Leadership

The co-CEO model is returning at some of the world’s most complex companies. Its success does not depend on whether two talented people can get along. It depends on whether a board can create shared authority without creating competing sovereignty.

By Frank Farnel | Responsible Public Affairs | September 8, 2026

Executive summary

• The co-CEO model has moved back into serious boardroom consideration. Spotify, Oracle and Comcast all formalized dual-chief structures that took effect in 2026, while Netflix has operated with Ted Sarandos and Greg Peters as co-CEOs since January 2023.
• Research on shared leadership is directionally encouraging but not conclusive for public-company co-CEOs. A meta-analysis of 42 independent team samples found a positive association with team effectiveness; an observational study of 87 co-CEO companies reported higher average shareholder returns than relevant benchmarks. Neither result proves that appointing two CEOs causes superior corporate performance.
• The decisive issue is not personality fit. It is authority design: a common mandate, complementary domains, a defined joint-decision zone, a fast tie-break process, one external position and an independent board capable of resolving disagreement.
• Netflix illustrates the value of a long apprenticeship and complementary expertise. Spotify is testing a related logic, but with an unusually active founder-chair above the pair. SAP and Salesforce show why boards should treat the structure as contingent, reversible and vulnerable to founder asymmetry or crisis pressure.
• A co-CEO arrangement should be selected because the strategy requires simultaneous leadership capacities—not because a board cannot choose between candidates or wishes to postpone a succession decision.

The return of the double act

The conventional defense of the single CEO is easy to understand. Markets want accountability. Employees want clarity. Boards want to know who owns the result. In a crisis, nobody wants to discover that the most important decision is waiting for two diaries to align.

Yet the job itself has become harder to contain within one person. A global chief executive is expected to allocate capital, understand technology, read geopolitics, hold the culture, satisfy investors, navigate regulators, protect resilience and remain credible in public. The answer cannot simply be a superhuman job description followed by disappointment when a human being fills it.

Several prominent companies are experimenting with another answer. Spotify appointed Alex Norström and Gustav Söderström as co-CEOs from January 1, 2026, as founder Daniel Ek moved to executive chairman. The company said the change formalized an operating reality in place since 2023: the two executives had already been leading much of strategy and execution as co-presidents.^[1] Oracle appointed Clay Magouyrk and Mike Sicilia as CEOs in September 2025, while Safra Catz moved to executive vice chair.^[3] Comcast named Mike Cavanagh to serve alongside chairman and co-CEO Brian Roberts, effective January 2026.^[4]

These examples do not establish a trend in the statistical sense, and their recent performance cannot yet validate the design. They do, however, make the governance question immediate: when does sharing the title expand the organization’s capacity, and when does it merely move conflict into the office at the top?

The theory: plural leadership inside a singular office

Academic research distinguishes “plural leadership”—influence exercised by more than one person—from the heroic model in which leadership is embodied in a single individual. A major review by Jean-Louis Denis, Ann Langley and Viviane Sergi showed that plural leadership includes several different arrangements: power may be formally shared, informally distributed, rotated by expertise or negotiated through interaction.^[15] These are not interchangeable.

A co-CEO structure is the hardest version because it places plurality inside the most symbolically singular office in the corporation. Both people carry the same ultimate title. Employees, investors, governments and partners may therefore infer equal competence over every issue even when the internal division is more precise.

The evidence supports neither reflexive enthusiasm nor dismissal. Marianne Döös and Lena Wilhelmson reviewed 67 empirical papers spanning 55 years of managerial shared leadership. They found a real organizational practice but a fragmented research field, with inconsistent terminology and limited cumulative evidence.^[13] Separately, Danni Wang, David Waldman and Zhen Zhang’s meta-analysis of 42 independent samples found a positive overall relationship between shared leadership and team effectiveness (ρ = .34), with the nature of the leadership being shared materially affecting results.^[14] That study concerns teams, not specifically listed-company co-CEOs.

A widely cited Harvard Business Review analysis of 87 public companies led by co-CEOs reported average annual shareholder returns of 9.5 percent during the shared tenures, compared with 6.9 percent for relevant indexes; nearly 60 percent outperformed.^[12] It is valuable counterevidence to the claim that dual leadership must fail. But it remains observational: boards may choose co-CEOs in distinctive firms, compatible pairs may self-select, and survivorship and industry effects are difficult to remove. The responsible conclusion is modest. Two CEOs can work. The title alone tells us very little about whether they will.

