Our thanks to Pete Foley, PTG guest blogger and innovation expert, for the third and final installment of long form blog posts that explore how psychology, behavioral and perceptual science can be applied to consumer research. Today, Pete addresses the important intersection of conscious and unconscious factors and how they influence consumer retail decisions.
This is the third in a series of blogs that explore how psychology, behavioral and perceptual science can be applied to consumer research. Today I’ll talk about some ways to design research that compliments noninvasive measures, thus helping us to get a better line of sight on the mixture of unconscious and conscious elements that influence retail decisions. I‘ll also talk about how these techniques can help us probe a couple of key perceptual mechanisms – deselection and affordances.
As I’ve mentioned in previous blogs, panelists over-thinking the decisions we study is one of the biggest enemies of predictive research.
Most decisions result from a mix of low and high engagement decision processes, and the mechanisms and outcomes of these are often different. Fast, time constrained decisions favor defaults, habits, and hence familiar brands and past choices. Longer, more thoughtful decisions open up the possibility of evaluating newer, perhaps riskier choices. Somebody rushing through a supermarket with many items on a shopping list may simply grab a familiar brand that they know they or their family will enjoy. It is a safe bet, and requires little effort. However, put them in research where it is obvious that we are very interested in their choice, and they may allocate more time to the decision, and consider new options. I’ve mentioned before that creating this kind of engagement may well be our goal if we are launching a new product, but if our data is to be predictive, it is critical that this high engagement thinking is triggered organically by our product or service, and is not an artifact of our methodology.
Restoring the Balance: There are things we can do to mitigate these potential confounds. Noninvasive, mobile techniques such as biometrics and eye tracking free us to test in more realistic contexts, and help prevent our measurement intruding on the test. However, all sorts of effects, ranging from lack of availability of prototypes, confidentiality, or lack of access to real retail environments mean that we often cannot test in a completely real context.
However, there are a couple of things we can do to augment noninvasive techniques, and mimic realistic levels of engagement, even if we cannot test in a completely realistic setting. For example, we can disguise the questions we are asking, and/or add enough ‘noise’ and cognitive load into a test so that the amount of time and effort people can allocate to the decisions we want to study approximates the real world.
Disguised Questions: For example, it is common in Behavioral Economics to go to great lengths to avoid people knowing exactly what question is being asked. Dan Ariely, the well known Behavioral Economist from Duke University has a terrific movie out at the moment called Dishonesty, and in it there are some wonderful and very creative examples of experimental techniques that disguise what data is being collected. My favorite is a modified shredder, that only shreds the very edges of documents, allowing his team to collect data that panelists believe they destroy because it is of no interest to the testers.
There are also many ways that we can disguise what questions we are asking. One of the simplest is to simply bury them in with a lot of other questions. In a retail context, for example, give people a long shopping list, a limited time to shop, and bury the decision we are really interested in somewhere in the middle of the list. Or give panelists a challenging, complex task that they think is the reason they are there, but bury the question we are really interested in another part of the process. Perhaps ask them to take part in a crowd-sourcing exercise, but offer a choice of products at the end of the test as their reward for doing the research. That is the decision we are really interested in. It’s not perfect, but they will reserve most of their mental effort for the part of the test that they think is being observed, which we may not even measure, and treat our target decision at the end with a level of engagement that more closely mirrors what they would do in the real world.
Signaling: Disguised questions not only help create a realistic balance of high and low engagement in another way. They help to avoid another potential artifact, another effect studied in Behavioral Economics called signaling. Part of the reason we buy designer brands is because they signal something about us to other people. The brand may be says something about our affluence, our values, or our lifestyle that we want to advertise. Think about this in the context of retail research. Are we more likely to buy a premium brand, or perhaps a smart, but expensive new product if we know people are watching us, than if we just run into a store, or browse a website where nobody knows us, and nobody is taking any notice of us?
