JKR

Weighing the pig more isn’t making it any fatter
It’s a curious paradox that as business grows more confident about the value of design, it’s becoming less confident in exercising judgement about it – to the extent that few organisations will today implement even cosmetic design changes without extensive pre-testing.

The prevailing culture of accountability and management by measurement suggest this trend is set to continue. While it has undoubtedly brought some benefits with it, creating a great body of work for design to be benchmarked against and rewarding the elevated status of design with its own studies, rather than a few minutes discussion at the wrap-up of an advertising focus group, it certainly hasn’t been making design better.

With the recession, many marketing budgets are being pruned, so design testing isn’t receiving incremental investment, but frequently being funded by the reallocation of money away from the insight gathering that might otherwise inspire more transformational design. Put bluntly, the current trend in research is at best making design safer, not better.

The research world shouldn’t regard the situation with complacency, however. The legacy of high profile failures of ‘extensively researched’ re-designs, such as Tropicana in the US and Mr Kipling cakes in the UK and a thus-far more quietly expressed disquiet about disappointing paybacks from positively tested design revisions herald an urgent reassessment of the role of research in design and how to improve it.

Paradoxically, this recession isn’t the time for market research to play it safe, it isn’t the time to be doing the wrong type of research a bit righter, it’s the time to invent techniques that exploit newly available technology and insights into human psychology to help organisations to design better.

Please learn from behavioural psychology
Over the last 30 years the evidence from behavioural psychologists like Nobel laureate Daniel Kahneman has finally persuaded economists to accept that what people say they’ll do and what they actually do are not one and the same. Even when fully informed and presented with seemingly rational alternatives, human beings persist in making irrational choices. Yet it’s as if the world of market research largely remains in a state of blissful ignorance about this.

In the context of packaged goods, this is especially true. We all know that supermarket shopping is both a blessing and a chore. The immense variety presents a remarkable wealth of choice, yet most of us purchase the same things this week that we bought last week, or the week before. Design may be called upon to cement this habitual behaviour or to try to disrupt it. Either way we have to recognise that, as humans, we don’t think as much as we like to think we think. We don’t make conscious assessments before each purchase, we learn to navigate sub-consciously by rote and habit, and we then sleepwalk our way around the store. Packaging is tangible evidence for Robert Heath’s theory of ‘low involvement processing’. Evidence that branded communication can be relevant and persuasive without requiring consciousness. Effective design modifies behaviour without one even realising it.

Why then do so many client research managers and market researchers persist in asking direct, explicit questions of unreliable witnesses about an implicit communication medium?

Qualitative researchers know well the importance of the order effect when testing multiple concepts and seek to avoid it by diligent rotation, yet they are serial perpetrators of a kind of bias effect by declaring an interest in design at the outset of an interview. The consequence of which is that in a desire to be helpful, respondents make a conscious effort to appraise design. This rational-thinking is precisely what we wish to avoid. In practice, all that is needed to capture spontaneous comment is to obscure their motives by reversing the traditional order of questioning. In the sight of mocked-up packs, preferably among their competitive set, a discussion about the category and the brands within it readily takes place. Anything important about a proposed design will surface and if little or nothing is said, that is not an indication of failure. The message received in sight of the design is the important criterion, not the message transmitted.

The order effect is so fundamental that we have witnessed, (suffered), a complete reversal of the conclusion about a proposed design by reversing the order of questioning between groups. Same moderator, same design, same competitive set, same demographic group; rejected out of hand by the early groups, lauded to the heavens by the later.

After 20 years of talking about this issue with researchers and client research managers, it frustrates us that the lesson still isn’t universally accepted. We urge you to send home the respondents when you hear a moderator welcoming a new group by announcing ‘thank you for coming everybody, we’re here to discuss the redesign of’…

Package design is about identification, not communication
Once one has accepted that grocery shopping is low involvement sub-conscious processing, then getting brands noticed becomes fundamental to being chosen, not liking of a design. Clients often seek to deny the implications of this principle, because they would benefit greatly from the use of packaging as a free advertising medium and become over optimistic about the communication of rational and emotional messages at the moment of truth.

