This Uber cool thing called Behavioural Design

How leveraging insights and methodologies from behavioural science can inform user centred design.
With the winter months well and truly upon us, last week I found myself staring out of the office window wondering just how I was going to get to south-east London, normally a 45 minute journey, in an ambitious 20. To make matters worse, I had a huge suitcase full of my belongings to take with me. With no tube stations near my desired destination, I decided Uber was the only option.
Now, I’m not afraid to admit it — I’m pretty tight with money. Maybe it’s the cost of living in London, or perhaps being a Scot it’s just in my blood. Regardless, the thought of getting an Uber is less than appealing. That is, until I saw the option of an UberPool — a ridesharing service with a slightly longer waiting time and a short walk to the pick-up location, in return for a cheaper fare. Being a sucker for a bargain, I took the shared service with two other passengers, arriving at my destination in just under 30 minutes (including a 3 minute walk at either end). Despite the distinct lack of passenger interaction (and carpool karaoke for that matter), overall I was extremely satisfied with the service. So, why isn’t everyone choosing to go for a cheap ride? Well in fact, it would appear they are. Diving deeper into the world of passenger pooling, more than 50% of passengers in US cities such as San Francisco are now choosing to pool, and shared journeys have saved over 1.7m miles being driven in London so far in 2019. 21st century customers it seems are placing greater value on their social and environmental impact.
But, what is interesting is how this shift in customer behaviour came to fruition. In late 2018, with customers being accustomed to quick-matching and door-to-door experiences, potential UberPool riders were found to be cancelling their trips in between request and match (with a driver) at higher rates than in relation to other products. To unpick this problem, Uber created a designated Labs team, responsible for leveraging and applying theoretical insights from behavioural science to build and design products. It was this insight which ultimately informed the future design of UberPool, in a way which was both functional and customer-centric.

Daniel Kahneman and Amos Tversky first introduced the concept of behavioural science in the mid-1970s. During this time, it was still widely accepted by neoclassical economists that individuals act rationally, carefully weighing up the benefits and costs of a decision. The two psychologists however discovered the opposite; when we are faced with decisions which are too difficult to comprehend, we instinctively focus on one key factor, rather than rationally assessing every possible detail. These mental shortcuts we take, or heuristics as defined by the duo, help us quickly solve a problem or make a choice. Though in most cases these heuristics are extremely useful in allowing us to quickly make sense of complex environments and situations, they can also ‘fail’ us when misinterpreted. When this failure occurs, the result is cognitive bias, misdrawn conclusions as a result of mistaken judgement.
In the late 2000s American Academics Richard Thaler and Cass Sunstein developed their own theory which would become commonly known as ‘Nudge Theory’. Nudge Theory aims to influence the behaviours and decision making of groups and individuals through positive reinforcement and indirect suggestions, without taking away individuals’ power to choose. Thaler and Sustein identified a number of nudges — intervention points which alter the environment so that when biased thinking occurs, the resulting choice will be the most positive or desired outcome. Make sense? If not, imagine being in a supermarket. Replacing sweets with healthy snack options at the checkout tills to reduce junk food consumption is an example of a nudge (it’s a real thing, check it out!). As with every theory, additional nudges have been proposed by various theorists to apply in wider social contexts.
So, why does this all matter? By being aware of the biases which influence behaviour, we can design products, services, communication or other interactions which are positively influenced by nudges, and drive a high customer value perception. To help us do so, Buster Benson and John Manoogian III developed the Cognitive Bias Codex in 2016 (Figure 1), a visual matrix of 175 cognitive biases grouped by the problem they are trying to solve. With such tools, we can begin to understand and appreciate what impacts our patterns of thinking, and try to influence it accordingly.



So, back to our tale of transit. Uber set out to apply theoretical insights from behavioural science to improve their UberPool product. With the desired outcome stated to change UberPool customer’s perception of the service post-request, the team set about researching and listing the behaviours needed to achieve this aspiration. Diving into insights of people’s perceptions of time and waiting, the following biases were pulled out:
- Idleness aversion: people dread idleness and want to be busy.
- Operational transparency: giving consumers transparency of what is happening behind the scenes increases their valuation of products.
- Goal gradient effect: people feel more motivated when they feel like they are advancing towards their objective.
Given their findings, UberLabs needed to match these behavioural biases to desired intervention types. The team recommended:
- Highlighting progress during waiting times to customers.
- Explaining each step going on behind the scenes, such as when other riders travelling the same way had been found.
- Providing additional information (behind an info icon), such as explaining the arrival time estimate calculation.
With this functionality added to UberPool in A/B testing, the team observed an 11% reduction in the post-request cancellation rate, the rationale being customer’s felt they had greater clarity over their journey lead time. Considering they hadn’t actually changed the nature or cost of the service, just the perception of the service, not bad, right?

Working within the world of experience design, employee engagement and creative communications, I began to wonder how behavioural science compliments the common creative approach to problem-solving, known as design thinking. Design thinking, in essence, uses design methodologies and thinking to solve problems. It brings together what is desirable from a human perspective, and what both is technologically and economically viable to create. Design Council’s double diamond (Figure 2), which represents the process of exploring an issue in depth (divergent thinking) and then taking focused action (convergent thinking), is likely the best known design processing tool used globally.

Given behavioural science’s focus on user-centricity, it may be unsurprising to hear the two methodologies complement each other no end. Though in behavioural science the behavioural challenge at hand is often more clearly defined beforehand, a similar stage of exploration is required to identify multiple behaviours which could influence the desired outcome. However, arguably without the influence of Design Thinking, this initial exploration stage would not be as in-depth or wide-ranging. For example, it would have been easy for Uber to assume customers were cancelling pool cars as a result of people not wishing to travel with strangers, the fundamental difference between this service and others which they provide. Therefore, they may have only explored biases associated with human interaction. However, through a more holistic investigation of the problem statement, unconscious decisions and manifested actions associated with the perception of time and waiting were discovered to have the greatest influence.
The application of Behavioural Science within Design Thinking also allows us to ask the deeper why’s of problem-solving. Whereas in Design Thinking, user insights may only be explored or defined at surface level, applying a behavioural insights-led lens puts greater attention on the deeper workings of the human brain, thus shifting the solution direction from how might we stop Uberpool riders cancelling trips post-request, to how might we change UberPool customer’s perception of waiting time post-request. (FYI — some companies, including Uber in other studies, have used this deeper understanding to immorally exploit driver’s thinking patterns). To me, the two methodologies are mutually interdependant — each need each other, whether being used explicitly or implicitly by behaviorists and designers. Therefore, if you’re like me and had never consciously considered biases at the beginning of a project, here’s a few of my favourites (picked from Coglode’s funky behavioural nuggets cards) to get you started, which I believe are applicable to any design-led campaign:
- Storyteller bias — we’re more persuaded by and better recall those who tell stories, and use narrative to create powerful, human connection.
- Collection bias — we have an emotional need to amass sets of related objects or experiences.
- Fluency shortcut — easy-to-repeat messages or statements that are easier to understand are more believable and more likely to be spread by others.
- Humour effect — we’re more motivated by and remember things which make us laugh.
Evidently from the case studies above, behavioural science is an extremely powerful tool, which when used morally, can positively influence the behaviours and decisions of individuals, groups or society as a whole. Through consistent consideration of heuristics such as the above, we as designers can continue to develop our methodologies and practices to ensure a true user-centred approach is taken, and delve into the deeper ‘whys’ of human thinking. Perhaps we need to start thinking about re-drawing out those diamonds…
*for anyone interested, 4.5 stars is my actual Uber rating.


