Get Your Weekly Digest

Receive the best podcast summaries and insights delivered to your inbox every week.

We respect your privacy. Unsubscribe anytime.

Freakonomics Radio

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

with Mark Feigen, Jim Balsillie, Mike Cannonbrook, Scott Farquhar, Jeffrey Sonnenfeld, Lori Williams

Published October 29, 2025
View Show Notes

About This Episode

Stephen Dubner revisits the question of whether companies run by co-CEOs perform better than those with a single chief, exploring both supportive evidence and strong skepticism. CEO advisor Mark Feigen and several current and former co-CEOs describe the benefits and pitfalls of shared leadership, while Yale professor Jeffrey Sonnenfeld critiques the model as creating role confusion and undermining decisive authority. Computer scientist Lori Williams adds evidence from pair programming, showing how working in pairs can improve quality and satisfaction, raising the broader question of when two leaders might truly be better than one.

Topics Covered

Leadership and management Co-founder relationships Entrepreneurship Data-driven decision-making Workplace dynamics Team culture Innovation strategy Business growth and scaling Failure and perseverance Scrappiness in business

Disclaimer: We provide independent summaries of podcasts and are not affiliated with or endorsed in any way by any podcast or creator. All podcast names and content are the property of their respective owners. The views and opinions expressed within the podcasts belong solely to the original hosts and guests and do not reflect the views or positions of Summapod.

Quick Takeaways

  • In a study of 2,200 large public companies, the relatively few firms that used co-CEOs delivered roughly 40% higher shareholder returns than those with solo CEOs, though the analysis lacks rigorous academic controls.
  • Co-CEOs can serve as each other's coach and check, reducing the risk of ego-driven, unchecked strategic bets but also risking ugly breakdowns when trust erodes.
  • The BlackBerry case shows how a strong co-CEO partnership can fracture over strategic disagreements, with the board forced to choose a side and the company ultimately collapsing in its core market.
  • Atlassian's co-CEOs describe shared leadership as a "superpower" that allowed them to divide responsibilities, take real time off, and stress-test decisions, though one later stepped down and the stock declined.
  • Skeptics argue that co-CEO arrangements create ambiguity and dilute bold decision-making, and that many so-called co-CEOs in practice function as a single leader plus a retained top lieutenant.
  • Pair programming research suggests that two people working together can produce higher-quality work with similar overall effort and higher satisfaction compared to solo work.
  • The success of co-CEO models often hinges on deep prior relationships, clear role division, and explicit mechanisms for resolving disputes.
  • Boards remain cautious about co-CEOs due to the high visibility of potential blowups, even as founder-led private companies commonly operate with de facto shared leadership.

Podcast Notes

Introduction and framing of the co-CEO question

Emerging trend of co-CEOs in big companies

Dubner notes more large firms are hiring two CEOs instead of one[1:08]
• Examples mentioned include Spotify, Oracle, and Comcast, indicating the practice is not just theoretical
Revisiting an earlier Freakonomics Radio episode[1:23]
• Dubner says they previously made an episode called "Are Two CEOs Better Than One?" and are replaying it as a bonus with updated facts and figures

Origin of the episode idea

Listener email from entrepreneur Zach Levine[2:13]
• Levine describes himself as an entrepreneur who has started and sold several successful tech startups and has long considered having a co-CEO
• He feels strong at most CEO tasks but recognizes there are areas where he is not as strong
Core question posed[2:29]
• Levine asks whether companies run by co-CEOs perform better than those run by solo CEOs

Episode roadmap

Preview of guests and perspectives[2:36]
• Dubner promises a CEO expert who supports co-CEOs and one who thinks the idea is absurd, as well as co-CEOs from successful and failed companies
Inclusion of broader research on working in pairs[3:00]
• He mentions research showing that working in pairs can make people more productive and happier

Pro-co-CEO evidence and arguments from advisor Mark Feigen

Who is Mark Feigen and what does he do?