Packaging in crowded, competitive categories is generally most effective when it symbolises what makes the brand different and noticeable, rather than framing it in the same manner as its competitors, or providing a canvas for detailed communication. Design provides a shortcut to brand values we hold in our heads. The task of advertising and other forms of communication is to tell you something about the brand; the task for design is to identify it and express consistent values, implicitly.

Consumers do not need to like design for it to be effective. They don’t need to like it for it to help them to see it in the crowded store. They do not need to see explicit taste cues to believe it to be tasty.

Assume Makes an Ass of U and Me
Designers assume that all research agencies devise methodologies to reflect the peculiarity of packaged good purchasing. But they don’t.

Marketers assume that their consumer insight colleagues will adapt the discussion guide for packaging versus more explicit communication research.  But they don’t.

Researchers continue to assume that tried and tested advertising techniques work as well for design. Researchers also assume the shopper is rational [we’re not], that a group discussion in someone’s front room in Suburbia replicates the solo trolley-dash at Tesco [it won’t], that there is a causal link between attractive design and effective design [there isn’t] and that asking consumers what they like will predict what they will buy [it doesn’t].

My plea to researchers would be to address the specific challenges of assessing design and to be more inventive in addressing them.

The future’s digital, but can we have it smarter, sooner?
Independent pioneers like BrainJuicer saw the potential of widespread broadband access to develop cheaper, faster techniques to gauge recognition and liking of packaging with multiple cells, each only seeing one manifestation of each design, combining to transform quantitative testing.

Although online research suffers from 2-dimensional representation of a 3-dimensional form and the limitations of screen size require some caution around interpretation, we are sufficiently confident that the results bear comparison with traditional hall tests to recommend its use to many of our clients. Indeed, with the incorporation of webcam into many laptops, we foresee open-ended qualitative feedback being built into most online quantitative research, giving it the potential to overtake qualitative techniques for standalone projects like design.

If costs come down further, we expect to see even larger panel sizes, allowing us to analyse discrete segments of the population without loss of statistical rigour. However, we mustn’t be blind to the fact that online quant is simply taking a sub-optimal technique and returning it to fashion by removing the main barriers to entry. Like all such developments, this may only be temporary. Sooner or later, its flaws will lead to a return to other forms of research.

Sadly, the faster and cheaper it becomes, the more tempting it becomes to suffer indecision and test alternatives that are, in essence, more or less bad versions of each other. Just because it is a quantitative test, it isn’t infallible – we know of several design evolutions diluted by quant testing to the point that they lost sight of their objective, and others where design that passed the normative concept approval hurdle, but proceeded to precipitate a sales decline in practice.

We contend that the next step in the digital revolution may yet take us in a different direction. Perhaps it’s time to exploit new technology in a return to observational research, whether in-home or in-store? Perhaps the most important lesson of the failed Tropicana redesign in the USA was that to be chosen, you must first be noticed – recognition of key visual equities is critical for frequently purchased brand leaders on a crowded shelf.

Given the tumbling cost of cameras, shouldn’t we be able to demonstrate the effect on shopper behaviour triggered by a new design by secretly filming consumers in store and witnessing their reactions to design in real life? There’s already at least one British supplier, Shopper Behaviour Explained, doing this and backing up their filmed observations by asking them about their motivations while viewing playbacks of themselves in exit interviews. When you read that the Tropicana re-design cost PepsiCo $35m in lost sales, it must make sense to wire up a representative cross section of stores to observe shoppers in situ. Perhaps digital technology can do for research what Big Brother did for our engagement with TV?

Let’s celebrate Ethnography as a helpful rebrand of Observational research, if that’s what it takes to get a new generation of research managers to consider it? We need research to reveal how people behave in the real world if we’re to be informed and inspired to creative alchemy.

So, can a new decade herald the redesign of design research?
I really hope so – but obviously the choice is yours.

This article was originally written by Andy Knowles for Research World.