The central concept is one center of gravity. A company may have two people exercising chief executive authority, but it cannot sustain two competing sources of organizational truth. Strategy, capital allocation, risk appetite and the external position of the company must ultimately cohere. Shared leadership is therefore not the absence of hierarchy. It is hierarchy redesigned around a relationship and made governable by rules.

The six clauses of a workable co-CEO mandate

Boards considering the structure should draft it with the discipline normally reserved for a merger agreement. Six clauses matter.

Clause	Question the board must answer	Failure signal
Common mandate	What strategic problem specifically requires two chief executives?	The rationale is retaining two candidates or avoiding a choice.
Complementary domains	Which decisions can each CEO make independently, and where is consultation required?	Each CEO informally rebuilds a complete parallel cabinet.
Joint-decision zone	Which matters—strategy, capital, top appointments, risk and public positions—must carry both names?	Major decisions migrate between domains to obtain the preferred answer.
Resolution protocol	What happens when good-faith disagreement persists beyond the decision deadline?	Silence, delay or political escalation through lieutenants substitutes for a decision.
One external voice	How will investors, employees, governments and partners receive a coherent company position?	Stakeholders shop between CEOs or interpret difference as fracture.
Board sovereignty	Who evaluates the pair, arbitrates exceptional disputes and can end the structure?	A founder-chair, dominant co-CEO or faction becomes the unacknowledged final authority.

This framework does not require both leaders to participate in every decision. That would double coordination cost and slow the business. The aim is selective integration: autonomy where expertise and speed matter, joint ownership where the company must be indivisible.

Case study: Spotify formalizes a partnership—and introduces a third pole

Spotify’s design is unusually instructive because it did not begin with two appointments on a blank page. Norström, previously chief business officer, and Söderström, previously chief product and technology officer, had served as co-presidents. Spotify said they had largely led strategic development and operational execution since 2023. The 2026 appointments therefore converted an apprenticeship into formal authority rather than asking two senior executives to discover a working relationship after receiving the title.^[1]

The complementarity is legible. One leader comes from the commercial side of the platform; the other from product and technology. In May 2026, they jointly presented Spotify’s investor-day case, publicly demonstrating that the pair could alternate between a common narrative and their respective expertise. Spotify’s first-quarter 2026 release reported 761 million monthly active users, 293 million premium subscribers, €4.5 billion in revenue and €715 million in operating income.^[2] Those figures are facts. It would be analytically unsound to credit a structure that had existed for one quarter with producing them.

The more consequential design question sits above the pair. Ek did not become a ceremonial former CEO. Spotify stated that, as executive chairman, he would determine capital allocation, map the company’s long-term future and support the senior team; the co-CEOs would report to him.^[1] That arrangement can be a strength. A founder can protect the long arc while the co-CEOs integrate product and business execution. It can also create ambiguity. If the founder sets capital allocation and long-term direction, the organization may reasonably ask where the ultimate CEO mandate begins.

Analysis: Spotify’s test is not whether Norström and Söderström collaborate. Their prior roles offer evidence that they can. The test is whether Ek’s active chairmanship clarifies the horizon or becomes a third executive pole. Employees must know when the co-CEOs decide, when the chair counsels and when the board governs. Without that boundary, apparent dual leadership can become a three-level approval system.

Case study: Netflix builds the relationship before naming the pair

Netflix offers the strongest mature case among the current large-company examples. Reed Hastings elevated Ted Sarandos to co-CEO in July 2020 while Greg Peters became chief operating officer. In January 2023, Peters joined Sarandos as co-CEO and Hastings moved to executive chairman. In announcing the change, Hastings emphasized that the three had worked together in different capacities for 15 years and that Sarandos and Peters had developed trust through shared successes and failures.^[5]

That history matters more than the symmetry of the titles. Sarandos brought deep content and creative-industry authority; Peters brought product, technology and operating experience. The arrangement was preceded by staged delegation, not followed by it. Netflix had time to observe the relationship under pressure before formalizing it.

The company’s financial record since the appointment is substantial. Netflix’s 2025 Form 10-K reported revenue of $45.18 billion, up 16 percent from 2024, and an operating margin roughly three percentage points higher year over year.^[6] Again, this is not proof that two CEOs caused the performance. Content, pricing, advertising, paid sharing and market conditions all matter. It does show that a co-CEO structure is compatible with scale, strategic change and financial improvement.

