Deploying the three Bs with Kristen Berman: to change behaviours, one must decrease barriers and/or increase benefits

On 23 February 2022, the Behavioural Innovations Society hosted behavioural science practitioner Kristen Berman for a discussion over Zoom.

Kristen co-founded Irrational Labs with Dan Ariely in 2013. Irrational Labs helps companies and nonprofits understand and leverage behavioural economics to increase the health, wealth and happiness of their users.

Kristen also co-founded Common Cents Lab, a Duke University initiative dedicated to improving the financial well-being for low to middle income Americans. Under Kristen’s leadership, Common Cents has launched over 50 experiments with companies, touching tens of thousands of people. Kristen’s work has been featured in The Stanford Innovation Review, TechCrunch, and Scientific American.

Prior to co-founding Irrational Labs, Kristen was on the founding team for the behavioural economics group at Google, where she hosted one of the top behavioural change conferences globally, StartupOnomics.

She has co-authored a series of workbooks called “Hacking Human Nature for Good: A practical guide to changing behavior,” with Dan Ariely. These workbooks are being used at companies including Google, Intuit, Netflix, Fidelity, and Lending Club for business strategy and design work.

Before Google, Kristen was a Senior Product Manager at Intuit and a startup, Lytro. In these roles, she built product management and marketing systems for small businesses and consumers, for domestic and international markets.

And in her personal life, Kristen lives behavioural science. She’s started a communal living environment called Radish, to design life for happiness.

Kristen opened the call by pointing out that she felt that the terms “behavioural science” and “behavioural economics” were largely interchangeable. Both leveraged the fact that, in practice, the environment influences our behaviour more than our beliefs do. She illustrated this with reference to M&Ms: if a packet were nearby with the lid off, she would eat more than if it were closed and out of sight. Her beliefs would not have changed but her preferences would have.

Irrational Labs practices applied behavioural science, designing products and services that align with natural behaviours. She gave the example of working with TikTok to mitigate the spread of disinformation. In this case, building on the work of David Rand from MIT, the recommendation was to add friction to the act of sharing that caused potential sharers to pause and reflect in a “cold” state. This friction took some of the “heat of the moment” out of the decision to share, and resulted in a 24% reduction in questionable content.

Generalising across her experience, she nutshelled it as “the three Bs:” to change behaviours, one must decrease barriers and/or increase benefits.

Kristen then described her professional journey. She started her career as a product manager at Quicken and Intuit. She reflected on her experience running focus groups to test a product redesign. A breakthrough moment was one time after she and the team had worked particularly hard to develop a new product aggregating financial information across accounts. In focus groups to test the new product, customers literally covered their eyes to avoid seeing the picture it presented. They hated it!

In retrospect, as a behavioural scientist, it was obvious why they hated it. The tool was showing people exactly what they did not want to see — how little money they had, and much money they were wasting — in a clear and beautiful way. This was “the essence of information aversion, like stepping on a [weighing] scale and seeing a bad number.” The team “nailed it so much that the customers ended up depressed and emotive.”

Understandably, Kristen became frustrated, and was considering her next step. Coincidentally, at around that time, Dan Ariely released his first book, “Predictably Irrational.” So she sent him an email and asked what she could do to help him, in order to learn about applied behavioural science.

With hindsight, she sees asking how she could help him as a powerful framing. She recommends that budding applied behavioural scientists emulate it when approaching leaders in their field. Channelling JFK perhaps, one should ask not what someone can do for you, but what you can do for them. Timing also played a role of course.

For the next two to three years, she worked for Dan for free, alongside her day job, even as she moved to a startup. Together they planned and executed conferences and research.

Again with hindsight, this was pretty much the only way to get into the field, as there were no Master programmes at that time and it was frowned upon to go into industry after a PhD.

As it happened, Google reached out to Dan wanting to hire him as close to full time as they could get him while he was at Duke. Together he and Kristen started up Google’s behavioural science function, working with adwords, self-driving cars and apps for example. There was a lot of experimentation, as there was no playbook to follow.

Dan and Kristen then founded Irrational Labs to take on the consulting work that came in. Each problem presents uniquely and each project is bespoke; while there is the common thread of behavioural science in all solutions, there is no single standard answer.

Over time the work at Google wound down. Kristen and Dan co-founded Common Cents Lab with the idea of focussing their efforts on financial health and well being. Common Cents’ sought grants to perform research, often with fintechs such as Digit. Many of the experiments addressed saving and debt reduction.

