Bias by Design
By Patrick Bach and Monica Zhang
¹Every morning you wake up and on your way to work (or school) you stop at your local coffee shop to grab a hot drink.
If you’re like most people you very likely purchase a medium-sized beverage without much thought.
Now let’s pause and think about that.
We believe that as human beings, we are in control of our choices and our decisions; we believe that we are rational, autonomous decision-makers, right? Let’s take a moment to unpack the decision to purchase the medium coffee: was that decision driven by your knowledge of exactly how many liquid ounces of coffee you need to feel satiated? Or could there be something else at play?²
Behavioural sciences teach us that there is a psychological effect that is driving us to choose Medium (over Large or Small) which has more to do with the mental shortcuts our brains leverage each and every day to help us navigate the thousands of decisions we make. This effect is commonly referred to as “The Goldilocks principle”.
In cognitive science and developmental psychology, the Goldilocks principle refers to an infant’s preference to attend to events that are neither too simple nor too complex according to their current conception of the world.
That same principle is at play here. “Medium” represents a safe, reliable choice. Not too much, not too little. And that safe, easy decision is what helps keep us from becoming overwhelmed by the amount of information we would ordinarily need to take in to inform even a simple decision like what coffee size to purchase. The way information is presented and how choices are framed deeply impact our ultimate behaviour.
Now imagine how you might leverage this knowledge as a designer. Whether brainstorming ways to frame a new product’s value proposition on your website or presenting clients with varying options and features for a service: How you choose to display this information will inevitably bias the recipient towards a specific outcome.
Now let’s take a step back. Before we can really understand behavioural economics, we need to look at traditional economic models. In traditional economics, we are told that human beings are fully rational, that we make decisions based on perfect information, that we carefully weigh and consider all the options presented to us, that we make decisions individually. However, in reality, we are exposed to so many choices and decisions in a given day that this quickly overloads our brain. Our brains have evolved to overcome this burden by developing shortcuts that we call heuristics. These shortcuts aid the decision-making process but in doing so leave us prone to a number of different biases.³
Like in our medium coffee purchase, the desire to make a quick and relatively safe choice about how much coffee to buy means we end up choosing the middle option regardless of whether it is truly “optimized” to our consumption needs. This is particularly true in cases where we don’t have regular feedback or reliable experience. With coffee, we have likely ordered coffee or similar beverages before, and, should our order be too much or too little, we can easily correct it next time. With other, perhaps less trivial and more complex decisions like choosing between investment portfolios or insurance options, we may not have that feedback or experience, and so will likely default to more quick and easy heuristics.
As designers, our role and purpose is to help our users achieve the best possible outcome in any given context or scenario. In order to live up to that very lofty aspiration, we must first come to terms with a few different truths and embrace them.
#1 Human beings are inherently biased and don’t always make logical decisions
How you choose to present information to them will impact the discussion and outcome.
A classic example of this might be the difference between presenting your user with a $25 item with a $5 shipping charge or a $30 item with a zero-dollar shipping charge. From a purely logical, economic perspective these are completely identical choices, however, behavioural science might suggest that clients would be far more likely to follow through with the free shipping option. The psychology behind this is actually quite simple. Human beings are inherently loss-averse and effort-averse. Shipping costs are viewed as extra “hassle” and are not viewed as providing any extra value.⁴
#2 There is no such thing as a neutral design, only an unconsidered design
As much as we might like to believe that we present things in an unbiased way to our users, the reality is that every presentation guides users towards one choice more than alternative designs might. Even something as banal as choosing cereal at the grocery store is inevitably influenced by which boxes are at eye level, which are at the ends of the aisle and which are in the middle.The same holds true for any digital experience. How and where you choose to promote your premium options, for example, absolutely impacts the outcome.
The above is a picture taken from TurboTax’s website displaying their various product offerings. Even in this small capture, a number of behavioural principles are influencing decisions. For example:⁶
- 1. Social proof: When making a new or complex decision, we tend to use the actions of others as a rule of thumb for our own choices. In this case, seeing the “most popular” tag can pull people to that particular option.
- 2. Power of Free and Sunk Cost: TurboTax has cleverly worded their buttons, “Start for free.” Paying is painful. By delaying payment, TurboTax accomplishes two things: first, free things are naturally enticing to people — just think of free samples at Costco — and people will often try things they might not otherwise just because they are “free”; second, by the time people pay, they will have invested time and energy, meaning that abandoning the process and not paying now has a perceived cost attached in wasted effort.
- 3. Goldilocks Effect: Much like with the coffee that opened this article, people gravitate toward the middle option, especially when unsure what product might actually best suit their needs.
There are almost innumerable ways the above website influences the decision of its users. One might feel like the use of all of these “psychological tricks” borders on mind control, but getting rid of them or changing them doesn’t solve the problem — it just guides users towards a different set of decisions while potentially creating confusion along the way. As designers, we need to consider how our designs are influencing user behaviour and what decisions our designs might inadvertently be encouraging.
#3 Most designers are ill-equipped to navigate the complexities of human behaviour
As designers, we often rely on our intuition to make good design choices. In many cases that intuition is a very valuable asset in our problem-solving toolkit. However, it may also be a double-edged sword. As human beings, we tend to be pretty poor predictors of behaviour (our own and that of others) which is why our intuition can often mislead us into creating designs that either don’t help achieve our desired outcome or worse, completely backfire.
This can lead to bad outcomes for our users and it can also lead to bad outcomes for our business.
