Lenny's PodcastUsing behavioral science to improve your product | Kristen Berman (Irrational Labs)
CHAPTERS
- 0:00 – 0:45
Behavioral economics in one minute: why “rational” models fail
Kristen Berman opens with the core premise of behavioral economics: people aren’t purely rational, and emotion, present bias, and social norms shape decisions. The upside is that these patterns are predictable, which makes them designable.
- •Traditional economics assumes rational, emotionless decision-making
- •Behavioral economics merges psychology + economics
- •Common drivers: present bias and social norms
- •Predictability enables intentional behavior change via design
- 0:45 – 1:46
Podcast setup: who Kristen is and what you’ll learn
Lenny introduces the show’s goal—helping builders improve products—and frames Kristen’s work as highly practical, example-driven behavioral design. He previews case studies and themes like biases, behavior change, and ethics.
- •Show mission: improve the craft of building and growing products
- •Kristen leads Irrational Labs and has worked with major tech companies
- •Episode promises tactical examples and measurable outcomes
- •Tease: TikTok misinformation reduction and other case studies
- 1:46 – 3:58
Sponsor messages (Flatfile, Whimsical)
A short ad break highlighting CSV onboarding pain (Flatfile) and collaborative product tooling (Whimsical).
- •CSV import/onboarding errors drive churn risk
- •Flatfile positioned as “flawless data onboarding”
- •Whimsical positioned for collaborative thinking and specs
- •Links and promos for listeners
- 3:58 – 5:45
What Irrational Labs does: behavior change as a product discipline
Kristen explains Irrational Labs as a behavioral science consulting and design shop focused on changing user behavior. She shares how it started with Dan Ariely and how the team now supports many industries and product types.
- •Behavioral science consulting + design + experimentation
- •Work: understand psychologies, design interventions, test outcomes
- •Founded in 2013 with Dan Ariely
- •Now a team of ~20 behavioral scientists
- 5:45 – 6:51
Defining behavioral economics and behavioral design
Kristen defines behavioral economics as a reaction to classical economics by incorporating psychology. Behavioral design applies these insights to real-world product and business problems to shift behavior predictably.
- •Classical econ vs. real human decision-making
- •Humans are emotional, present-biased, and norm-following
- •Behavior is predictably “irrational”
- •Behavioral design operationalizes these insights in products
- 6:51 – 11:40
Fintech budgeting experiment: what users say vs. what they do
A fintech app’s most requested feature—budgeting—showed no measurable impact in a large experiment. Kristen explains why a behavioral diagnosis makes the failure predictable and offers alternatives like defaults and rules of thumb.
- •10,000-person experiment: budgeting features caused “nothing” to change
- •Behavioral diagnosis: map every step required to reduce spending
- •Budgeting demands too much repeated cognitive work
- •Alternatives: defaults (e.g., auto-enrollment) and simple heuristics/rules of thumb
- 11:40 – 16:09
When more friction increases conversion: questions that motivate action
Kristen shares a counterintuitive pattern: adding steps can boost conversion when those steps prompt users to actively imagine benefits. She explains how questions in onboarding can increase motivation—if they’re easy to answer.
- •Asking a question inserts an idea and shifts user attention
- •Quiz-like onboarding can increase conversion (examples: Apartment List, Trunk Club)
- •Case: TytoCare quiz increased purchase rate among completers
- •Avoid high-effort open-text questions; prefer dropdown/multiple choice
- 16:09 – 17:59
Kristen’s path into the field: Intuit, Dan Ariely, and Google’s BE team
Kristen describes discovering behavioral economics through Ariely’s work while she was a PM at Intuit. She then helped build and embed behavioral economics capabilities inside Google across many teams.
- •Reading Predictably Irrational sparked a “light bulb” for product work
- •Realization: product teams reinvent insights a research field already knows
- •Irrational Labs embedded at Google for ~3 years
- •Helped seed Google’s internal behavioral science practice
- 17:59 – 24:47
The 3B model: Behavior, Barriers, Benefits (and cognitive friction basics)
Kristen outlines Irrational Labs’ practical framework used across clients: define a specific behavior, reduce barriers, and increase immediate benefits. She distinguishes logistical vs. cognitive barriers and gives examples like uncertainty aversion and status quo bias.