Feigen's role as CEO advisor[3:41]
• He describes his firm as supporting CEOs with everything they need to be successful, including coaching, governance advice, and investor relations advice
Reputation as "CEO whisperer"[3:58]
• Dubner notes that Fortune magazine has called Feigen the CEO whisperer, highlighting his influence among leaders

Feigen's initial skepticism and research design

Prompt to investigate co-CEOs[4:27]
• A client asked about co-CEOs; Feigen admits he shared the usual skepticism and knew of no examples offhand
Sample of companies analyzed[4:31]
• His team examined 2,200 companies from 1996 to 2020, covering the S&P 1200 and the Russell 1000
• They found only 95 instances of co-CEOs among these large public firms, showing how rare the model is
Examples of public co-CEO firms[4:53]
• He cites Chipotle, SAP, and Research in Motion (RIM), maker of the BlackBerry, as firms that have had co-CEOs

Performance findings from Feigen's analysis

Headline result: shareholder returns[5:52]
• Feigen says CEO pairs delivered annual shareholder returns nearly 40% higher than companies with solo CEOs
• He notes that if a hedge fund could reliably get 40% better returns, it would become extremely large, underscoring the size of the effect
Limitations of the study[6:16]
• Dubner emphasizes that Feigen is not an academic economist and the study lacks the robustness checks of peer-reviewed research
• The analysis did not control for many other factors that might explain the performance differences, such as underlying firm characteristics
• Still, Feigen is impressed and believes the co-CEO model deserves a fresh look given the complexity of modern CEO roles

Historical and conceptual justification for shared leadership

Critique that co-leadership is rare in politics[7:05]
• Feigen says critics ask why there are no co-presidents of countries if co-CEOs are so good
Roman co-consuls as precedent[7:16]
• He cites ancient Rome, where the Senate appointed two co-consuls responsible for administration and the military
• Each consul could veto the other, forcing agreement and effectively making them partners
Evolution toward single strong leaders[7:34]
• Feigen argues that as countries became more militaristic, a top-down male figure became the norm, influencing firms' leadership structures as well
Contemporary team-based rhetoric versus single leaders[8:08]
• He notes that modern organizations talk about being team-based, suggesting a team at the top could be consistent with that philosophy

Feigen on adoption barriers and succession planning

Existing CEOs' likely reluctance[8:20]
• Feigen expects most current CEOs of top firms to say no if asked whether they would like a co-CEO
Greater openness when thinking about successors[8:36]
• He says more CEOs are open to handing their companies to co-CEO successors than might be expected, after hearing the arguments
• Some have asked Feigen about best practices rather than dismissing the idea outright
Framing as doubling capacity, not halving the role[9:03]
• Feigen stresses that co-CEOs are not about splitting the job in half but about doubling capacity in the most important role

Private versus public firms and prevalence of pairs

Why co-CEOs are more common in private firms[9:08]
• Feigen says many private companies are small and started by two people who effectively run the firm together, like siblings or friends
• He tells a story of his mother Lois and her best friend Ellen, who started a travel business and were partners until Ellen died
Power of collaboration in small enterprises[10:32]
• He suggests that informal co-CEO arrangements are everywhere in small businesses like bakeries and pubs, even if not titled that way

Key benefits of co-CEOs according to Feigen

Mutual coaching and grounding[10:41]
• Feigen says co-CEOs coach each other, collaborate, and act as sounding boards for ideas, especially "kooky" ones
• He quotes a co-CEO who said the best part is they can "call bullshit" on each other
Mitigating ego-driven overreach[11:12]
• Feigen links many CEO failures to becoming too aggressive with an idea tied to their identity and past success
• A co-CEO may encourage more cautious risk-taking and ensure both leaders are aligned before big bets
Family analogy for shared power[11:37]
• He compares co-CEOs to parents in many families, arguing most of us grew up with two de facto co-CEOs at home

