Freakonomics Radio

Are Two C.E.O.s Better Than One? (Update)

October 29, 2025

Key Takeaways Copied to clipboard!

  • Research by CEO advisor Mark Feigen suggests that large public companies with co-CEOs delivered annual shareholder returns nearly 40% higher than those run by solo CEOs, despite the model being rare in public markets. 
  • Successful co-CEO partnerships, like that of Atlassian's founders, rely on compatible but non-identical skill sets, shared values, and established mechanisms for resolving inevitable disagreements. 
  • Skeptics argue that the co-CEO model suffers from role confusion and a lack of unity of command, especially during crises, while proponents suggest it doubles leadership capacity and provides crucial mutual accountability that sole CEOs lack. 

Segments

Introduction to Co-CEO Trend
Copied to clipboard!
(00:00:00)
  • Key Takeaway: Recent high-profile appointments at Spotify, Oracle, and Comcast signal a renewed interest in the co-CEO leadership model.
  • Summary: The episode revisits an earlier discussion on co-CEOs, prompted by recent corporate announcements. The central question posed is whether companies led by co-CEOs perform better than those with solo leaders. The segment promises expert opinions both for and against the model, alongside real-world examples.
Feigen’s Research on Performance
Copied to clipboard!
(00:03:46)
  • Key Takeaway: Mark Feigen’s study of 2,200 public companies found co-CEO pairs delivered annual shareholder returns nearly 40% higher than solo-led firms.
  • Summary: Feigen analyzed 95 instances of co-CEOs in large public companies between 1996 and 2020, noting the rarity compared to private firms. The significant performance difference led him to believe the model deserves a fresh look due to the increasing complexity of the CEO role. However, the study lacked robustness checks to rule out other confounding factors.
Benefits of Partnership and Ego Check
Copied to clipboard!
(00:09:49)
  • Key Takeaway: Co-CEOs coach each other, ground kooky ideas, and prevent the ego-driven over-aggression common when sole CEOs are certain of their success.
  • Summary: The primary benefit of co-CEOs is having a partner to bounce ideas off and call out flawed thinking, mitigating the risk of an unchecked CEO becoming overly invested in a bad idea. This partnership structure is naturally common in small private firms and even mirrored in family dynamics. The model may lead to more cautious risk-taking, but ensures alignment when risks are taken.
BlackBerry Co-CEO Failure Case
Copied to clipboard!
(00:12:14)
  • Key Takeaway: The RIM co-CEO partnership between Lazaridis and Balsilli dissolved into strategic impasse over hardware versus services strategy when facing existential threat.
  • Summary: Jim Balsilli described his 20-year partnership with Mike Lazaridis, where they divided roles between commercialization and engineering, as highly intimate and successful during growth. The partnership failed when the board sided with Lazaridis’s hardware focus over Balsilli’s view during the iPhone/Android disruption, leading to Balsilli’s departure. Balsilli maintains no bitterness, viewing the experience as a privilege despite the company’s subsequent decline.
Atlassian Co-CEO Success Factors
Copied to clipboard!
(00:28:00)
  • Key Takeaway: Atlassian’s co-CEOs, Cannon Brooks and Farquhar, succeeded by having overlapping but non-identical skills and a shared value system, allowing them to take breaks and share executive support.
  • Summary: The Atlassian founders used their partnership as a superpower, noting that two three-quarter CEOs are better than one singular CEO. They explicitly divided responsibilities (product/engineering vs. go-to-market) and used their relationship to navigate crises like the 2008 financial panic without internal sniping. However, Scott Farquhar’s 2024 departure suggests the model is vulnerable to relationship breakdown or succession issues.
Skepticism from Jeffrey Sonnenfeld
Copied to clipboard!
(00:33:20)
  • Key Takeaway: Jeffrey Sonnenfeld argues that co-CEOs create ambiguity in command, citing historical examples like Microsoft under Gates/Ballmer where a unitary leader (Nadella) ultimately proved more effective.
  • Summary: Sonnenfeld believes CEOs need community for candid feedback, which co-CEO status doesn’t inherently solve, and that unity of command is crucial. He dismisses Feigen’s performance data, stating that authentic co-CEO arrangements are rare, often being ‘in name only’ for retention, and points to disastrous outcomes like Nordstrom’s shared leadership. He emphasizes that history favors monuments to bold individuals, not committees.
Pair Programming Analogy
Copied to clipboard!
(00:43:52)
  • Key Takeaway: Research on pair programming shows that two programmers working together produce higher quality code with only a slight increase in time, effectively debunking the ’twice as expensive’ myth.
  • Summary: Lori Williams’s research demonstrated that pairs produced significantly fewer defects than solo programmers, making the overall economic outcome positive when factoring in defect-fixing time. The practice also resulted in higher satisfaction because humans are social and benefit from constant communication and idea checking. This research provided ‘permission’ for companies to adopt the practice despite initial cost concerns.
Future Outlook and Conditions for Success
Copied to clipboard!
(00:48:05)
  • Key Takeaway: The co-CEO model is likely to remain rare in the Fortune 100 due to board fear of failure, but success is predicted when partners have time to build rapport before taking the top job, as KKR demonstrated.
  • Summary: Feigen predicts that while boards fear the risk of a co-CEO blow-up, the model will grow slowly, potentially reaching 5-10 instances in the Fortune 100 within a decade. A key success factor is giving potential co-CEOs time to work together in subordinate roles (like co-presidents) before officially sharing the top title. This prior experience helps nurture the necessary partnership and shared vision.