Netflix also benefits from an organizational culture that explicitly favors judgment over elaborate process. That makes a paradox visible: a decentralized company can sustain two people at the top only if those people are exceptionally clear about context. Freedom lower in the organization increases the cost of contradictory signals from above.

Practical lesson: appointing the pair is the last step, not the first. Boards should observe how prospective co-CEOs make decisions together, share credit, handle disagreement and protect enterprise interests before conferring equal authority. Chemistry described in an interview is not evidence. Joint work under real constraints is.

Case study: SAP chooses unity when the operating context changes

SAP appointed Jennifer Morgan and Christian Klein as co-CEOs in October 2019. The structure ended in April 2020, early in the COVID-19 shock, when Klein became sole CEO and Morgan agreed to depart. SAP’s half-year report records the transition; the company’s public rationale at the time stressed the need for clear, unambiguous steering during an unprecedented crisis.^[11]

It is tempting to label the episode a failure of shared leadership. That conclusion goes beyond the evidence. Six months is too short to isolate the effect of the model, and the pandemic abruptly changed the decision environment. A structure designed for transformation under normal strategic conditions was judged against a crisis demanding rapid integration.

The case is nevertheless important because it challenges the idea that governance architecture should be permanent. The right leadership design can change with the strategic situation. A co-CEO model may add information-processing capacity during portfolio transformation, international expansion or a convergence of technologies. In a compressed crisis with tightly coupled decisions, its coordination costs may rise.

Practical lesson: boards should define review triggers at appointment—an acquisition, activist campaign, liquidity shock, regulatory intervention or severe operational crisis—and state whether the dual structure will be reassessed. Reversibility is not an admission of doubt. It is a design feature.

Case study: Salesforce and the problem of unequal equality

Salesforce has twice experimented with a co-CEO alongside founder Marc Benioff. Keith Block was appointed co-CEO effective August 7, 2018.^[7] He stepped down in February 2020, leaving Benioff again as chair and sole CEO.^[8] Bret Taylor became vice chair and co-CEO in November 2021.^[9] Salesforce announced one year later that he would depart effective January 31, 2023, after which Benioff would again hold the chair and CEO roles.^[10]

Neither departure proves that founder/non-founder pairs cannot work. The official announcements do not establish that ambiguous authority caused either exit. What the sequence does establish is structural: in both experiments, the continuing founder was the stable axis and the other co-CEO role was temporary.

Analysis: formal equality does not erase accumulated authority. Founders hold narrative legitimacy, historical relationships, voting influence and symbolic ownership that a title cannot replicate. When one co-CEO is also chair, founder and permanent public face, colleagues may continue to treat that person as the real principal. The second CEO can then carry operational accountability without equivalent institutional sovereignty.

This is the most dangerous version of shared leadership because the hierarchy is both powerful and denied. The organization receives two formal answers and one cultural answer. Senior executives learn to wait for the founder, route around the partner or use one leader to appeal the other. Resentment grows even when the individuals remain cordial.

Practical lesson: boards must map informal power as seriously as delegated authority. If the asymmetry is intentional, name it and design a president/COO or CEO/deputy structure that reflects reality. Equal titles should not be used to dignify unequal mandates.

What these cases actually teach

First, complementarity is necessary but insufficient. “One understands product; the other understands customers” sounds persuasive, but every major decision eventually crosses that line. Product architecture affects revenue. Commercial commitments constrain technology. The pair needs both divided domains and a joint-decision zone.

Second, trust must be operational. Boards often describe co-CEOs as close colleagues. The relevant question is whether they have reliable practices: a daily exchange without staff, a single briefing pack, a shared chief of staff or tightly connected offices, agreed rules for copying one another and a prohibition on surprise commitments.

Third, speed comes from preauthorization, not constant consensus. Each leader should be able to decide within a defined field. Matters that require both should have a deadline and a route to resolution. “We decide together” is a value; it is not a protocol.

Fourth, the board becomes more important. A weak board can hide behind a powerful solo CEO. It cannot safely hide behind two. Directors must assess the health of the relationship without becoming a court for routine disagreements. They need direct access to both leaders, common performance objectives and evidence of whether the organization is experiencing clarity or faction.

Finally, external stakeholders will test the seam. Investors will compare language. Regulators will seek commitments. journalists will look for disagreement. Major customers may appeal a decision to the other CEO. The pair must disagree honestly inside and speak consistently outside—without manufacturing a false theater of unanimity.

