Is it possible to change behavior? Behavioral Science applied to digital products
This article focuses on “CREATE,” a behavioral design framework conceptualized by Stephen Wendel and discussed in his book “Designing for behavior change.” My goal is to describe the framework through various examples while highlighting its connection to the digital product development cycle, starting from the discovery phase.
Over eight years have passed since Designing for behavior change was published (originally in 2013, revised in in 2020). Nevertheless, it remains one of the main references for the application of Behavioral Sciences in product development. Its author, Stephen Wendel, is the Head of Research at HelloWallet, a startup that offers financial guidance and helps people put these tips into practice, using behavioral design strategies — a very interesting multidisciplinary intersection:
And it was based on this work, and the analysis of several other companies — such as FitBit, Nest, and Opower — that he conceived the CREATE framework. Its goal:
- Understand how our brain makes decisions, and how it affects habit and behavior change.
- Identify which specific behaviors to target, taking into account both our own objectives and those of our clients.
- Design products around those insights; and
- Refine the impact achieved through metrics and data analysis loops.
But what does “behavior change” actually mean? And how can we do it?
Wendel emphasizes a need to understand the nuances involved in the context and decision-making processes, and then act upon them — that’s why he prefers the verb “untangle” rather than “break” throughout the text.
Imagine that each behavior is the result of a combination of two simultaneous elements — motivation and ability —then activated by a trigger. Thus, Wendel considers that the starting point for a product discussion should be the action or behavior that we are trying to change, not what the product should do to achieve that goal. When it comes to digital products, this is a significant paradigm shift for the discovery phase.
01. Understanding
We tend to think that all of our actions and decisions are intentional. For example, we wake up at 8:30 tomorrow because it’s a quieter time to write; we’ll attend the next Neuroscience conference because we need to update ourselves in the field; we’re eating this burger to satisfy our hunger.
However, reality is way more complex than that.
I’ve mentioned before that we spend most of our time in an autopilot mode— the result of the limbic dominance of our brain (Damasio, 2012) and also of what Daniel Kahneman classifies as “System 1 thinking,” that is:
a set of past experiences and simple rules that almost immediately point us in a direction to follow; a kind of “instinctive” evaluation that we generally make effectively in familiar contexts where these experiences are relevant.
Our brain classifies the situations we are exposed to as either a threat or a reward, which can limit our field of sight, cognitive ability, propensity for teamwork, and creativity. Furthermore, these resources, as well as memory and attention focus, are restricted by competing and simultaneous stimuli. All of this combined makes us seek out shorter (or familiar) paths to decision-making, saving energy for the body. Thus, habits can be understood as a consequence of this system, functioning predictably and outside of our conscious control:
Based on these observations, Wendel developed the “CREATE” framework, an acronym for:
- Cue
- Reaction
- Evaluation
- Ability, and
- Timing
When these five pre-conditions are present, a person tends to take action. The problem is that throughout each of them, there are countless distractions that can discourage action — whether because we shift our attention, or because we do not even notice an important element of the context.
Thinking about digital products, we can create solutions to assist this process:
- Through “brute force” (in the author’s own words), supporting conscious choices every time the person has to make a decision.
- Through habits, which are related to the more automatic and intuitive part of the brain. Or
- Through small tools (Wendel uses the term “cheating”), which transfer the burden of a task previously performed by the user to the product — that is, once the user gives their consent, the product itself makes a certain choice without the user having to take any action at all.
02. Discovery
Based on the understanding of any nuance that characterizes our decision-making process, the next step in Behavioral Design involves a clear vision of:
- Results: what will be different if our product is successful? Here, the focus is on behavioral issues. For example, an app that emphasizes the importance of maintaining financial reserves, but does not change the savings behavior of its users certainly did not succeed in its proposal.
- Agents: for whom this change is designed? Usually, users of our product.
- Action: how will this change be made (the behavior involved)? In the previous example, related to long-term savings, how can our users achieve this goal? There are several possibilities: automatic enrollment in private pension programs, at the time of signing a job contract (as Thaler & Benartzi did in the USA, through SMarT); emotional associations related to creating a better future for children; etc. This needs to be done without any ambiguity: if we seek to improve the health of our users, will this be achieved by practicing more exercise or choosing healthier foods? Or both? Lack of clarity can result in dispersion and, consequently, inertia regarding the stimulus provoked by our product.
At this stage, Wendel suggests starting with paths of least friction, that is: what actions our users already perform and that can somehow lead them to new results. Thus, we start from a set of already consolidated habits or behaviors and amplify them towards a better outcome.
In addition, we need to learn about these people: what motivations do they have to act or avoid acting? Do they already have experience with similar situations or products? Who will be our target audience, since different groups may respond differently to the stimuli we are creating? How much are they already interested in the topic and willing to make a change?