Distraction: Disguising our question helps stop people from putting disproportionate time against it, but our ideal is to get them into a flow where they largely forget that they are being tested. Cutting edge techniques like eye tracking in the form of eye-glasses, noninvasive biometrics and wearable technologies all help with this, as they become almost invisible to the wearer after a short period of time. Giving people multiple, complex tasks as in the disguised choice testing described above also helps. A third, complimentary approach is to add non task related distracters. In a retail environment, add muzak, the smell of the bakery, other shoppers, shelf stockers, all of which add cognitive load that mimics the real world. Multisensory distractors like sound and smell are particularly useful.
Selection, Deselection and Grouping: In my last couple of blogs I talked about the heuristics that guide and capture attention. Typically, we want attention. We don’t want our new initiative to become an invisible gorilla – big, hairy, but largely invisible to the people who we want to notice it. However, not all attention is good, and if we do need it, attention alone is not always enough.
Why would we not want attention? Attention, in general, activates thinking and engagement. If we want to disrupt a market, or if we are a small player who wants to grow, this is a good thing. We want shoppers to stop and consider us. That means breaking through their habits and defaults, which are largely low engagement processes, and instead engaging them in high engagement thinking that can lead to new behaviors. But if we are already the market leader, this is risky. Certainly, we still need people to find us, but it is better if this occurs primarily via low engagement processes like past experience and memory. If we are located where they’ve found us before, and look like we’ve always looked, then they can find us without having to pay much attention. And provided we deliver quality performance, that is often enough to make the sale. Making people think more about their decision only opens the door to other options.
Attention is not always enough: If we are not a market leader, and want to grow share, then we do need to change behavior. Attention is part of what we need, but it is not enough on its own. The heuristics that guide visual search have evolved for efficiency, and a big part of that efficiency comes from not wasting time and resources on stuff that isn’t useful. If you read my first blog, I talked about saccades, and how our visual system uses what amounts to an attentional spotlight to focus our vision on information that is important. The counter to this is that we process very little detailed attention on areas that are not salient to us. For example, if we are driving down the freeway hungry, feature gain effects make the spotlight more sensitive to things that can provide food, and so we notice restaurants more than if we’d just eaten. But because our attentional resources are finite, this means that we also pay less attention to other things like electronics stores or gas stations. We will notice them, but because they don’t fit with our goals, we quickly deselect them. Something similar happens in a supermarket. If a shopper is looking for a soda, and their preferred brand is Coke, feature gain effects mean their attention is tuned to red. A high contrast section of, for example, blue may still grab their attention, but only briefly. As with the gas station or electronic store, their attention will quickly move on. They may spend just enough time on it to draw a boundary around the blue section, creating a whole section that they can ignore. This grouping and deselection is very efficient for the visual system, but can be very damaging if it happens to our brand. Obviously it is really important to know if it is happening to us, and if yes, mitigate it. Noninvasive eye tracking is about the only way to find this out, and test ways to address it.
Embodied Cognition: In the last couple of blogs I’ve talked a lot about tapping into heuristics and unconscious drivers of behavior. Noninvasive techniques, and the kind of test designs I’ve just described above allow us to get a much better line of sight on these. However, there is one heuristic I haven’t mentioned which may especially benefit from these techniques. We unconsciously reach for, touch, and hold objects that fit with our bodies. For example, we grab or pull a handle, cradle a curved object in our hand, and pull away from something sharp, or with hard angles. Called affordances in psychology, these effects are both subtle and powerful, and operate almost exclusively below awareness. A handle that is too small for a hand, or that butts up against another product on a shelf, and is hence difficult to grab, can cause shoppers to unconsciously deselect a product. Facing a choice between a product with a handle, and a very similar one without, shoppers will often unconsciously grab the one with the handle. Non-invasive methods allow us to more effectively evaluate, and iteratively design for these affordances.
So in summary, once we are armed with noninvasive techniques, it opens doors to both measure, and optimize all sorts of important, but otherwise hidden effects than can have a significant impact on what shoppers purchase. However, it is also important to pair noninvasive techniques with experimental designs that compliment them, and test shoppers who are making decisions with similar levels of engagement to those which would apply in the real world.
Building upon last week’s thoughts from PTG guest blogger and innovation expert, Pete Foley, we are pleased to share the second in a series of long form blogs where Pete explores psychology, behavioral and perceptual science and how it can be applied to consumer research. Today’s installment addresses the power of noninvasive research and the unconscious drivers of consumer behavior.