More recently, Irrational has been working with Intuit, Microsoft LinkedIn, Airbnb and others, applying a behavioural science lens to product and marketing roadmaps.

On the topic of “rounding up” features in fintech apps, her perspective was that it was a smart way to encourage people to save, as they find it easier to save small numbers than large numbers. Moreover, “rounding up” was a familiar concept from the gas pump, for example. That said, most of the time, the amounts saved are too small. So people think they’ll create an emergency fund but might end up with only $100. Her client Digit solves this problem fairly well by taking larger amounts. It hurts a little more but adds up to a more useful sum.

Asked about Radish, Kristen described a misalignment between what we spend money on in housing and what makes us happy. Typically we spend money to on better school districts, bigger kitchens, and nicer bathrooms. While these are all desirable things, it is likely that we’ll adapt to them — in the sense of taking them for granted — sooner rather than later.

A global finding on happiness is that it is driven by social support and relationships. So if you were to attempt to design a life to maximise happiness, you would look to increase the number of quality relationships. Building on behavioural science insights, you’d make it easy to do this, so that the path of least resistance was to have more quality relationships.

In practice, this is the exact opposite of what we do: when we graduate from college we move away from our friends and family to wherever we get a job; when we have children we move even further away from our friends. David Brooks has a saying that “the more money you have the more loneliness you can buy.”

Radish is a reaction to this. Radish is a house with a mix of apartments and a communal kitchen in which 14 people live together. Most nights the residents share dinner. Food is ordered jointly through a Slack channel. Occasionally Kristen might want some more alone time, but in general there’s more upside in the form of closer relationships. Kristen summed it up as: “Radish is an attempt to design life so we live closer to each other and relationships are easier.” It’s hard to do, and has consumed a lot of time and money, in part because our societal built infrastructure, such as zoning laws, is not designed to accommodate communal living.

Kristen shared a post describing Radish.

The audience were then asked “what do you believe is the most pressing behaviour change problem that society has?” The results appeared via Menti as a word cloud, with obesity, misinformation and climate change featuring prominently.

Kristen shared her perspective on the question, starting with the universal human tendency to indulge in motivated reasoning to appease our confirmation bias. She summarised this as: first, seeing something we think to be true; then creating reasons why it must be true. We find it easy to believe that we are right and others are wrong, in part because we “see” what we’re thinking and intending but we don’t see other people’s thoughts or intentions. It’s much harder for us to see where other people are coming from. Across society, this tends to polarise views.

Asked how this might be counteracted, she suggested taking a third party perspective, or — better still — actually talking to a third party. Over time, this would train the brain to think of the other side of an argument.

These are difficult things to do in practice, particularly in the heat of the moment. Kristen recommended Julia Galef’s book “The Scout Mindset” and her own article in Scientific American.

Asked about how behavioural science might be relevant to climate change, Kristen shared her view that it was a supply chain and policy issue rather than a consumer behaviour change problem. Part of the solution could come from innovation, for example Impossible Burger. And of course we should all turn the lights off when we leave a room and take shorter showers …

The above notwithstanding, there are small behavioural changes that could cumulate to big difference in climate impact. For example, a large number of people pressuring elected representatives to make relatively modest policy changes could have a large impact.

With respect to obesity, Kristen pointed out that it’s both a significant challenge and a really hard fight. 70% of the US population is overweight or obese. However, it’s hard to blame people for their choices given that unhealthy food is the norm and tastes better, while access to healthy food can be limited. Moreover, compared to “one-off” behaviour changes like auto-enrolment for a pension, changing eating behaviour at and between every meal is much harder (barring a switch to Soylent).

To address the problem, we may have to regulate sugar more than we currently do. This shouldn’t be too hard, given the existence of a clear negative externality. In addition, we may have to change norms around eating. The emergence of new “micro-movements” including veganism and organic farming at the very least lets people talk about food in a different way.

The audience were then asked “what behaviour change area or problem are you working on, or want to learn more about?” The results appeared via Menti as a word cloud. In this instance, there was a broad spread of topics.

Kristen felt that this spread was reflective of the work of Irrational, in that they work on many issues at once.

One area of current focus was digital healthcare. For example, data suggest that patients are more likely to show up for a virtual appointment than a physical one. Likewise, adherence to prescriptions is found to be lower in digital environments, except when delivered to the home. Conversely, there’s a broad benefit to enhanced healthcare access, especially in rural areas.

