Many of my posts identify some cognitive bias and offer advice on how best to avoid it. In this case I identify one that we can get to work in our favor.

Selection Bias is the property where a group of people fails to be representative of a broader population in some way. If you offer free haircuts the population that shows up will have a lower proportion of bald people. If you have better medical care you will identify diseases at a higher rate but that doesn’t mean they occur in greater frequency. If you are doing research or analysis and hope to generalize your results to a broader population it is important to avoid this kind of bias.

But selection bias can also be a good thing.

We often say at Meta that awesome things are built by the people who show up to build them. Sometimes what you need is an army of the willing marching towards a single goal. Infighting is far more damaging to progress than almost any technical or competitive obstacle. I value productive tension but sometimes the time it takes to address the part of the team pulling in the wrong direction is just not worth the value you gain from their additional perspective. When an organization is under pressure it is often more valuable to move in a uniform direction with full force than it is to spend time triangulating the exact right vector.

This idea must be applied very carefully, though. If you ask your existing population to refer candidates to hire there is a likelihood you will not acquire more diverse viewpoints. If you enforce too strong an ideological hegemony you’ll lose your highly valuable malcontents. When the organization is running more smoothly and has more degrees of freedom it benefits to invest more in exploration by inviting more dissent.