This is the second in a series of blogs where I explore how psychology, behavioral and perceptual science can be applied to consumer research. In the first blog I talked about the value of running research in real contexts, and of using noninvasive techniques to unearth unconscious drivers of consumer behavior. Today I’ll talk in more detail about how this can help us to better understand and leverage the cognitive and perceptual biases that often drive behavior, making research a more powerful iterative and prototyping partner for innovation and design.
Recording what people do is of course an important part of research. However, as I mentioned in my last blog, if we want our research to be an integral part of optimizing product, package or communication design, then we need to dig deeper, and understand the mechanisms behind those decisions. That means exploring the cognitive and perceptual biases and heuristics that often drive actions and decisions. By combining knowledge of these with noninvasive techniques that measure both nonconscious and conscious behavior, we can help to design communication, products and services that better resonate with how people really make decisions.
Mixed Conscious-Unconscious Decisions: Most human behavior is driven by a mixture of conscious and unconscious mental processes. Take driving as an example. Many of us will probably at some time have experienced arriving home from our daily commute, only to remember very little detail about the journey. This realization is often a little surprising, even disturbing. However, our autopilot, or unconscious brain is actually quite capable of handling the complex, but routine task of driving a familiar route. It can direct where we look, and our actions, and in many cases does so more effectively than our conscious self. However, with driving, as with most tasks, our conscious and unconscious minds work together. Our unconscious, or autopilot, handles the routine stuff, with our conscious mind taking over when our habits, defaults and automatic responses can no longer handle a situation. In this case, autopilot can usually manage a familiar freeway drive, freeing us up to listen to the radio, or mull over an important meeting from earlier in the day. However, if something unusual happens, and provided we are not overly distracted with something like texting, autopilot will hand over control to our conscious mind as needed. For example, if someone pulls out in front of us, our unconscious will likely slam on the brakes, because it is faster to action, but then hand over to our frontal cortex to help deal with any consequences. If road work closes our normal route, or if it starts snowing, then we may get a similar handover, as autopilot does not have the experience to deal with atypical situations.
This interplay between fast, routine, largely unconscious mental processes, and more considered, slower, ones plays out in many human behaviors from playing sports to shopping. As I mentioned in my last blog, Daniel Kahneman in his Nobel Prize winning work on Behavioral Economics labeled them System 1 and 2 thinking in the context of human decisions. However, we also see similar effects in perception – where we look is often automatic, but we can also evoke manual override, and direct our vision as well.
Underestimating Automatic Decisions: An important insight for research that the driving analogy illustrates, but that applies to a wide variety of tasks is that we tend to underestimate how big the unconscious component of our thinking is, and often discount it. However, understanding the unconscious biases and heuristics that guide perception and decision can open up all sorts of opportunities to help people find what they need. In fact, this can be a more powerful tool than persuasion, as biases are often more consistent than more thoughtful preferences.
This is potentially powerful, but not simple. There are a lot of biases. I keep a list of over 200 cognitive, memory and perceptual biases, and that is far from comprehensive. One way to simplify is to recognize that some biases are more often relevant to retail than others. For example, “following the crowd” (herding in Behavioral Economics), “preferring familiar choices” (the mere exposure effect in psychology), and “choosing the middle of a choice distribution” (Compromise Effect) often drive time constrained decisions on the web or in the supermarket.
But the real world is messy, and in virtually every real situation, multiple biases and heuristics compete. To understand each unique situation, we therefore still need to run research. But the better we understand how these biases work, the better we know what to look for, what ‘questions’ to ask, and hence the better research we can design.
Retail Application: The same applies for perception and attention. Take a new product in retail as an example. If we want a shopper to notice and try something new, we first have to ‘snap’ them out of autopilot, which favors past habits and defaults. Understanding perceptual biases is key to creating the attention cues that will grab attention in this way. However, a common intuition is to think there is a fixed hierarchy of cues that we can tap into in order to do this. For example, we intuitively feel that our attention is drawn to things like movement, bright colors, or cues associated with sex or food. These are important, and all else being equal, we are primed to attend anything to do with threats to our survival, eating, or mating. But reality is more complex than this, and context plays a huge role. The cues that grab attention are those that are salient at that time, and that stand out against local background. For example:
1. Differentiated vs. Local Context: A bright, flashing light will catch our attention, but only if it is not surrounded by other bright flashing lights. We need to be differentiated versus the immediate surroundings. Add another flashing light to Times Square and it is probably not going to grab our attention.