Dangers and failure modes of co-CEOs

Co-CEO breakups can be ugly[12:25]
• Feigen lists potential pathologies: distrust, formation of internal camps, indecision, and rumor-spreading between co-CEOs
Chipotle example[12:26]
• Founder Steve Ells brought in his best friend Monty Morant, and they became co-CEOs when going public in 2006
• After a 2015 E. coli crisis, they fought and reportedly never spoke again
BlackBerry example foreshadowed[13:07]
• Feigen mentions that when facing iPhone and Google competition, BlackBerry's co-CEOs were close but did not perform when under pressure and their partnership fell apart

Case study: BlackBerry and the breakdown of a co-CEO partnership

Rise of Research in Motion and BlackBerry

RIM's origins and BlackBerry's dominance[13:07]
• Dubner explains that RIM was founded in 1984 in Waterloo, Canada and developed multiple wireless technologies before the BlackBerry smartphone became its breakout product
• By 2007, RIM was Canada's most valuable company with a market cap of $67 billion, and its co-CEOs were widely celebrated

Jim Balsillie on forming the co-CEO partnership

How Balsillie met Mike Lazaridis[13:44]
• They met when Lazaridis was a supplier to the company where Balsillie worked; Balsillie was impressed by Mike's engineering and innovation computing skills
Decision to share the CEO role[14:26]
• They became partners in 1992 and agreed to share the CEO title because of their distinct skills: Balsillie on commercialization and financing, Lazaridis on engineering and product
• Balsillie says they talked every day, often multiple times daily, and does not recall a day in 20 years when they did not communicate, even on weekends
The grind of building the business[15:03]
• He describes them as a typical "10-year overnight success" in technology, emphasizing long, hard work before rapid growth

The intimacy and functioning of the partnership

Shared office and constant coordination[15:15]
• For the first couple of years they shared an office, discussed meetings before and after, and even had each other's voicemail passwords
• When traveling, they shared hotel rooms, illustrating how close and intertwined their work lives were
Anecdote about dodging an angry CEO call[15:43]
• Facing an irate customer CEO, Balsillie hid under his desk and told their assistant Lisa to say he was not there and that Mike was available instead
• Lazaridis simultaneously hid under his desk, and they flipped a coin to decide who would take the call, showing both camaraderie and the reality of shared unpleasant duties

Strategic shifts and disagreement over BlackBerry's future

Competitive terrain changes with Apple and Google[16:19]
• Balsillie describes two simultaneous shifts: Apple launched a high-end phone with a rich e-commerce ecosystem, and Google introduced a subsidized model with cheaper handsets in exchange for data
• RIM operated in the mid-market and carriers pressured them to slash prices or risk being too expensive
Divergent strategies: hardware vs. services[17:32]
• Balsillie believed the phone business was effectively dead and advocated focusing on RIM's strong messaging services business
• He says Mike wanted to double down on hardware, which Balsillie called a "suicide march"; Mike viewed Balsillie's stance as heresy
Board decision and end of co-CEO relationship[17:38]
• The strategic impasse went to the board, which sided with Mike's hardware-focused plan and dropped the services business emphasis
• Balsillie describes that decision as the end of the relationship; he left in 2012, and Mike left a year later; sales later collapsed from $20 billion to $2 billion

Lessons from BlackBerry about co-CEOs and boards

Response to the claim that crises require one captain[19:46]
• When Dubner suggests crises might call for a single captain, Balsillie counters that the board should have forced the co-CEOs to work through their disagreement rather than pick a side
Balsillie on the value of co-CEOs[19:40]
• He says he agrees with many of Mark Feigen's points about advantages of co-CEOs and felt Feigen "spot on"
Specific benefits highlighted by Balsillie[20:20]
• He calls the ability for co-CEOs to be in two places at once "gold," recounting chronic travel-related fatigue but noting they could cover more territories and key meetings
• He notes that major partners expect to see the CEO as a sign of respect, and co-CEOs allowed them to meet this expectation across both supplier and carrier relationships
• On retention, he argues that co-CEOs create two strong leadership "force fields" for key executives, aiding in the retention of the whole executive suite