- •B1: Pick an uncomfortably specific behavior (not “log in”)
- •B2: Barriers—logistical (forms, credit card) vs. cognitive (biases)
- •Examples: uncertainty aversion and status quo effect
- •B3: Benefits—emphasize immediate rewards due to present bias
- •Immediate motivators: completion bias, social desirability, progress feedback
- 24:47 – 28:53
Ethics and incentives: how behavioral tools can become exploitative
Kristen and Lenny discuss the “dark side” of persuasion and how incentive design can push teams toward harmful outcomes. Kristen’s LendingClub story illustrates how optimizing one metric can distort decision-making, and she recommends aligning KPIs to customer-friendly behaviors over longer horizons.
- •Ethical risk often comes from incentives, not the tools themselves
- •Story: performance-based pay made her “think like a predatory lender”
- •Recommendation: set incentives on customer-aligned behaviors, not just conversion
- •Extend incentive time horizons to avoid short-term hacks
- •“Deadlines are a gift” can help people prioritize (when aligned with user goals)
- 28:53 – 32:03
TikTok case study: reducing misinformation sharing by adding friction
Kristen explains how TikTok partnered with Irrational Labs to reduce misinformation spread by targeting shares specifically. The intervention combined labels and a “Are you sure?” confirmation to slow users down, reducing shares by 24%.
- •Define the behavior precisely: reduce shares of misinformation videos
- •For “do less,” add barriers rather than remove them
- •Interventions: unverified label + share confirmation prompt
- •Mechanism: slows hot-state behavior; adds logistical + cognitive friction
- •Result: 24% reduction; aligned with broader platform best practices
- 32:03 – 35:32
How they designed it: literature review, hypothesis, and de-risked testing
Kristen details the methodology behind the TikTok work, starting with research synthesis and then testing multiple prompt variants in quantitative studies before launching limited in-product conditions. She emphasizes comparing conditions rather than asking users if they “like” a design.
- •Start with literature review and expert knowledge (e.g., accuracy reminders)
- •Timing matters: intervene at the point of sharing, not before/after
- •Rapid testing via platforms like Prolific to compare multiple variants
- •Use relative performance across conditions, not single-design preference tests
- •Pre-testing helps choose best options when in-product experiments are limited
- 35:32 – 47:06
One Medical case study: onboarding that drives immediate appointment booking
One Medical wanted “more engagement,” which the team translated into a specific onboarding behavior: booking an appointment immediately after signup. By reducing choice, recommending a provider and near-term time slots, and nudging toward a quick virtual visit, they increased bookings by 20%.
- •Clarify what “engagement” means—teams often disagree internally
- •Target behavior: book a doctor appointment right after signup
- •Diagnosed barriers: choice overload, scheduling effort, unclear mental model
- •Intervention: ask health questions → recommend one provider + limited times (tomorrow)
- •Result: 20% increase in appointment bookings during onboarding
- 47:06 – 52:01
DIY behavioral design: behavioral diagnosis, getting started, and the Bootcamp
Kristen shares how teams can apply behavioral design without a consultancy: align on a specific behavior, then run a behavioral diagnosis (a journey map “on steroids”) using detailed screenshots and psychology overlays. They also describe Irrational Labs’ Behavioral Design Bootcamp and resources like the “Sweet 16” biases.
- •Run a team workshop to define the target behavior with high specificity
- •Behavioral diagnosis: map every step; identify where/why people drop
- •Create a detailed screenshot deck and annotate with psychologies/barriers
- •Use Google Scholar to research what has worked; start with correct keywords
- •Bootcamp: self-paced 8-week course with modules, homework, Slack, office hours
- 52:01 – 56:14
Lightning round: books, media, hiring, and favorite thinkers
A fast set of personal recommendations and opinions: Kristen’s go-to behavioral science books and podcasts, a favorite show, hiring philosophy, and a respected thought leader. She also notes evidence that interviews don’t strongly predict job performance compared to trials and assessments.
- •Book recs: Predictably Irrational, Influence, and Robert Frank’s work
- •Podcast recs: Science of Change and No Stupid Questions
- •Favorite show: The Rehearsal (and discussion of Nathan Fielder)
- •Hiring: interviews don’t predict performance well; prefer skills assessments/trials
- •Thought leader shoutout: behavioral scientist/designer Chris York
