One of the relatively few responsibilities that almost exclusively falls on my shoulders is allocating scarce resources. We have direction defined by our mission. We have a thesis on how to achieve our goals as spelled out by our strategy. But none of that dictates how to balance the short term against the long, risk against reward, or urgent against important.

People enjoy the simplicity of stack ranked priorities. Decisions are easy when every conflict can be solved by simply looking at a list. Unfortunately, such a priority list rarely withstands contact with the real world. Just because something is a priority doesn’t mean you invest in it infinitely. Just because breathing is the most important thing to my survival doesn’t mean I starve.

The goal is to maintain a balanced portfolio of investments. Even with a strong sense of priorities we have tough calls to make. Some work is higher risk than others. Some work will yield on longer timelines than others. We must have an idea in our heads of how we want to balance all those aspects. Done well, a portfolio allows us to protect critical long-term investment without neglecting work that is immediately valuable.

For example, Google is said to aspire to a ratio of 70% core work, 20% related work, and 10% new work. When I worked on ads my goal was for 60% of our work yield within a year, 20% within two years, and 20% further out. Moving to Facebook Reality Labs demanded a very different portfolio mix with more risk and longer timelines.

I start by thinking through what the mix of work, risk, and yield should be in the abstract. You can imagine a large spreadsheet with the number of rows and columns in each section determined by how much I have decided to allocate. Then I take my priorities and start to fill them in without allowing myself to overrun the allotted space. As with portfolios in investing, uncorrelated risk is preferable and periodic rebalancing is important. And of course I’m free to shift the allocation if I see a particularly attractive opportunity.

Portfolios are scale invariant. So within Facebook, Facebook Reality Labs likely represents a high risk long term bet, but within FRL we have our own frame of reference and our own portfolio allocation. This can help contain urgent work from overtaking important work because it can’t grow beyond its investment allocation.