03. Design
The framework determines three parts of the decision-making context that can be shaped to generate behavior change:
Action
Break down the action into smaller steps that are compatible with the user’s level of experience and interest. They should be simple enough to remove as much work as possible from these people and offer immediate feedback to make it clear whether they are on the right track or not.
Environment
Work on people’s motivation by reminding them of their reasons for acting and changing a certain behavior; create additional incentives if necessary. Highlight actions that need to be taken or place small reminders that can be associated with other contexts with which they are more familiar. It is necessary to provide updates on the progress of the process (status): what stage we are at, who is taking care of it, what is the deadline for completion. And, whenever possible, remove elements that may, in some way, compete for the attention of users or impose obstacles — remember the action funnel above and the various stages in which the person may feel stimulated to… simply do nothing!
User’s readiness
Wendel suggests a three-step process:
- Narrative: work on how people see themselves. Make them perceive the action as a natural extension of their identity and remind them of other important changes they have already made in the past.
- Associations: make clear the progress towards the expected result and try, whenever possible, to link this change to other things that they are interested in.
- Education: make sure people have all the information they need to act, and that they know exactly what — and how — to do it.
The combination of these three steps outlines a Behavioral Plan, a document that will be very useful when defining the customer experience map and, especially, when writing user stories.
For the Behavioral Plan to work, however, a comprehensive mapping of our target audience is essential, resulting in a well-defined persona. Otherwise, we will either not have enough information to apply design guidelines, or they will be implemented in a generic way — and therefore, turn ineffective.
04. Refinement
When it comes to human behavior, we have an enormous range of possibilities, making it impossible to predict them in their entirety. Therefore, we must be realistic in our expectations. Despite tests, experiments, and research, something is unlikely to work 100% as expected.
As Wendel (p. 293) states, our main goal is to “minimize these errors.” One way to achieve this, which aligns with the agile development cycle, is through continuous processes of inspection and adaptation (or refinement). Here are some tips:
- Define clear metrics for the results and action triggers.
- Measure the impact of your product in relation to those results and triggers. If this cannot be directly evaluated in the product, you will need to resort to experiments and/or controlled tests (such as RCT) to establish statistical bridges that prove the causal effect.
- Use data to your advantage. Identify where users tend to get stuck in screens, steps, or circumstances. With a customer experience map, you can understand the overall context of these situations and optimize them. After all, an unfinished step can be frustrating and demotivating, potentially causing users to stop using the product.
- Consider behavioral impacts alongside business and technical objectives. This means seeking a different perspective of your product from the most important part of it: people who are going to use it.
- Test it, over and over again. Even if something seems like an excellent idea, you need to be as certain as possible that the solution will have the desired effects once implemented in the real world. There are several examples of apparently good ideas that have resulted in even bigger problems. For instance, there was a case of an Israeli daycare center that faced tardiness issues. Parents picked up their children after the agreed time, disrupting the school’s dynamics and overloading teachers, nurses, and aides. The obvious idea was to apply a fine to those who were late. The non-obvious result was that parents quickly did a cost-benefit analysis and realized that, for a relatively low amount, they could leave their children there for much longer, without having to deal with the ethical dilemmas they faced before regarding the teachers — it was an advantageous trade-off that significantly increased the number of delays.
What’s next
This is one of the many frameworks available for working on behavioral aspects. What caught my attention in Stephen Wendel’s book, and encouraged me to start testing with “CREATE” and write this article, was its adherence to the digital product development cycle. All the steps and tools described here are related to creating solutions and apps, but they add an extremely important perspective, which is often neglected. In the end, the impact on behavior, and its implications for the decisions we make, is what matters.
Feel free to share your experience putting this model into practice in the comments section below 😉
Further references
Akbas, M.; Ariely, D.; Robalino, D. A.; Weber, M. (2016). How to help poor informal workers to save a bit: evidence from a field experiment in Kenya. Institute for the Study of Labor (IZA), Bonn, Germany.
Benartzi, S., & Thaler, R. (2004). Save More Tomorrow: using Behavioral Economics to increase employee savings. Journal of Political Economy, Vol. 112, No. S1, pp. S164-S187.
Damásio, A. (2005). Descartes’ Error: Emotion, Reason, and the Human Brain. Penguin Books.
Gneezy, U., & Rustichini, A. (2000). A fine is a price. Journal of Legal Studies, Vo. 1, Jan.
Laskey, A. (2013). TED Talk: How behavioral science can lower your energy bill.
Kahneman, D. (2013). Thinking, Fast and Slow. Farrar Straus Giroux.
Wendel, Stephen. (2013). Designing for behavior change. O’Reilly.
