Play your position, you way out of character
--Nas, You Know My Style
In Jim Collins' best selling management book Good to Great, he demonstrates through massive research and comprehensive analysis that when it comes to CEO succession, internal candidates dramatically outperform external candidates. The core reason is knowledge. As I discussed in Why We Prefer Founding CEOs, knowledge of technology, prior decisions, culture, personnel, et al tend to be far more difficult to acquire than the skills required to manage a larger organization. Collins does not, however, comprehensively explain why internal candidates fail. I will attempt to do so here.
In my post How Andreessen Horowitz Evaluates CEOs, I explored two core skills for running an organization: 1. Knowing what to do and 2. Getting the company to do what you know. While being a great CEO requires both skills, most CEOs tend to be more comfortable with one or the other. I call managers who are happier setting the direction of the company Ones and those who more enjoy making the company perform at the highest level Twos.
What Ones like and don't like
Ones like spending most of their time gathering information from a broad variety of sources from employees to customers to competitors. Ones love making decisions. Although, they prefer to have comprehensive information when they make a decision, they comfortably make decisions with very little information when necessary. Ones have great strategic minds and enjoy nothing more than a good game of 8 dimensional chess against their best competitors.
Ones sometimes get bored with many of the important execution details required to run a company such as process design, goal setting, structured accountability, training, and performance management.
Most founding CEOs tend to be Ones. When founding CEOs fail, a significant reason why is they never invest the time to be competent enough in the Two tasks to direct those activities effectively. The resulting companies become too chaotic to reach their full potential and the CEO ends up being replaced.
What Twos like and don't like
Twos, on the other hand, thoroughly enjoy the process of making the company run well. They insist upon super clear goals and strongly prefer not to change goals or direction unless absolutely necessary.
Twos like to participate in strategic discussions, but often have difficulty with the strategic thinking process itself. Where a One might be perfectly comfortable spending one day a week reading, studying and thinking, doing so would make a Two very nervous, because it would not feel like work to a Two. A Two would get antsy at the thought of all the processes that might be improved, people that might be held accountable to achieving the standard, or sales calls that could be made while he was wasting time just thinking about strategy.
Big decisions worry Twos much more than they worry Ones. Circumstances often force both Ones and Twos to make critical decisions with insufficient data, but Ones generally feel fine about doing that and do not get overly anxious about the consequences. Twos, by contrast, can become highly agitated about such things and sometimes over complicate the decision-making process in order to provide a false feeling of thoroughness about the choice.
CEOs that are Twos, despite their love of action, can sometimes slow decision making in a company to a halt.
You need both characteristics to be a good CEO
While people tend to be Ones or Twos, with discipline and hard work natural Twos can be competent at One tasks and Ones can be competent at Two tasks. If a CEO ignores the dimension of management she doesn't like, she generally fails. Ones end up in chaos and Twos fail to pivot when necessary.
Often Two executives act as Ones for their functions, but Twos as members of the executive team. For example, the head of sales might easily make all the decisions that are local to the sales organization, but prefer to take direction with respect to the overall company plans. This is the best kind of multi-layer leadership possible, because directions are clear and decisions are made rapidly with precision.
How organizations tend to be constructed
The primary purpose of the organizational hierarchy in a company is decision-making efficiency. It follows then that most CEOs tend to be Ones. If the person at the top of the decision-making hierarchy doesn't like making extremely complex decisions, the company's processes will be slow and unwieldy.
If you're a One, it can be counter productive to have another One on your staff, because she will want to set her own direction rather than follow yours. This kind of strategic contention can confuse the organization and send employees in opposing directions. As a result, many great One CEOs employ primarily Twos and Functional Ones on their staff.
What happens at succession?
This brings us to the question of succession. Since most organizations are run by Ones and have a team of Twos (sometimes Functional Ones) reporting to them, replacing the CEO can be extremely tricky. Do you promote someone from the executive staff even though they are likely a Two? Microsoft did this in 2000 when they replaced Bill Gates, a prototypical One, with Steve Ballmer, literally his number two. Or do you reach deep into the organization and pull a One from a level lower where they are likely to exist? GE famously did this with Jack Welch in 1981. It was an incredibly bold move by GE--not only did they promote an executive two levels down in the organizational chart past all of his superiors, but in doing so they named the youngest CEO in the history of GE. It's difficult for most board members to even conceive of the possibility that there is a One deep in the organization that is more qualified to run the company than anyone on the executive staff.
Both methods can be problematic. The first approach leaves the company in the charge of a Two. As the company faces forks in the road, decision-making may slow down and the company may lose its edge. In addition, the natural Ones (in Microsoft's case, stellar executives such as Paul Maritz and Brad Silverberg) will eventually leave.
In scenario two, by promoting someone past everyone on the executive team and making them CEO (as GE did), you will likely cause massive turnover of the executive staff. In fact, in very short order, almost none of the original GE executives remained under Welch. In a diversified conglomerate like GE, this kind of rough transition is possible. For companies in the highly dynamic technology business, the super high-turnover scenario is more dangerous.
The big conclusion
The big conclusion will be a big disappointment for those looking for an answer. The answer is there is no easy answer. CEO transition is hard. If you bring people in from the outside, you lower your chances for success. If you promote from within, you must deal with the One/Two phenomenon. Ideally, you'll promote a One and the rest of the executive team will be glad you did. Too bad things are rarely ideal.
Reprinted from Ben's Blog
Ben Horowitz is the cofounder and General Partner (along with Marc Andreessen) of the venture capital firm Andreessen Horowitz based in Menlo Park. Follow him at @bhorowitz, or get updates on his venture capital firm @a16z.