353. How to Optimize Your Apology

Episode Summary

Executive Summary: The episode examines the economics of apologies through John List and Ben Ho’s large-scale Uber field experiment, showing that apologies alone rarely change behavior unless they are perceived as costly and sincere. It also compares private, corporate apologies with public ones, finding that structure, timing, and accountability strongly affect forgiveness and reputational repair.

Main Topics: Field experiments and the study of apologies (Priority: 5/5): John List explains why field experiments in real-world settings are better than lab studies for understanding behavior, and frames the Uber apology project as a natural experiment on customer repair. Ben Ho’s theory of apologies as costly signals (Priority: 5/5): Ho’s model argues apologies are not just cheap talk; they work best when they carry a visible cost, signal trustworthiness, and are used in ongoing relationships. Uber’s experiment on rider bad trips and apology treatments (Priority: 5/5): List, Ho, and colleagues tested eight apology/coupon combinations on 1.6 million riders to measure future spending after a bad ride and whether apologies reduced churn. Public apologies: format, sequencing, and effectiveness (Priority: 4/5): Rutgers sociologist Karen Cerullo’s study of 183 highly visible apologies found that successful apologies start with the victim, end with remorse, and avoid self-justifying context. When apologies backfire or must be used sparingly (Priority: 4/5): The episode argues repeated or poorly framed apologies can worsen outcomes by signaling incompetence, weakness, or insincerity, especially in politics and repeated offenses. Case study: Quayshawn Stewart’s viral apology (Priority: 4/5): A young man’s anti-Semitic video and subsequent heartfelt apology became a real-world example of an effective apology that included victim focus, remorse, and restitution.

Key Arguments: Apologies by themselves are often insufficient; on Uber, the apology-only treatments produced little consistent effect on future spending or trip frequency. Apologies work better when they are costly to the apologizer and when recipients can clearly see that cost, whether financial, reputational, or behavioral. A $5 coupon sent quickly after a bad ride helped recover some lost spending, suggesting that concrete compensation can do more than words alone. Apologies should not be overused; repeated apologies can create negative expectations and can backfire, making the firm worse off than if it had said nothing. The most effective public apologies begin by centering the victim, then express remorse, and ideally include restitution; self-focused or contextualized apologies are usually weaker. Public apologies are shaped by audience and strategic incentives; politicians may avoid apologizing because admissions of fault can be politically costly. A sincere apology can sometimes de-escalate conflict and repair relationships, as shown by Stewart’s response to his viral incident and subsequent community engagement.

Data Points: Bad-trip revenue loss: 5% to 10% - Uber estimated that riders who had a very bad trip spent less on future rides over time. Study sample size: 1.6 million riders - Total riders accumulated across U.S. markets during the Uber field experiment. Treatment groups: 8 groups - Riders were randomly assigned to control, apology types, and coupon combinations. Coupon value: $5 - Future-trip coupon used as the compensation treatment. Tracking window: 84 days - Customer behavior was observed for 84 days after the bad trip and email. Settlement speed increase: around 20% - In states with “I’m sorry” laws, malpractice settlements reportedly moved faster. Moderate injury settlement reduction: around $20,000 - Average decrease in final settlement size for moderate injuries after apology-protecting laws. Major injury settlement reduction: $50,000 to $60,000 - Average decrease in final settlement size for major injuries, including quadriplegia and death. Apology-only effect: weak/inconsistent - The Uber experiment found it was hard to detect significant behavior changes from apologies alone. Public apology effectiveness in sample: less than one-third - Cerullo and Rouhan found fewer than a third of the studied public apologies were effective. Study sample of public apologies: 183 apologies - Visible apologies from 2000 to 2012 with full text and poll data. Viral apology views: 500,000+ views - Stewart’s apology video quickly spread and drew substantial engagement.

Pivotal Quotes: "you know what is the worst part about getting a bad trip? Is I never received an apology." — John List: List explains the customer experience that motivated the Uber apology study. "you really have to squint hard to make the case that apologies by themselves work." — John List: Summary of the main experimental finding from the Uber field study. "what you say first and what you say last goes a long way in whether or not people forgive you." — Karen Cerullo: Her conclusion from analyzing the structure of public apologies.

Implications: For companies and public figures, apologies work best when they are immediate, costly, victim-centered, and paired with real restitution. Empty or repeated apologies risk backfiring; credible repair matters more than rhetoric.