Does having two startup co-CEOs work?

One of the key lessons I learned when I was CEO was to hire more experienced people, sooner, and delegate key decisions to their expertise. However, there are many things which need to be decided quickly or with the authority of “the boss”. Whether these are high level strategy, having someone to take ultimate responsibility or just to come to a final decision, this seems like something which can only be decided by a single person.

That means that I’ve always found it odd when I hear about a company having two co-CEOs. Startups are not democracies and someone needs to be responsible for everything that happens.

In the early days of a startup, there are many decisions. Questions about culture, who to hire, whether to seek investment and how much, how communication should work and even the original product vision.

In some contexts, consensus decision-making can work, as can deliberation or committee, but as we know from government it is also notoriously slow and laborious. It might produce a politically optimum outcome for a broad constituency of national views, but is that the right basis for startup decision-making?

When I’m leading through a tough decision, I try to say, at the outset, “I want all of your opinions, but I’m going to be the one who ultimately makes the decision.” Or in some cases, I will say, “I don’t know if I’m the right decision-maker. I need help exploring what the decision vectors are, and I need all of your help. And then I will let you know how we’re going to make the decision once we’ve talked about it.” If you don’t give people that guidance, which is I think a common mistake, you’re likely to run into trouble.

High Growth Handbook, Elad Gil

Which successful companies have co-CEOs? #

Very few high growth startups have two CEOs. It often starts out as a way to make the early founders feel better rather than decide on one leader, but to me it just sounds like indecision. Roles need to be divided but the key decisions need to be made by one person.

But there are counter-examples. Oracle, SAP and Atlassian currently have co-CEOs, with Atlassian having this structure from the beginning. Despite this, it seems that investors are skeptical of such arrangements. What happens if the two people don’t agree? Who do investors work with? How does it work with board voting arrangements? Who do the key executives actually report to?

Founding a company is difficult but that is why the most successful companies have multiple co-founders. They support each other and provide a forum for difficult discussions. Having two people with equal power and authority makes things difficult for everyone.

Even in the Roman Republic (509 to 27 BC), which had two elected consuls serving together as the highest level of office. They shared “imperium” over the whole of Rome and its provinces, rotating on a monthly basis. This avoided conflicts because only one consul was actually governing each month, but the other consul still had a veto power over the other. A great example of practicality and safeguards (until the end of the Republic, of course).

Academic literature on two co-CEOs #

A 2011 study – It Takes Two: The Incidence and Effectiveness of Co‐CEOs. Financial Review, 46 – looked at 111 US public companies with two co-CEOs. It found the following interesting results from the data:

• Most co-CEO arrangements happen as a result of M&A. The buyer and seller CEOs combine forces in the new company.
• Having two CEOs has advantages in allowing the combination of complementary skills in a role that often has to oversee a wide range of areas, particularly if there are multiple locations or timezones that require senior leadership representation. However, I’m unsure why this couldn’t be covered by a senior executive team and video conferencing. Leaders are not expected to be experts in everything they run, they’re expecting to be experts in running teams of experts.
• Several sources were cited (Mintzberg (1989), Hackman (2002) and Alvarez and Svejenova (2005)) which suggest that coordination problems and interpersonal conflicts are common in co-CEO arrangements due to strong egos. Exploring this was one of the goals of the study and they actually found that co-CEO arrangements last for about 4.5 years, the same as single CEOs. This suggests that such arrangements are just as stable.
• Co-CEOs receive significantly less incentive compensation than individual CEOs.
• Firm value is affected positively. Announcements of co-CEO appointments are usually met with a positive market reaction and the subsequent market to book (M/B) value of the company increases over time with co-CEOs.
• Co-CEO arrangements resulted in lower effort from CEOs and suggested that CEOs preferred to be solitary rather than work together (Stein (1988), Aghion and Tirole (1997), and Almazan and Suarez (2003)).

However, the study specifically says that this is not an argument for appointing co-CEOs. There is correlation but not necessarily causation:

our findings regarding corporate value do not imply that more firms should adopt a shared leadership structure. We base our valuation analysis on propensity score matched samples and therefore compare the firm’s M/B with co‐CEO leadership to the unobserved M/B that would have occurred with a solitary CEO. Improvements in M/Bs only show that firms with shared leadership improved their valuation. Our results do not imply that any firm could appoint co‐CEOs and enjoy higher value.

Arena, M. P., Ferris, S. P. and Unlu, E. (2011), It Takes Two: The Incidence and Effectiveness of Co‐CEOs. Financial Review, 46: 385-412. doi:10.1111/j.1540-6288.2011.00305.x

Conclusions #

Only 15% of the companies in the 2011 survey had founders as as co-CEOs. This is the same trend we see with the tech company examples above – Oracle and SAP which supports the notion that older, established public companies might run fine with co-CEOs but does not say anything specific about startups.

Atlassian seems to be the only high growth, tech company that has successfully started, grown and gone public with a co-CEO model, and it is not the only thing it has done differently. It is therefore an outlier and its model should not be copied.

Public companies are (and should be) run quite differently from new startups, usually because they have hired CEOs that approach things very differently from founder CEOs. My experience suggests that startups need a small, core (mostly technical) team that have clearly defined roles and are working together towards a single goal. As the company grows, decision making is delegated based on outcomes to an experienced team led by a single CEO.

Having two co-CEOs for a startup does not work.