What Leaders Should Do Now

1. Start with the strategic burden. List the distinct leadership capacities the next chapter requires. Choose two CEOs only if simultaneous, peer-level ownership is genuinely superior to a CEO supported by strong executives.
2. Run a live apprenticeship. Give the prospective pair joint responsibility for a consequential cross-enterprise issue before appointment. Observe decisions, not presentations.
3. Publish an internal authority map. Employees should be able to see which decisions are individual, joint, board-reserved or delegated below the top team.
4. Define the tie-break before the tie. Set a decision clock, an escalation route and the narrow circumstances in which the chair or board intervenes. Avoid giving one co-CEO an invisible permanent veto.
5. Use shared enterprise measures. Individual domain metrics are useful, but compensation and evaluation should be dominated by outcomes both leaders own. Otherwise the structure rewards optimization of halves.
6. Audit informal power. Ask senior executives confidentially whose answer they believe is final. A consistent gap between the charter and lived authority is an early warning.
7. Protect one narrative. Coordinate investor, employee, government and crisis communications. Different voices can carry one position; identical scripts are not required.
8. Design the exit. Specify review dates and transition options if one CEO leaves, performance diverges or strategic conditions change. The company should not have to improvise its sovereignty during a rupture.

Conclusion: share the office, not the sovereignty

The single-CEO model survives partly because it converts complexity into a reassuring picture: one person, one mandate, one answer. The picture was never entirely true. Every serious organization already depends on plural leadership across functions, geographies and expert communities.

The co-CEO model makes that dependence visible at the top. Visibility can be an advantage. It acknowledges that strategy now requires different forms of authority at the same time. It can also expose the company to delay, court politics and conflicting commitments if the board mistakes personal goodwill for institutional design.

The choice is therefore not between one leader and two. It is between implicit plurality and governed plurality. Spotify, Netflix, SAP and Salesforce point to the same conclusion from different directions: shared titles work only when the enterprise has one center of gravity. Two people may carry the office. The organization must still know where authority joins, how disagreement ends and who remains accountable for the whole.

Key Evidence

• January 1, 2026: Spotify’s Alex Norström and Gustav Söderström became co-CEOs after operating as co-presidents; Daniel Ek became executive chairman.^[1]
• 87 public companies: an observational HBR analysis reported 9.5% average annual shareholder returns during co-CEO tenures versus 6.9% for relevant indexes; nearly 60% outperformed.^[12]
• 42 independent samples: a peer-reviewed meta-analysis found an overall positive association between shared leadership and team effectiveness (ρ = .34); it did not specifically test public-company co-CEOs.^[14]
• $45.18 billion: Netflix’s 2025 revenue, up 16% year over year; operating margin increased by about three percentage points. These results show compatibility, not causation.^[6]
• Two reversions: Salesforce returned to Marc Benioff as sole CEO after the Keith Block and Bret Taylor co-CEO periods ended in 2020 and 2023, respectively.^[8][10]

Glossary

Co-CEOOne of two people who formally hold the chief executive title and share ultimate executive responsibility.Plural leadershipAn umbrella concept for leadership influence distributed or shared among multiple people rather than concentrated in one individual.Joint-decision zoneThe defined set of matters that neither co-CEO may decide alone, typically including enterprise strategy, major capital allocation, risk appetite and top appointments.Center of gravityIn this article, the coherent source of enterprise direction created by a common mandate, shared facts and a final resolution mechanism.Founder asymmetryThe informal power difference that can persist when one co-CEO is also the founder, chair, controlling shareholder or enduring public symbol of the company.