Marriage analogy and emotional aftermath

Comparing co-CEO partnership to marriage[22:08]
• Balsillie says many executives wished their marriages were as good as his relationship with Mike, emphasizing that each partner felt better because of the other and that "one and one is three" when the partnership works
Absence of bitterness[22:55]
• Asked if he feels regret or bitterness, Balsillie says no, noting he prospered, "changed the world," and was extremely tired by the demands of the CEO role
• He sometimes wonders what would have happened had his strategy been followed but concludes the alternative would mainly have been more years of intense travel and work despite already having more money than he imagined

Further advocacy and conditions for co-CEOs, plus Atlassian example

Feigen on skepticism and selective adoption

Goldman Sachs co-leadership example[26:29]
• Feigen recounts how Goldman leaders John Whitehead and John Weinberg became co-CEOs despite McKinsey's Marvin Bauer predicting failure; instead, it worked for 25 years
Co-CEO not a universal prescription[27:36]
• He stresses that it would be silly to claim co-CEOs always work; they are a choice that should not be dismissed just because of skepticism

Success conditions: willingness and conflict resolution

Need for both leaders to be willing[27:25]
• Feigen says a key predictor is that both people willingly choose power-sharing, a reason founder co-CEOs often work well
Planning for inevitable disagreements[27:36]
• He advises co-CEOs to explicitly discuss how disagreements will be resolved, recognizing that conflict is natural
• He likens unresolved conflict to parents fighting, which is unpleasant for observers, underscoring the need for clear rules

Atlassian founders as co-CEOs

Founding story and initial lack of titles[28:48]
• Scott Farquhar recalls meeting Mike Cannonbrook in a co-op program where they had internships at "uninspiring" large tech employers in Australia
• After graduation, Mike emailed about starting their own company; early on, they simply did all tasks without formal titles
• They only adopted formal titles years later and realized co-CEO was the natural description
Division of responsibilities at Atlassian[29:44]
• Farquhar says Mike runs product, engineering, and design, while he runs go-to-market functions like sales and marketing
• Both have at different times handled "the whole job," suggesting flexibility in their roles
Benefits: rest and continuity[29:38]
• Farquhar notes he was able to take a proper honeymoon and time off because Mike could fully run the business
• He contrasts this with peers at public firms who have not taken more than a week off in a decade or more than 24 hours totally offline
Decision-making dynamic[30:23]
• Farquhar says if he cannot convince Mike something is a good idea, there's probably a good reason; he sees this check as a benefit
• Cannonbrook says they are overlapping but not identical; they share values like openness and honesty but have different skill sets, making one redundant if they were identical or conflict-prone if too divergent

Atlassian under stress and later leadership change

Handling a rocky period together[31:51]
• Farquhar recalls their first really rocky period, saying they quickly made tough decisions together, such as freezing pay rises and stopping hiring temporarily
• He describes the period as one where two people in the boat helped navigate tricky waters without sniping or tension
Subsequent departure of one co-CEO[32:34]
• Dubner notes that in April 2024, Farquhar announced he was leaving Atlassian, amid rumors of a falling out with Cannonbrook that both denied
• He states that since then the company's share price has fallen by about 25%, leaving room for different interpretations about the co-CEO model

Skeptical view: Jeffrey Sonnenfeld on the problems with co-CEOs

Sonnenfeld's background and mission

Role at Yale and CEO institute[33:13]
• He is a senior associate dean and professor at Yale School of Management and runs the Chief Executive Leadership Institute, described as the first school for incumbent CEOs
Purpose of his programs[33:25]
• Sonnenfeld says accomplished CEOs can atrophy in high office; his programs aim to keep them fresh by offering short, pragmatic learning segments matching their attention spans
• He references a Greek adage that "the fish rots from the head" to stress the importance of leaders staying current and informed

Loneliness at the top

CEOs with effectively no friends[34:38]
• Sonnenfeld says a CEO of a large consumer goods company and a CEO of a top-three bank each told him they have no friends, despite being affable
• He notes their lieutenants are ambitious aspirants, making full vulnerability difficult, and regulatory scrutiny constrains what they share with boards and even family
Consultants and hidden agendas[35:24]
• He warns that consultants can act like "corporate intestinal tapeworms" who never leave, and may have commercial incentives to prolong dependence