For example, while working on a project in the insurance space, our team had made a reasonably well-informed decision that letting our customers “build” their own packages would be preferable to simply having one “recommended” by the provider.
After all, there’s lots of research out there that supports the idea that a sense of endowment can go a long way in creating brand loyalty and stickiness. Well, we put this concept to the test, and lo and behold it backfired! We found that when users have been presented a “recommended” offer vs one they could “build” themselves, they were far more likely to trust that this was the best product for their needs, which in turn translates into intended purchase behaviour. How could this be? Well, it turns out insurance is such a complex product that most users feel uncomfortable making more detailed choices about their coverage. They would prefer that someone with specialized knowledge (the insurance broker or provider) make those decisions for them, based on demographic and historical data.
Apples to Apples?
So far, we’ve explored a number of relatively low-stakes examples of how biases can impact an outcome. But a fundamental misunderstanding of customer psychology or conversely an overreliance on our own intuitions can sometimes backfire in a colossal way.
Let’s explore an exemplar of what not to do in the world of retail. In 2011, US retailer J.C. Penney wanted to improve their flat sales and underperforming stocks. To do so, they brought Ron Johnson of Apple retail operations fame onboard in an attempt to shake things up. Johnson was a superstar in the world of retail having built the concept of the modern Apple Store — a concept that was so successful that Apple stores would go on to be the most lucrative retail outlet per sq ft (nearly doubling the revenue of the 2nd place retailer Tiffany’s).
Johnson had a vision for a radically different kind of retail store, one that offered everyday pricing that is “fair and square.” In other words, an item constantly on sale from $15 to $10 would just consistently be listed at $10. While understandable, customers hated it and the strategy backfired. Sales declined by 22% and stock prices dipped more than 50%.⁶ So what went wrong?
First, there was a misunderstanding of context. Johnson had used a similar strategy at Apple and assumed that the same strategy would work for J.C. Penney. He didn’t understand that J.C. Penney consumers were different from Apple consumers — or rather, he didn’t understand that people behave differently when shopping at J.C. Penney compared to shopping for Apple products. When shopping at J.C. Penney, people were anchored to the high list price and felt a thrill and immense satisfaction comparing their actual purchase price to that high anchor. For J.C. Penney consumers, finding a “deal” was part of the value and attraction of shopping at J.C. Penney in a way that it wasn’t when shopping for Apple products.
Second, there was overconfidence in the strategy. Overconfidence is a frequently cited classic of behavioural science; people are generally overconfident. In this case, Johnson was overconfident that his strategy would work based on his one data point from Apple, a data point of questionable analogous value given the different context mentioned above. Johnson never ran any pilots or experiments, he assumed that his strategy would work and went all in, severely costing the company he was trying to help. Johnson’s experience suggested his strategy would work, but because he never tested his intuitions, he never got the evidence that might have disproved his hypothesis and saved J.C. Penney millions.
Ethics in Behavioural Design
These examples can help you avoid bad design in the typical understanding of the phrase, but we must also be wary of bad design in the moral sense. A common critique of Behavioural Economics and behavioural science generally is that it can steer people towards decisions that they don’t want to make. It can feel manipulative and morally icky — think addictive designs like the social media death scroll or the dark patterns commonly found in cancelling service or unsubscribing from a newsletter. To that end, there are a few ethical design principles that can help guide your use of these heuristics and biases to design products and services for end-users.
Ensure Designer/User Interests are Aligned: Are you designing for what the user would want you to design? For example, are you designing to maximize spend? Or are you designing to help your users pick the product that is right for them?
Be Up-front and Transparent: Don’t mislead users. Be honest about what you’re doing and why. Intentional framing is not the same as lying or omitting critical information.
Provide an (Easy) Way Out: Keep in mind that designs should be connected with choice and your role as a designer is to make choices easier for users and not to make the choice for them.
Where do I begin?
There are hundreds if not thousands of biases. There are biases that affect how we remember things and what we remember, biases that affect how much information we can process at once, and also biases that push us to act faster. So how do you as a designer make sense of all of this?
Step 1: Be self-aware. Being aware that all of your designs are inherently biased is the best place to start.
Step 2: Familiarize yourself. You can’t know ALL of the biases, but there are many that are so frequent and so common, that they’re worth knowing in and out. Here’s a helpful resource: https://thedecisionlab.com/biases/
Step 3: Experiment. Never rely on your intuition alone. In fact, never rely on what you read the biases either. Each bias exists and impacts our behaviour in a unique way in each context. Often multiple biases are at play and influence us in contradictory ways. Use an evidence-based approach to testing your designs both qualitatively and quantitatively to ensure you’re actually achieving the outcomes you value.
In conclusion, the reality is that there is no neutral design, only unconsidered design. Whatever you design will influence behaviour so as a designer you are obligated to consider how your design shapes user choices and ensure that your influence is ethical.
[1]: Coffee cups by www.brando.ltd via Unsplash
[2]: Coffee Cup Sizes Farzaneh Ab via Pinterest
[3]: https://cooltool.com/blog/behavioral-economics-why-a-psychologist-got-a-nobel-prize-in-economics
[4]: https://blog.rjmetrics.com/2014/07/14/the-psychology-of-free-shipping-why-it-works-as-a-marketing-tool
[5]: Image source: https://turbotax.intuit.ca/tax/software/personal/online
[6]: https://www.reuters.com/article/us-jcpenney-shares-idUSBRE91R0UT20130228
