How well does Atlassian’s co-CEO model work out in reality?

Co-CEOs are one of those corporate ideas that sounds either brilliantly modern (“collaboration!”) or wildly chaotic (“two captains, one ship, and a lot of emails”). Atlassian made it more than a management experiment. For over two decades, co-founders Mike Cannon-Brookes and Scott Farquhar ran the company togetheran unusually long run for a leadership setup that most boards treat like a temporary bridge, not a permanent address.

So how did Atlassian’s co-CEO model work out in reality? The honest answer is: it worked remarkably welluntil it didn’t need to anymore. And that ending is part of the lesson, not a footnote.

First, what “co-CEO” means (and what people fear it means)

In theory, co-CEOs share the top job with equal authority. In practice, the model can fall into one of three buckets:

• True partnership: Clear roles, shared strategy, one voice to the company.
• Political détente: Two leaders to keep the peace during succession or internal rivalry.
• Slow-motion tug-of-war: Decisions stall because nobody knows whose “yes” counts.

Research and reporting on co-CEO structures tends to land on the same conclusion: it’s rare, it can outperform, and it’s fragile. Studies cited by Harvard Business Review have found co-CEO-led public companies often performed well on shareholder returnswhen the pairing is designed for complementary leadership rather than compromise. That “when” is doing a lot of work.

Atlassian’s co-CEO story: a rare case of “built for two”

Atlassian wasn’t a company that accidentally ended up with co-CEOs. It started that way. Cannon-Brookes and Farquhar co-founded the business in 2002, and for years they operated as a matched set: two founders, two leaders, one missionbuilding tools that help teams work better together.

That origin matters. Most co-CEO arrangements are retrofits: a board wedges two leaders into one role and hopes their calendars align. Atlassian’s model was more like a long-running band: two people who built the music together, toured together, argued about the setlist, and still showed up for soundcheck.

Reality check: the co-CEO era ended

Atlassian’s shared leadership model didn’t crash; it concluded. Farquhar stepped down as co-CEO effective August 31, 2024, with Cannon-Brookes continuing as sole CEO afterward. Atlassian communicated the transition publicly and tied it to long-term planning rather than a sudden rupture. That matters because the ending reveals something important: the model worked for a long stretch, but it wasn’t treated as a sacred artifact. When the company and the founders’ lives changed, the structure changed too.

Why it worked: the five ingredients Atlassian had (that most companies don’t)

1) A “pre-existing trust account” (built over decades)

Co-CEO success requires unusually high trust. Not “we agree on slide fonts” trustmore like “I’ll let you represent the company to the world without hovering behind you like a nervous parent” trust. Atlassian’s founders had years of shared wins, shared losses, and shared context. That makes conflict less existential. You can argue about strategy without secretly drafting a coup memo.

2) Complementary strengths, not duplicated ego

Strong co-CEO pairs usually divide leadership energy in a way that’s legible to the organization: one leans outward (customers, partners, market storytelling), the other leans inward (operations, product execution, systems). Even if the split isn’t formally labeled on an org chart, employees can feel it: “When the topic is X, I know which CEO will drive it.”

In interviews and public commentary over the years, Atlassian’s founders projected that kind of complementaritywithout constantly reminding everyone they were “complementary,” which is the executive equivalent of saying, “We’re not fighting,” in the middle of a family dinner.

3) Clear strategic anchors: product-led growth and long-term bets

Atlassian built a reputation for product-driven growth and a team-centric philosophy. Those aren’t just branding words; they’re decision filters. The more a company has shared filters, the less co-CEOs must negotiate every choice like it’s an international treaty.