References and Further Reading

Corporate and regulatory primary sources

1. Spotify, “Spotify Announces Leadership Evolution: Daniel Ek to Become Executive Chairman, Alex Norström and Gustav Söderström to Become Co-CEOs in January 2026,” Spotify Newsroom, September 30, 2025.
2. Spotify, “Spotify Reports First Quarter 2026 Earnings,” Spotify Newsroom, April 28, 2026.
3. Oracle Corporation, “Oracle Corporation Announces Promotion of Clay Magouyrk and Mike Sicilia to CEOs; Safra Catz Appointed Executive Vice Chair,” September 22, 2025.
4. Comcast Corporation, “Mike Cavanagh to Join Brian Roberts as Co-Chief Executive Officer,” September 29, 2025.
5. Reed Hastings, “Ted Sarandos and Greg Peters Are Now Co-CEOs of Netflix, With Reed Hastings as Executive Chairman,” About Netflix, January 19, 2023.
6. Netflix, Inc., “Annual Report on Form 10-K for the Year Ended December 31, 2025,” U.S. Securities and Exchange Commission, filed January 23, 2026.
7. salesforce.com, inc., “Appointment of Keith Block as Co-Chief Executive Officer,” Form 8-K, U.S. Securities and Exchange Commission, August 6, 2018.
8. Salesforce, “Keith Block Steps Down as Salesforce Co-CEO; Marc Benioff Is Chair and CEO,” February 25, 2020.
9. salesforce.com, inc., “Appointment of Bret Taylor as Vice Chair and Co-Chief Executive Officer,” Form 8-K, U.S. Securities and Exchange Commission, November 30, 2021.
10. Salesforce, “Bret Taylor to Step Down as Salesforce Vice Chair and Co-CEO,” November 30, 2022.
11. SAP SE, “SAP 2020 Half-Year Report,” 2020.

Research and theoretical works

12. Marc A. Feigen, Michael Jenkins and Anton Warendh, “Is It Time to Consider Co-CEOs?” Harvard Business Review, July–August 2022.
13. Marianne Döös and Lena Wilhelmson, “Fifty-five Years of Managerial Shared Leadership Research: A Review of an Empirical Field,” Leadership, vol. 17, no. 6, 2021, pp. 715–746.
14. Danni Wang, David A. Waldman and Zhen Zhang, “A Meta-Analysis of Shared Leadership and Team Effectiveness,” Journal of Applied Psychology, vol. 99, no. 2, 2014, pp. 181–198.
15. Jean-Louis Denis, Ann Langley and Viviane Sergi, “Leadership in the Plural,” Academy of Management Annals, vol. 6, no. 1, 2012, pp. 211–283.

Source and methodology note

This article was researched through September 8, 2026. It prioritizes company announcements, securities filings and annual reports for appointments, dates and financial results, then uses peer-reviewed research and a clearly identified practitioner study for conceptual and comparative evidence. Company statements describe their own rationales and should not be treated as independent evaluations.

The available evidence has important limits. Public-company co-CEO arrangements are uncommon and heterogeneous; titles do not reveal the actual allocation of power. The 87-company return comparison is observational and cannot establish causation. The team-level meta-analysis is not a direct test of co-CEOs. Netflix’s financial results demonstrate that dual leadership can coexist with strong performance, not that it produced that performance. Spotify, Oracle and Comcast have operated under their latest structures for too short a period to support durable conclusions. Interpretations concerning founder asymmetry, organizational clarity and the “one center of gravity” framework are the author’s analysis, not findings attributed to the companies.

Suggested internal links

• The CEO Handoff Is the Leadership Test: Why Succession Must Begin Before the Search — after the discussion of apprenticeship and transition design.
• The Right to Stop: Why Real Leaders Build Systems That Can Overrule Them — in the section on escalation and board sovereignty.
• The Face on the Screen Is No Longer Proof — where the article discusses coherent external authority and verification.
• The Apology Is Not the Repair: What Leaders Must Rebuild After Institutional Failure — in the discussion of accountability for the whole enterprise.

LinkedIn post

Spotify, Oracle and Comcast have all moved to co-CEO structures. Netflix has already shown that two people can share the top office through strategic change and strong financial performance. SAP and Salesforce remind us that the same architecture can also prove temporary.

The useful question is not whether two heads are better than one. It is whether two executives can exercise authority from one organizational center of gravity.

That requires more than trust. Boards must define individual domains, the decisions that require both leaders, a fast resolution process, one external position and an honest account of informal power—especially when a founder remains chair.

My latest Responsible Public Affairs article examines four cases and sets out six clauses for a workable co-CEO mandate. The conclusion is deliberately unsentimental: share the office if the strategy requires it, but never leave sovereignty ambiguous.

#Leadership #CorporateGovernance #SharedLeadership

Newsletter teaser

Can two CEOs create more capacity without creating two companies? Spotify, Netflix, SAP and Salesforce reveal that shared leadership succeeds or fails less through chemistry than through governance. This week’s Leadership essay sets out the six clauses every board should define before dividing the top office.

Professional hashtags: #Leadership #CorporateGovernance #SharedLeadership