Rejection of co-CEOs as the solution

Advocating external communities instead[35:51]
• Rather than co-CEOs, Sonnenfeld believes CEOs need off-the-record communities like his institute to provide candid feedback
Concerns about role confusion and unity of command[36:15]
• He argues co-CEOs introduce ambiguity over who is in control and who is the lead spokesperson, undermining unity of command

Critique of Feigen's evidence and real-world cases

Doubts about the 40% outperformance finding[36:21]
• Asked whether he believes Feigen's data, Sonnenfeld answers "No, not at all," rejecting the claim that co-CEO firms substantially outperform
Co-CEO arrangements in name only[37:05]
• He says at firms like Netflix and Salesforce, co-CEO titles were used mainly as retention strategies for talented executives, while it remained clear who truly ran the company
Examples where co-leadership failed or muddled[37:41]
• He claims co-leadership was disastrous at Nordstrom, and at Microsoft it was often unclear whether Bill Gates or Steve Ballmer was really calling the shots
• He contrasts that with Satya Nadella as a clear, unitary CEO who is bold but not autocratic

Questioning the reality of pairs in creative fields

Beatles documentary insight[38:10]
• Sonnenfeld references a long Beatles film showing that Paul McCartney was effectively driving the band while George Harrison and John Lennon had less influence than assumed, and Ringo Starr tried to placate
Paul Simon on Simon & Garfunkel[38:35]
• He recounts Paul Simon saying that Simon & Garfunkel was not a true collaboration because "Artie just showed up and sang," undermining the pair mythology
Great individuals versus committees[39:36]
• Sonnenfeld notes there are few public monuments to committees or task forces compared to monuments for bold individual leaders
• He argues entrepreneurs need courage and boldness; while boards should backstop them to prevent recklessness, clear authority is essential

Pair programming as an analogy for co-leadership effectiveness

Lori Williams's background and solo programming experience

Working alone at IBM[41:11]
• Williams recalls sitting in her office "suffering in silence" at IBM, stuck on problems without help
Transition to academia and discovering pair programming[41:39]
• She left IBM to pursue a Ph.D. at the University of Utah, where her advisor became excited about pair programming in 1998
• She notes that when an advisor's eyes light up, that topic becomes the student's interest too

What is pair programming?

Driver/navigator structure[42:31]
• Pair programming typically involves two programmers at one computer, with one as the "driver" typing and moving the mouse, and the other as the "navigator" watching and guiding
• She emphasizes that it's constant chatter, not a passive observer situation

Experiment comparing pairs versus solo programmers

Study design in a university class[42:58]
• In a 15-week course with 41 students, 28 students formed 14 pairs and 13 worked alone
• Over two-week assignments, pairs had to complete two programs while solos completed one, and they measured quality, time spent, and satisfaction
Results: time, quality, and economics[44:41]
• Pairs spent slightly more total person-hours (about 11 versus 10 for a solo student) but produced work with far fewer defects
• Accounting for time to fix defects, the pairs were economically advantageous overall, not twice as expensive as skeptics feared
Effects on satisfaction and practice adoption[45:18]
• Pairs reported higher satisfaction, both in the study and in industry examples; Williams attributes this to humans being social and getting feedback while working
• She says many practitioners thanked her for providing "permission" to use pair programming by showing it did not cost double, helping overcome managerial resistance

Future of co-CEOs, structural considerations, and closing reflections

Feigen on the potential trajectory of co-CEOs

Predictions for adoption in large firms[49:41]
• Feigen expects co-CEOs to remain common in smaller companies but says large-company boards will be wary of visible blowups
• He does not expect the Fortune 100 to quickly reach high co-CEO penetration but wagers that within about 30 years, perhaps 25% might have co-CEOs
De facto partnerships beneath the title[50:19]
• He anticipates many organizations will have effective partnerships like CEO-president or CEO-CFO pairs who work together constantly, even if only one has the CEO title