Consider a few big, noisy transitions Atlassian navigated during the co-CEO era:

• Cloud migration and platform evolution: Atlassian ended support for many Server products on February 15, 2024, pushing customers toward Cloud or Data Center. That kind of shift is operationally brutal and emotionally unpopular with some customersexactly the sort of thing that can break leadership alignment if the top isn’t locked in.
• Distributed work at scale (“Team Anywhere”): Since 2020, Atlassian has allowed many employees to choose where they work, building a distributed-work operating model instead of treating remote work as a temporary patch.
• Acquisitions that widened the audience: Atlassian’s Trello acquisition (announced in 2017) brought a more visual, broader-market work style into the portfolio. Later, the Loom acquisition (announced in 2023) doubled down on async video communicationanother “teamwork” bet aligned with the company’s narrative.

These are not “set it and forget it” decisions. They require sustained leadership conviction, especially when customers, employees, or investors grumble. A co-CEO model can help hereif the two leaders reinforce the same direction instead of offering the org two competing interpretations of reality.

4) One culture message: collaboration isn’t optional

Atlassian sells collaboration software. That doesn’t automatically make its leadership collaborativebut it raises the cultural stakes. If the co-CEOs were constantly in visible conflict, it would undercut the company’s credibility in a very public way. In that sense, the business model itself nudged the leadership model toward “practice what we sell.”

A useful way to think about this is: Atlassian’s co-CEO structure wasn’t just a governance choice; it was a culture signal. “We build for teams, we lead as a team.” That consistency helps employees tolerate the occasional complexity of dual leadership.

5) Founder-led governance reduces the usual co-CEO landmines

Many co-CEO arrangements fail because of power ambiguity: who really has the board’s confidence, who sets compensation, who is the “heir,” who gets blamed when numbers miss? Founder-led companies can avoid some of that because the legitimacy is baked in. Atlassian’s founders also maintained significant influence through governance structures, whichlove them or hate themcan reduce day-to-day leadership uncertainty.

The trade-offs Atlassian still had to manage

Two voices can confuse the room (unless you work hard to sound like one)

Employees and investors typically want clarity: who decides, who owns the outcome, who is accountable. The co-CEO model can deliver clarity, but it doesn’t do it automatically. You need explicit decision rights, aligned messaging, and a shared communication cadence so the company doesn’t become a “choose your own adventure” novel where every department follows the CEO they last spoke to.

Speed isn’t guaranteedalignment is

The common fear is that co-CEOs slow everything down. The deeper risk is different: co-CEOs can be fast, but only after investing heavily in alignment. When alignment is strong, decisions can move quickly because debate happens early, privately, and thoroughly. When alignment is weak, the organization experiences the worst of both worlds: slow decisions and unclear direction.

Succession is harder than it looks

A long-running co-CEO structure raises a question every board eventually must answer: what happens when one leaves? Do you replace them and keep the model? Do you revert to one CEO? Do you split into CEO/president? Atlassian ultimately chose a clean answermove to a single CEOafter Farquhar stepped down.

That outcome doesn’t mean the model failed. It means the model did its job for a specific era of the company and then evolved. In fact, one of the most “real world” signs of health is that leadership structures change without the company collapsing into a reality TV reunion special.

So… did Atlassian’s co-CEO model “work”?

If your definition of “work” is: create durable growth, stay strategically coherent, and avoid leadership implosion, Atlassian’s co-CEO run is one of the strongest examples in modern tech. The partnership lasted far longer than the typical co-CEO tenure discussed in management research, and the company navigated major product, platform, and workplace shifts during that era.

If your definition of “work” is: remain co-CEOs forever, then noAtlassian ultimately ended the arrangement. But permanence is a strange benchmark for leadership design. The more practical benchmark is whether a model delivers results and can transition cleanly when circumstances change. Atlassian cleared that bar.

Lessons other companies can steal (without stealing the headaches)

1) Don’t use co-CEOs to avoid making a decision

If the board is appointing co-CEOs because it can’t choose one leader, it’s probably choosing a future argument. Atlassian’s model worked in part because it wasn’t a compromiseit was the original design.

2) Define “who owns what” in plain English

Co-CEO success depends on decision rights that are boringly explicit. Not “we both own strategy,” but “you lead product strategy and GTM integration, I lead execution cadence and org design” (or whatever the pair decides). The org should be able to answer: “Which CEO do I go to for this?”

