Embrace your customers’ unconscious decisions

Embrace your customers’ unconscious decisions or they are gone in .06 seconds, writes Gavan J Fitzsimons.

We are waking up to unconsciousness. Over the past 20 years there has been an explosion in academic research in psychology and marketing exploring the role of the ‘unconscious’ in consumption domains. The title of my own research lab, the Unconscious Consumer Lab, would have been highly controversial in the corridors of the world’s top business schools at the turn of the century. Today, gratefully, it seems relatively pedestrian, thanks in part to recent popular press books such as Professor Danny Kahneman’s excellent bestseller Thinking, Fast and Slow. No longer do we wonder if consumers are influenced outside their conscious awareness, but rather we wonder to how great a degree they are.

Money for nothing

Accompanying this academic research has been a wide array of research tools and insight opportunities designed to capture non-conscious consumer processing in the commercial research space. And yet, it feels like a virtually daily occurrence when I hear from an insights director or a VP of marketing that their most recent – typically quite expensive – foray into the world of the non-conscious/neuroscience/psychophysiology has flopped, and they are under attack from their colleagues in finance for poor return on investment yet again. Why the disconnect from the growing academic acceptance that non-conscious processing is critical in consumer settings, and the lack of strong support and movement towards non-conscious research tools in practice?

The unconscious paradox

A powerful domain to explore non-conscious consumer processes has proven to be that of the consumer-brand relationship. Years ago, we demonstrated that the strong associations consumers held with brands could be non-consciously activated and lead to changes in behaviour. One of our first such demonstrations involved subliminally priming participants with one of two brands (either the Apple logo or the IBM logo) and then examining their subsequent behaviour. Across a range of studies and contexts we found that consumers who had been unwittingly exposed to an Apple logo were approximately 30% more creative in post-exposure behaviour than their counterparts who had been subliminally primed with the IBM logo. Rosie Ferraro, a former Duke PhD student now at the University of Maryland, showed that this influence extended to actual brand choice as well. She found that participants who were unwittingly exposed to Dasani bottled water in a series of photographs were about twice as likely to buy Dasani water shortly after exposure. We followed up that result and showed that consumers who fast-forwarded past advertisements when watching TV shows they had recorded at home were in many cases more positively disposed to the brand than if they’d watched the actual ads in real time. In other words, non-conscious exposure to brands led to substantial changes in behaviour. Contrast that to the very small average effect of being consciously exposed to traditional advertisements: a recent study estimated an advertising elasticity of only 0.13, or in other words for every 1% increase in advertising spend, a firm should expect only a 0.13% increase in sales. We have argued that the non-conscious pathway to behaviour is potentially so substantial that consumers’ defensive mechanisms are not activated in these settings. If we have positive associations with a particular brand, non-consciously activating them leads to a more positive predisposition to act if given an opportunity. By comparison, consciously thinking about a brand not only leads to activation of the consumer’s set of associations with the brand, but also leads to the activation of our defensive screen: “Brands are attempting to persuade me to act; I do not want to be influenced; thus I will consciously resist.” These examples are just the tip of the non-conscious iceberg. And so, this should lead to great success in the commercial world of all things non-conscious, correct?

Unfortunately, that hasn’t proven to be the case. Take as an example forays into using neuroscience, broadly construed to predict consumer behaviour. Some research techniques in the neuroscience category can be extremely cost- and time-intensive (for example, fMRI requires extremely expensive equipment and takes an hour or more per participant for straightforward studies). Early efforts to use fMRI in the consumer insights space led to some very mixed results, as documented by a recent Advertising Research Council study. Anecdotally, I’m repeatedly surprised to hear that firms have dedicated multimillion-dollar budgets to their efforts in the space, with grand takeaways that distil down to “we found that the reward centre of the brain was more active for our loyal versus casual users during consumption”.

Implicit measures

Another example might include the use of implicit measures (e.g., the Implicit Association Test) to predict consumer behaviour. Such measures ask respondents to categorize a word or image as quickly as they can. For example, you might be asked to categorize a series of brands as either ‘good’ or ‘bad’. Which category you put the brand into obviously matters, but the implicit piece of the measure is the speed between when the brand appears on the screen and when you respond. This millisecond response time is taken as an indication of how strongly you hold the opinion that the brand is good or bad, with faster response time indicating more strongly held beliefs. Once again, while there have been some successes, there have been just as many commercial flops attempting to utilize such techniques to predict consumer behaviour.

For example, a large retailer I partnered with relayed its experience of attempting with limited success to segment and target its customer base using such implicit response-time measures. Lest I appear to simply be a curmudgeon about implicit and neuroscientific research approaches, note that some of the most interesting research I’m currently involved in uses such techniques (e.g. fMRI, mobile eye-tracking and response-time measures). The research questions these techniques allow me to answer are unique and different from those I could answer using purely behavioural research approaches. For example, we’re currently pairing mobile eye-tracking data with response-time measures to explore how consumers are impacted by in-store primes. These tools permit us to examine millisecond differences in approach to certain items on a shelf as a function of background imagery (signage above the shelves).

