realise they were emotional, implicit, multi-dimensional, whimsical, instinctive, intuitive… and far, far less predictable than we had always assumed.

So, to truly appreciate the prescience of David Ogilvy’s words, we need to deconstruct his quote into its component parts.

People don’t think how they feel

It is strange to think that, until less than a decade ago, the idea that our emotions were more influential in our eventual behaviour than our conscious, rational thought processes was considered outlandish.

Perhaps that has something to do with the times; before the economic crash, cold-hearted classical economics still ruled the roost and the emerging digital industry appeared to reflect this thinking with its simplistic response models and binary view of the existing media landscape.

We now know that our implicit, intuitive, emotional brain has a much greater part to play in advertising success than we ever could have predicted. We also know that the thinking, rational brain tends to act more as a post-rationaliser than a driver of behaviour. To quote the eminently quotable Rory Sutherland;

“The conscious, rational brain is not the Oval Office; it isn’t there making executive decisions in our minds. It is more like the Press Office, issuing explanations for decisions we have already taken.”

The problem would be less challenging if people could simply report their emotions and implicit processing via established research methodologies; the problem is that they cannot. Not only is the implicit mainly below our level of conscious awareness, but according to leading clinical neuroscientist Professor Richard Silberstein, much of it is inaccessible to the parts of the brain that are active when we are responding to market research questions.

In other words, not only do we not think how we feel, but we can’t directly or consciously communicate how we feel.

People don’t say what they think

We have always known that there has been a disconnect between what people really think and what they say in research interviews or focus groups.

The issue of self-representation has been known for a long time. It is the main reason why research respondents over claim on the amount of documentary programming they watch, or their readership of quality newspapers over tabloids.

The media industry has long been aware of this, which is why we tend to play down the responses in favour of ‘quality’ content and upgrade the numbers who claim to prefer more populist fare.

I distinctly remember, in the run-up to the launch of the UKTV channels, the research telling us it would be UK Horizons – the documentary channel – that would sweep the ratings, whereas UK Style – the reality/lifestyle channel – would struggle to bother the scorers; of course, UK Style ended up generating double the audience share of its upscale sibling.

Of course, there are many other reasons why respondents might not say what they think. One of my concerns with the increasingly self-selecting and incentivised world of online research is the possibility that respondents might just be saying what they think they can get away with in order to keep each survey they complete as short and financially rewarding as possible.

Or the context inhibits them; focus groups are a particularly vulnerable methodology in this respect, with people who often have a really interesting viewpoint being overwhelmed by more forceful or extrovert group members.

Perhaps they don’t know what they think, but feel they ought to offer an opinion nonetheless. Or what they think has changed from one moment to the next; we know that human thought can be highly fluid, if not mercurial.

There are myriad reasons why what people actually think, and what comes out of their mouths or keystrokes, ends up taking us in very different direction. That is why many of the more rational-based research methodologies have been straining so hard to remain relevant in recent years.

People don’t do what they say

What people do is, ultimately, the endgame for market research; we are all trying to monitor, diagnose or predict consumer behaviour in some way or other. The problem is that the gap between what people do, and what they say they do (or are going to do), is perhaps the biggest gap of all.

If people did what they said they would do, intentions to purchase would be sufficient to define every future consumer market, but it is actually quite a poor predictor of actual behaviour.

There are papers going back to the early 1970s identifying the differences between saying and doing; and if intentions failed to match actions back then, the differences are likely to have increased in line with choice, technology and complexity.

For example, the ‘purchase journey’, which must have seemed so simple to those 1980s marketing theorists, can now be seen to be less a funnel and more like a pinball machine, pinging us all over the place. So, it is hardly surprising that people’s stated intentions are often laid by the wayside.

The same is true for media usage. Perhaps the best example can be seen in the shockingly inaccurate predictions about the death of broadcast television; indeed, just about any established medium.

Respondents would say they had watched far more on-demand, or non-broadcast channels than they had, and far less ‘traditional’ TV. In fact, a combination of poor research design and faulty, skewed memory processes produced a series of outlandish claims from the digital disruptors that simply never materialised.

In part this was a result of asking skewed samples of context-sensitive questions and applying misguided analysis to generate misleading conclusions; but it was also due to respondents being much more likely to remember that relatively novel interactivity, and being completely incapable of accurately representing the hours of broadcast TV they would routinely watch in all of the surveys in which they participated. They said that they watched hours more on-demand or online TV at the expense of broadcast, but that was a far cry from what they actually did.

Cutting out the middleman

So, what to do in the face of so much confusion and contradiction? The obvious answer is that we should cut out the middleman – getting people to say what they think – and instead concentrate on the two extremes – how they feel and what they do.

How they feel provides the engine for future behaviour; after all, one of the standout findings from the IPA’s analysis of 30 years of advertising effectiveness awards entrants is that whether or not people like a campaign is a much better predictor of the brand’s performance than many of the traditional measures, such as awareness, recall or even intentions to purchase.

The problem is, of course, that so much is invested in getting people to say what they think; an analysis by a major research agency recently suggested that this was responsible for well over 90% of global research spend.

Methodologies aimed at measuring our emotions or implicit attitudes are still considered somewhat ‘left-field’ by many in the industry, although the body of work identifying the closer relationship between these measures and subsequent consumer or media behaviour is growing impressively.

Meanwhile, what about behaviour? Surely, we have more and better measures of what people actually did than at any time in modern history. Of course, that is true, but the data often comes from diverse sources; is mainly based on the behaviours rather than the people or influences behind them; is often less than comprehensive; and can occasionally be contradictory, as anybody with a sensitive analytics dashboard will tell you.

That said, the increasing convergence of behavioural data and in-depth quality of insight into the influences behind those behaviours means that we are more able than ever before to cut out the middleman, relying less on asking people to say what they think and more on monitoring what they do and how they feel.

It might even have persuaded Mr. Ogilvy that this is no longer “the problem with market research”, and could be its saviour; especially in the fields of advertising and media research.