3) Build a single narrative for employees and investors

People can handle two leaders. What they can’t handle is two realities. Atlassian’s public communications consistently framed the company as founder-led with a unified mission, which reduces the risk of internal factions.

4) Treat the model as a tool, not an identity

Atlassian’s shift to a single CEO after 2024 is a reminder that leadership structures should serve the business, not the other way around. If the organization outgrows the modelor life changes for the leadersadapt without drama.

Real-world experiences: what co-CEO leadership feels like (500+ words of practical perspective)

If you’ve never worked in a company with co-CEOs, here’s the lived reality people often describewithout pretending it’s all sunshine or all chaos. Think of it like having two senior editors on a newsroom: when it works, the story gets sharper; when it doesn’t, you get two rounds of conflicting redlines and a deadline that starts sweating.

The “two doors” problem (and how good companies fix it)

The most common employee complaint in dual leadership structures is simple: “Which door do I walk through?” If two CEOs can greenlight a project, teams may shop for the easier yes. That can quietly rot execution because decisions become political instead of strategic. Healthy co-CEO organizations counter this by making ownership visible. Employees can usually tell when it’s working because meetings sound like this: “That’s in CEO A’s lane, but CEO B and I aligned on the principle.” The key word is aligned. Alignment becomes an operating system, not a quarterly goal.

When it’s great: faster decisions after better debate

One underrated upside is decision quality. In strong pairs, the hardest argument happens early, between the co-CEOs, and the rest of the org gets a crisp direction afterward. That feels surprisingly efficient: instead of ten executives debating in public, two leaders do the heavy lifting, then communicate a unified call. This is especially valuable in product companies where trade-offs are constant: customer experience vs. scalability, speed vs. reliability, innovation vs. technical debt. If co-CEOs have complementary instinctsone more visionary, one more execution-focusedthe debate can produce “yes, and” outcomes rather than “either/or” stalemates.

When it’s rough: meetings become a choreography lesson

The downside is that co-CEO alignment is work. Communication overhead isn’t just more meetings; it’s more choreography. Leaders have to decide who speaks first, who closes, who owns follow-ups, and how disagreements are handled in front of the org. In weaker pairings, employees notice subtle cues: one CEO interrupts, one CEO “corrects,” one CEO goes quiet. That’s when uncertainty spreads. People don’t need leaders to agree 100%but they do need to believe disagreements are resolved privately and communicated clearly afterward.

Crisis mode reveals the truth

The real stress test is crisis: an outage, a security incident, a product migration backlash, or a sudden market shift. In a crisis, co-CEOs can be a superpowerone runs external communications and customer reassurance while the other drives internal execution and rapid decision-making. But if trust is shaky, crisis becomes a spotlight on leadership fractures. Companies that handle this well typically have a pre-agreed “crisis protocol” that answers: Who is the spokesperson? Who owns the war room? What decisions require both signatures? What’s the escalation path when minutes matter?

What employees remember after the model changes

When a long-running co-CEO structure ends, employees usually remember two things: whether the transition felt respectful, and whether the strategy stayed consistent. A clean handoff signals maturity. It tells the org, “We’re not built on a personality cult; we’re built on principles and systems.” If the company continues shipping product, supporting customers, and communicating direction, the leadership change becomes a chapter, not a cliff. Atlassian’s arc illustrates that co-CEO leadership can be a durable advantage when it’s founder-built, role-clear, and culturally reinforcedand it can still be wise to evolve the model when the company enters a new season.

Bottom line

Atlassian’s co-CEO model is one of the rare examples where the arrangement wasn’t a temporary patchit was a long-term operating design. It likely worked because the founders were aligned on mission, complementary in style, and disciplined about presenting one direction to the organization. The fact that the company later transitioned to a single CEO doesn’t weaken the case; it strengthens it. The real world isn’t a leadership textbook. The real win is building a model that performsand can evolve without breaking the business.