Using behavioral science to improve your product | Kristen Berman (Irrational Labs)

Episode Summary

Executive Summary: Kristen Berman explains how behavioral economics and design help products change user behavior by targeting specific actions, removing barriers, and creating immediate benefits. Through examples from TikTok, One Medical, fintech, and lending, she shows that product teams often overestimate rational decision-making and should instead diagnose behavior, test interventions, and align incentives to improve outcomes ethically.

Main Topics: Behavioral economics vs. traditional economics (Priority: 5/5): Berman contrasts the assumption that people are rational decision-makers with the behavioral science view: people are emotional, present-biased, and driven by predictable psychologies that products can design for. The 3Bs framework (Priority: 5/5): Her core framework for behavior change: define the specific behavior, identify barriers (logistical and cognitive), and create immediate benefits that motivate action now. Behavioral diagnosis and experimentation (Priority: 5/5): Teams should map the user journey step-by-step, observe what people actually do, and test multiple variants rather than relying on interviews or intuition alone. Case study: TikTok misinformation reduction (Priority: 4/5): TikTok reduced misinformation sharing by combining labels, friction, and a confirmation prompt at share time, showing how slowing users down can change behavior. Case study: One Medical onboarding (Priority: 4/5): One Medical increased doctor bookings by simplifying choice and guiding users to an immediate appointment during onboarding, making the product's value concrete right away. Ethics and incentives in behavioral design (Priority: 5/5): Berman warns that the same tactics can be used manipulatively if teams optimize the wrong KPIs; customer outcomes and longer-term incentives matter most. Practical guidance and learning resources (Priority: 3/5): She recommends starting with a behavior workshop and behavioral diagnosis, using research tools like Google Scholar, and learning through Irrational Labs' boot camp and frameworks.

Key Arguments: People are not fully rational; product design must account for psychology, emotion, present bias, and social norms. The most important first step is defining the exact behavior you want to change; vague outcomes like "engagement" or "login" are insufficient. Reducing friction usually helps, but sometimes adding deliberate friction can improve outcomes when it creates reflection or attention at the right moment. Questions in signup flows can increase conversion if they make users think about the product's benefits before committing. Behavior change works best when benefits are immediate and concrete, not just long-term or abstract. Behavioral diagnosis helps teams map real-world user steps and identify where barriers actually occur. Incentives shape how behavioral tactics are used; measuring the right customer outcome prevents dark-pattern behavior. Default settings and automation are often the strongest interventions in finance because they reduce both logistical and cognitive effort.

Data Points: Irrational Labs team size: 20 behavioral scientists - Berman says the company has grown to a team of 20. Google embedded engagement: around 3 years - She says Irrational Labs was embedded within Google for roughly three years. Google team coverage: over 25 teams - Work at Google spanned teams from self-driving cars to YouTube. TikTok misinformation share reduction: 24% - A label plus share-time confirmation pop-up reduced misinformation shares. One Medical booking increase: 20% - Onboarding changes increased doctor's appointment bookings. Credit Karma recurring deposit setup increase: 18% - Behavioral design work improved recurring deposit setup. Fintech budgeting test size: 10,000 people - Large experiment used to test budgeting variants vs. control. Budgeting experiment result: 0 change in average spend - Budgeting features did not change spending behavior. Budgeting experiment result: 0 change in spend variability - No reduction in variability was observed. Open-text field removal impact: 40% page-over-page conversion improvement - Removing an open text field improved conversion. Quiz completion impact: 53% vs. 37% purchase rate - People who completed a quiz were more likely to buy a title-care device. Trunk Club conversion lift: 133% - Berman cites a public example of question-based signup improvements. Lending Club target bump: 5-point bump - She was incentivized to improve borrower conversion by five points. Lending Club achieved result: 4.5 points - They nearly hit the target but missed by half a point. Elevator study delay: 16 seconds longer - Making elevator doors close more slowly increased stair use. Course length: 8 weeks - Irrational Labs' self-paced boot camp spans eight weeks. Course modules: 16 modules - The boot camp includes 16 modules and homework with Slack feedback.

Pivotal Quotes: "people make decisions with lots of emotion. We are present bias. We overweight our present selves." — Kristen Berman: Defining behavioral economics and why it differs from classical economics "the only wrong answer here is log in" — Kristen Berman: Explaining that teams must define the specific behavior they want to change "We are what we measure and it really matters what you measure." — Kristen Berman: Discussing incentives and the ethical risks of behavioral design

Implications: Product teams should stop optimizing vague metrics and instead design for specific behaviors, using diagnosis, testing, and immediate value. The same tactics can help users or manipulate them, so incentives and customer outcomes must guide implementation.