2. Salient and Intuitive: If something isn’t salient to our goals, we may notice it, but will quite likely unconsciously deselect it. The same applies if it isn’t intuitive, especially in a time-constrained environment. Unless we have been primed to look for it, we may well notice, but then deselect or reject something that we don’t understand, before it even gets consciously considered.
Testing in realistic contexts and levels of engagement and distraction are key to measuring this bias. An environment that triggers unrealistic engagement in a visual search or decision is likely to overestimate the appeal of the new, novel, or unusual. This is why noninvasive research in realistic context is so important.
Mobile Cognition: In the case of bricks and mortar retail, we also need to account for mobile cognition. Real shoppers spend much of their time on the move. What we notice varies enormously as a function of distance, and the amount of attention we dedicate to a scene changes if we are moving, rather than standing still. We therefore need to build mobility and autonomy into our methods, and measure behavior at the distances where a shopper will naturally make decisions.
Leveraging the Shopping Cycle: This is made even more important because most retail decisions do not occur in a vacuum. Contextual effects such as goal, emotional state, degree of engagement, and surrounding visual and conceptual information all influence decisions, but so does time, and when a decision occurs in a shopping trip. Cognitive effects like priming and decision fatigue can strongly influence engagement, attention and ultimately, decisions. These effects change as we progress through a shopping trip, be it in a supermarket or on the web.
The Shopping Cycle and Decision Fatigue: For example, what we’ve already purchased can influence us in many ways. It may make us more likely to buy complimentary products. Or if a shopper is already close to their mental budget, and have depleted the ‘slack’ that is normally built into it for unexpected or forgotten items, they may be less likely to try new products, even if they notice them.
Another well known, and pertinent effect in psychology is called ‘Decision Fatigue’, or Ego Depletion. The energy we have available to make decisions is limited, and the more decisions we make, the less of that energy we have. The effects of this can be quite dramatic. A study of Israeli judges showed that they consistently gave parole to about 70% of people at the start of the day, but this rate drops to near zero towards the end of the day. Breaks for lunch or coffee briefly bumped it back up to 70%, but the trend was very much down as the day passed. In this case, the easier, or default decision is to not give parole, and so the more decisions they made as the day went on, the less likely they were to grant parole, as they fell back on the easier default of not granting parole.
Judges are professional decision makers, so imagine if decision fatigue has this kind of impact on them, how much it can impact our everyday lives. A shopping trip also involves many decisions, and so defaults are more likely to kick in towards the end of a shop. We see this play out with candy offered at checkout. The mental energy that we use to make decisions is the same as that used for self-control. We are therefore susceptible to grabbing a candy bar at checkout, even if we are not being cajoled by our children, because our self control has been depleted by all of the choices we’ve just made. And, this goes beyond just the checkout. For example, market-leading brands are often habitual or default choices, while buying a new, innovative, or growing brand typically requires more thought. For retailers and CPG manufacturers, where in the shopping cycle a category is placed can therefore be very important, with new products probably favored near the beginning, and big, established brands near the end.
Once again we can only effectively measure this if we keep context, engagement and autonomy in our tests. If shoppers know they are being tested, not only does it tend to make their decisions more considered and less automatic, this also increases the amount of energy that goes into them, in turn increasing decision fatigue. Forcing shoppers into an artificial context with too few decisions, asking them to follow what for them might be an unnatural route, or simply overtly observing them all change the amount of decision energy used.
It is always challenging to completely eliminate artifacts caused by the testing process. However, techniques like noninvasive eye tracking and biometrics can go a very long way towards helping us to avoid the act of testing influencing our results.
In my next blog I’ll talk about test designs that can compliment noninvasive measurements, how noninvasive techniques can help us design products and packages that invite us to handle them and take it off the shelf, and I’ll dig a little deeper into a concept I briefly mention here, selection and deselection.