Importance of prior joint experience

KKR as an example of staged co-leadership[50:44]
• Feigen cites KKR appointing Joe Bay and Scott Nuttall as co-presidents in 2017, then moving them to co-CEOs four years later after they successfully worked together
• During their joint leadership, KKR's market cap tripled and assets under management doubled, which Feigen presents as evidence that testing the partnership first can work well
Value of long-term personal relationships[51:15]
• He again cites his mother and her best friend, who had known each other since age six before co-founding a business, as an example of how long shared history can underpin trust

Feigen addresses potential conflict of interest

Does advocating co-CEOs help his business?[51:30]
• Dubner asks if promoting co-CEOs simply increases Feigen's client base; Feigen replies that it's actually bad for his business because lonely CEOs are the ones who seek out advisors
• He says he has "zero commercial interest" in co-CEOs and is simply fascinated by the surprising research findings and wants to share them

Episode conclusion and listener engagement

Invitation for listener feedback[51:59]
• Dubner asks listeners what they think about co-CEOs and invites emails to radio@freakonomics.com
Closing remarks on multiple episodes[52:13]
• He suggests that, like co-CEOs, two episodes in a week can be better than one, promising a brand new episode shortly

Lessons Learned

Actionable insights and wisdom you can apply to your business, career, and personal life.

1

Complementary partnerships at the top can improve decision quality, coverage, and resilience, but only when both leaders willingly share power, trust each other, and have clearly differentiated strengths.

Reflection Questions:

  • • Where in your organization or projects could pairing people with complementary strengths create better outcomes than relying on a single "hero" performer?
  • • How might your decision-making improve if you had a peer who could credibly challenge your assumptions before big bets?
  • • What specific steps could you take this month to build or formalize a trusted partnership around your own role, even if titles stay the same?
2

Power-sharing arrangements require explicit mechanisms for resolving disagreement; without pre-agreed rules of engagement, conflict between co-leaders can paralyze the organization or lead to destructive splits.

Reflection Questions:

  • • What are the current unwritten rules you rely on to resolve high-stakes disagreements with peers or partners, and are they actually working?
  • • How could you introduce clearer escalation or tie-breaking processes so disagreements are settled quickly without damaging relationships?
  • • Which upcoming decision would benefit from writing down a simple protocol for how you and key stakeholders will handle not reaching consensus?
3

Leadership roles are structurally lonely, so effective leaders deliberately cultivate external communities and honest feedback channels rather than assuming a co-leader will automatically solve their isolation.

Reflection Questions:

  • • In what situations do you find yourself withholding your real doubts or worries because you lack a safe audience to share them with?
  • • How could building a small, trusted peer group outside your organization change the way you process major decisions and stress?
  • • What is one concrete move you could make this quarter to create a more candid feedback loop around your leadership (for example, a mentor, peer forum, or structured debriefs)?
4

Empirical tests of unconventional structures, like pair programming or co-CEOs, can reveal that intuitively "inefficient" setups sometimes produce better quality, satisfaction, and long-run economics than traditional solo models.

Reflection Questions:

  • • Where are you currently rejecting a nontraditional structure or process because it "seems" inefficient, without having actually tested it?
  • • How might running a time-bound experiment with a paired or shared role help you gather real data instead of relying on intuition alone?
  • • What small-scale pilot could you design in the next 60 days to compare a collaborative model against your current solo approach?
5

Succession planning and shared leadership work best when partnerships are tested over time, with clear role boundaries and progressive responsibility, rather than suddenly dropping two people into a shared top job.

Reflection Questions:

  • • How are you currently identifying and stress-testing potential leadership pairs or teams before they take on major authority?
  • • In what ways could you stage roles (for example, co-presidents before co-CEOs, or shared project leads before formal promotions) to observe how people really work together under pressure?
  • • What is one key leadership transition coming up where you could deliberately give potential successors a trial period of working together before assigning final titles?

Episode Summary - Notes by Tatum

Are Two C.E.O.s Better Than One? (Update)
0:00 0:00