Decision by consensus is a common fallback for leaders when dealing with uncertain situations, but it almost always leads to sub-optimal choices. It is usually used by mediocre leaders who lack the competency to make the call by themselves, and the confidence to delegate the decision entirely to a subordinate. Great leaders either decide as a dictator, or trust and empower their reports to make the decisions, as dictators of their own. Of course, this does not mean not listening to feedback from the team and going blindly in one direction despite the rest of the team objecting. However by falling back to the group consensus, we often get the worst outcome: a bad decision while nobody feels empowered, and loss of confidence in the leader.

The fallback to consensus only feels needed in the face of difficult problems. For the simple, straightforward, “no-brainer” decisions, there is no need to resort to the group consensus mechanism. Everyone knows what needs to be done and can just get to work. However, in challenging situations, when the leader feels the most doubt and hesitation, is when the temptation to fallback to group consensus the strongest. Unfortunately this is also precisely the time least suited for group consensus based decision making.

The ancient Romans knew this more than two millennia ago, and had a mechanism in place in times of extreme turmoil to switch to a dictatorship (Julius Caesar being the most famous of such dictators). More generally, at the societal level, decision by mass consensus is arguably one of the biggest weaknesses of democratic systems. Almost all well-functioning democratic countries have tried to skirt around this fatal flaw by one way or another, such as with representative systems. But still, remnants of the problem remain. We put up with it because the alternatives to democracy are much worse in many other areas. But we shouldn’t be fooled into thinking that decision-making by group consensus, especially under stressful situations, is somehow not so bad.

In private organizations, this problem doesn’t have to exist. Companies are already effectively dictatorships, with the CEO holding the absolute authority as the King. Here, a hierarchy of highly competent “dictators”, each having absolute authority over their own work, all the way down to the lowest level worker (who completely own the work that they do), is much better than layers of bureaucrats, each seeking consensus from one level down. Decision by dictators allows the execution of more cohesive visions (even though they are inevitably flawed), concentrates accountability, and separates the competent from the incompetent quickly. Decision by consensus on the other hand tends to scatter the vision, diffuse accountability (one of the reasons why it’s so attractive in times of high uncertainty), and hides the poor performers behind the wall of the group.

Of course, reality is often messy and operates on spectra instead of dichotomies. Under certain circumstances, it may be beneficial to skew more towards a consensus-driven decision making process. As Ben Horowitz discusses in Peacetime CEO/Wartime CEO, companies during peacetime often operate better under a less dictatorial, more harmonious leadership style. They are not struggling for survival, but rather are looking to broaden and expand their existing opportunities. But during wartime, when faced with extreme adversity and high uncertainty, decision by consensus should be avoided whenever possible. Either make the call as a dictator, or fully delegate it to the next level of dictators.