Kristen described Irrational’s work with a home heart rate monitor. The idea is that instead of patients having to visit a doctor simply to have a heart checkup, this can be done at home. In practice, setting up the device was too hard for many users. As a result, Irrational recommended changing the onboarding process and redesigning communications, leading to a significant increase in device utilisation within three months. This was despite not being able to conduct a full RCT as there was simply “too much noise in the system.”

Commenting on the “organisational behaviour” topic, Kristen explained how Irrational advises some clients on how to set up their hybrid working models. In general, the solutions align with the intuition that more choice is helpful, and that people will figure out what works best for them and optimise their choices.

That said, “you can work when you want, where you want,” is probably not the right answer. For example, a meeting that his half-virtual and half-physical is likely to be a worse meeting, as those on the phone feel excluded, and those in the room feel slowed down. What’s important is coordinating times at which team members do “deep work.” This takes away some choice about the timing of deep work, but it provides coordination such that team members know that meetings will not be scheduled in those time slots. Similarly, norms about punctuality enable coordination.

As a more general point, Kristen pointed out that as systems get bigger, there’s a tendency to add rules. This reduces experimentation and increases bureaucracy. One way they work with companies is to try to help them kill excess meetings and avoid the tendency to share everything all the time. Some of this simply involves using tools such as Slack and email fully and appropriately. Kristen referenced Leidy Klotz’s book “Subtract” as a countercultural guide to improvement through subtraction rather than addition.

More conventional Q&A followed, starting with “to what extent do you use insights from neuroscience in your work?” “Almost never!” was Kristen’s response. Having spoken to neuroscientists, she felt that in general we simply don’t know enough about how the brain works. Yes, we know how dopamine and some parts of memory work; but we don’t know enough to understand how to change behaviour yet. Elon Musk’s latest venture, Neuralink, is an exciting development, but it’s only “a little bit” of mind reading — replicating electronically what someone is thinking when they’re trying to move a limb. Applications of AI and ML will rapidly change the state of our understanding.

In response to a follow up question on the reverse direction — neuroscientists leveraging progress from behavioural science — Kristen responded that neuroscience felt less applied. Neuroscientists work with five or ten brains over years, whereas behavioural scientists work with very large sample to estimate average effects. Both sides could learn a lot from each other — but don’t, yet.

Considering the question “what are your thoughts on the ethics of nudging in a consumer context?” Kristen pointed out that many if not most aspects of a consumer retail experience are already “manipulated” (for example, what is on which shelf in the grocery store). As a result, they tend not to look at the ethics of design itself. Instead they look at the incentives of the designer to drive certain behaviours.

She then talked about her work with Lending Club several years ago. At that time it was a peer-to-peer co-lending platform. Irrational proposed that their compensation be proportional to the increase in conversion … “and I turned into a predatory lender!” A different incentive environment changed her behaviour quickly. Working for Peloton, where corporate and personal incentives are more aligned, was easier.

Asked about “the growth of information technologies leading to the modification of the concept of truth,” Kristen responded that her greatest concern was the idea that there was in fact “one truth,” as most of the time there’s nuance and multiple viewpoints. Specifically, her worry with “a globalising noisy world is that the nuance of truth is lost, and [even as that happens] people are less curious and more definitive.” Nuance is not rewarded on Twitter. She illustrated how competence is judged by competence, so that a doctor that conveys higher confidence, regardless of the base rate, is perceived to be better than one who conveys the true odds. Her remedy would be to increase the rewards for nuance and questioning.

The final question concerned the environmental impact of behaviours in an increasingly online world. Kristen responded with a higher level point about changing the environment rather than trying to intervene in every daily decision.

There was a follow up question on the differences in behaviour online vs in person. Kristen noted that there were clearly differences by channel. For example, if you ask someone to do something, they’re more likely to say yes if asked in person than by text. More generally, people are nicer — act with higher reciprocity — when not anonymous, and changing behaviour often comes down to reducing friction and increasing present bias in a given context.

Kristen closed by sharing a post on how to get into behavioural science and mentioning Irrational’s bootcamp. In exchanges after the webinar, she kindly shared a discount code that is available on request.

The moderators, Duncan Martin and James Newport, thanked Kristen for her time and insight and the audience for their interest and questions.