Are your customers revealing their true preferences?

Without a warm up, what they say may not be what they want

The warm-up is a practice long familiar to "live" TV audiences, sports stadium crowds and the athletes themselves – it's all about getting people ready for the experience. And, when trying to glean accurate market insights and product preferences, warming up customers is now strongly recommended for researchers, brands and others.

According to a recent study, putting consumers in the right frame of mind to consider what they really want is particularly important for product developers and those who help them uncover customer preferences, not least because a mistake can be extremely costly when new products flop or fall wide of the mark.

To figure out what they really want and to make the right call, consumers need knowledge and time to think deeply or "self-reflect" on their choices before making a decision, say researchers Songting Dong of UNSW Business School, John R. Hauser of MIT Sloan School of Management and Min Ding of Pennsylvania State University.

Their study finds that asking consumers to state preferences cold, without the benefit of sufficient self-reflection, is a risky business, particularly when the method used is qualitative and unstructured – and this includes some old research favourites: focus groups and interview discussions.

Self-reflection is crucial for discovering the true voice of the consumer, and it may lead to reduced development costs and, ultimately, more successful products, the researchers claim. In their paper, Self-Reflection and Articulated Consumer Preferences, the trio uses the example of the automotive industry where it takes several billion dollars to develop a new model car.

Informed choices endure

There's a multitude of ways to discover customer preferences, from preference elicitation methods such as conjoint analysis, in which researchers estimate how people value a product's features, functions or benefits, to preference articulation methods such as Casemap in which research participants state how they value product features by rating its attributes. Many also rely on the qualitative approaches of focus groups, interviews and discussions.

"All these methods are used assuming that what consumers say is what they really want and that the preferences revealed will be relevant when it comes to making a purchasing decision," notes Dong, a senior lecturer in the school of marketing at UNSW Business School and a member of its digital enablement research network.

'You need to warm people up before you ask participants to state their preferences in unstructured tasks such as focus groups and depth interviews'

SONGTING DONG

"Our study shows unstructured preference articulation methods, such as the popular focus groups and depth interviews, can fall short for accurately discovering true consumer preferences."

The problem is not with the research methods themselves but about the need "to warm people up" with information required to make their purchase decisions "and giving them a chance to think about their needs carefully", Dong says.

In their paper, the researchers draw on the real-life example of Maria who thought she would ditch her 1995 Ford Probe for "a sporty coupe with a sunroof, not black, white or silver, stylish, well-handling, moderate fuel economy, and moderately priced", but after scouring the web and reading consumer reports, Maria changed her mind.

While her early thinking had her potentially lining up for a Hyundai, Nissan or Infiniti coupe, after learning more about the options, her true preferences emerged as a secondhand BMW or Audi. And when her Probe eventually succumbed to old age, Maria purchased the latter, so her informed choice endured.

A cool down period

In a study of 204 consumers in the US automotive industry, the research team provided participants with a strong incentive to think carefully about their preferences by offering them the chance to win the car of their choice. Much was at stake – wrong call, wrong car.

For comparison, the team rotated three different measurement methods: a choice-based conjoint analysis task; a Casemap feature-rating exercise; and an unstructured preference articulation task that asked participants to write an email about their preferred car choice to a friend who would act as their "agent" to buy a car for them if they won the prize.

One week later, their preferences revealed by the three tasks were validated with a follow-up product choice task.

"The one week delay was like a cool down period," Dong explains. "After one week, most people would have almost forgotten what they had told us in the three measurement tasks, making their final choice unbiased by what they said previously."

The study showed the order in which people had undertaken the methods had an impact on outcomes. Which task came first mattered – and those who had significant time to self-reflect first, via conjoint analysis or the Casemap tasks, were more likely to state their true preferences in the writing-an-email task. Their choice endured after the cool down period.

"Sufficient warm-up had helped them provide more accurate statements," says Dong. The early choices of consumers who wrote the email first (without sufficient warm-up), however, were less reliable.

The implications are clear. "You need to warm people up before you ask participants to state their preferences in unstructured tasks such as focus groups and depth interviews; sufficient warm-up may lead to much better data quality," Dong insists.

Multiple data sources

Even giving focus group and depth interview participants time upfront to read survey questions and consider their answers may not be enough, cautions Dong. A subsequent study of the Hong Kong mobile phone market showed a tendency to hang on to early-stage preferences.

"Even though people have more information and more time to reflect later, some errors in their initial thoughts go unrevised," confirms Dong. Reasons for this may be due to the way the short-term memory works, he says, or the determination "to stick with what you have decided".

'The study results show that sufficiently challenging decision tasks work well'

SONGTING DONG

Is there a better way to warm up? "The study results show sufficiently challenging decision tasks work well," Dong believes. He suggests asking research participants to make thought-provoking choice decisions first "and ensuring those decisions have real consequences to keep participants fully engaged". 

So, what of today's inclination to leap to the internet – an environment that often lends itself to instantaneous rather than reflective behaviour – to glean customer preferences?

Indeed, Dong spots many problems with the proliferation of online surveys, in particular bias issues – from selection bias in which "some of your customers who do not search the internet will not be covered" to hypothetical bias in which people do not think carefully or answer truthfully because "there are no real consequences to my answers". There's also response bias as less than 1% of people accept the invitation to fill in online surveys.

To help navigate this difficult survey environment, Dong has some useful pointers:

Try to link survey questions with real consequences and prompt consumers to think by including incentives. "Not just plain payment, the participants' reward should be related to their answers," he says.

Cross-check research results: "Combine results from multiple research studies, that way we can be more sure of the results."

And harness big data. "Merge multiple data sources. These days we have more and more information available. For example, if you want to know what people really think about your brand, try validating their answers with real purchase behaviour," says Dong.