When working with my commercial counterparts, I try to emphasize that we first need to figure out the key question we are trying to answer with any research endeavour. Is it whether people attend to a promotional effort of some type? Is it whether their set of associations with our brand changes? Is it whether they’re more likely to pick up our product off a store shelf and investigate for more information? Or is it simply whether a particular promotional effort will lead to more sales of our product? Counterparts on the finance side of the firm typically focus on the last question, but any of the above would be examples of legitimate research questions. Once we’ve got our key question laid out, we can then dig a little deeper and explore what type of research method might be most appropriate. Let’s assume we’re going to explore whether a promotion (an ad, a feature display and so forth) leads to increased sales. I argue that the key determinant of whether a more traditional, conscious research tool fits the bill prior to launch, or whether a newer, non-conscious tool might make more sense, comes down to one key dimension: the degree to which the consumer choice is largely deliberative versus largely spontaneous.

Deliberative decisions are often made in consumer settings. For example, when purchasing an automobile, most consumers are not rash, impetuous decision-makers. Rather, they gather information, compare options, weigh strengths and weaknesses etc. Smaller ticket items might also involve highly deliberative decision processes. For example, selecting a present for a loved one hopefully involves some degree of conscious processing. Selecting a product in a category that is new to the consumer would also typically involve fairly deliberative processing (e.g., purchasing an over-the-counter medication for an illness). In such categories, tried and true research techniques, such as day-after recall, attitude toward the brand, and purchase likelihood, are going to be good predictors of whether a consumer is likely to purchase one brand versus another.

Just as deliberative decisions are often made by consumers, so too are more impulsive or spontaneous ones. Choices made in categories that are selected every week might be simple fulfilment choices, and require essentially no attention or focus. While standing in the checkout queue at a grocery store, a consumer might decide whether or not to throw a package of chewing gum onto the register belt. Or it might simply be that a particular consumer is coping with three small children as he or she attempts to pick up a few items. In each of the above examples, there is very limited deliberation going on. Rather, the consumer is using their non-conscious system to determine what to do.

It’s worth noting that many spontaneous or non-conscious decisions in consumption settings are choices that consumers once made in a deliberative fashion, but now need dedicate no resources to. Then there is a category of spontaneous choices that build off a set of associations with a brand, for example, that has been acquired through conscious means. Which gum you select at the register may be based on seeing ads for the gum, learning it’s dentist-recommended, and so forth, even if you haven’t purchased chewing gum in years.

And finally, some spontaneous choices are truly novel with no prior input – for example, upon entering a store for the first time, do you gaze left or right first? It is in these situations research methods that tap into the non-conscious can have much more predictive success.

A mixed picture

How do you know whether your customer is going to be making a largely deliberative or a largely spontaneous decision? As with most questions in field settings, the answer is that you should expect heterogeneity on this dimension. Customers new to a category or customers with no time constraints are likely to be more deliberative, as are those shopping for expensive goods, or those that are high stakes on other dimensions (e.g., medicine). By contrast, those making choices they have made repetitively in the past, or those making choices in low-stake settings, are likely to be making more spontaneous choices. Of course, these types of rules of thumb are made to be broken. I would have thought that choosing a romantic partner would ordinarily be a fairly deliberative decision, and yet technology has created the millisecond swipe to categorize potential mates.

The key managerial decision is to work out which of your potential customers, or customer segments you are most interested in targeting, and to match the research approach accordingly. If I’m selling treats for dogs, for example, many of my customers will be regular customers that make a largely spontaneous choice. But new dog owners are likely to be highly deliberative. Acknowledging that we can’t design an effective approach to gathering insights without first deciding which of these two segments we’re interested in will be critical.

Of course, one solution is to simply take measures designed to tap into the conscious side of the consumer’s mind, as well as measures designed to tap into their non-conscious side. I occasionally hear people say that the explicit measures tell you what will happen, while the implicit tell us why they happen.

However, this strikes me as a cop-out, and possibly an implicit acknowledgment of the fact that consumers often don’t really know how much of what goes on in their minds is occurring outside their conscious awareness. And nor, to be frank, do firms.

Rather than investing heavily in all forms of research methods, and thus very likely wasting a substantial amount of money, better to ask the deliberative-spontaneous nature of the task your consumer faces. If we can answer this question well, we can manage the resource allocation across our insights efforts more effectively.    

— Gavan J Fitzsimons is Edward S & Rose K Donnell professor of marketing and psychology at Duke University’s Fuqua School of Business