Most of the time the processor in your computer is doing very little work. In fact, as you read this, it is probably sitting close to idle. The work demanded of it is often tiny compared to what it is truly capable of when it runs full tilt. Modern processors have base clocks and turbo modes to reflect these dual demands.

When work is light a processor sits at its base clock trying to be as energy efficient as possible. When work is hard it runs at its turbo frequencies for as long as it can manage until the work is complete. This often results in a very spiky existence for the processor as it oscillates between dreadfully long periods of doing nearly nothing and often relatively brief (but sometimes continuous) bursts of all out demand for optimal performance.

On the other hand, if a processor has insufficient cooling or suffers from other environmental factors (like being directly in sunlight for a long time), it will begin to thermal throttle. This is a feature designed to prevent the chip from harming itself, and potentially other valuable components, if it is going to exceed its thermal capacity for too long.

This means that the processor eventually finds equilibrium at maximum sustained performance given its thermal realities. In a perfect world (with good cooling) it can and will run at full tilt forever as it sheds heat into an environment capable of dissipating it effectively, but in the worst case it will shut down all together (such as when the heatsink is removed) which stops any work from being done.

Technology often mirrors life since it is made by humans for humans, and these concepts translate directly into how people operate. In knowledge work the demand is often spiky like that of a processor. We may be chugging along performing our work as normal before a new initiative kicks off and soaks up all of the available bandwidth from a team. Depending on the circumstances, this may cause people to go from from their normal work output into turbo mode. If teams or individuals run in turbo mode for too long, they burn out unless throttling is introduced.

This scenario is actually an example of optimal stopping theory. If a team or individual operates at 100% of their ability at all times then there is no scenario in which they can respond to a spike in demand or change of circumstances without dropping something else they were doing which was already important. However, you also don’t want anyone sitting around idle.

How much capacity should generally be utilized? When do you introduce throttling?

This depends on the nature of the work and the environment the team or individual finds themselves in. If the work is well understood and consistent then maybe you can survive closer to the 100% threshold, but the less clear and less consistent the work the more you need to build in some reserve capacity. This idea was somewhat infamously explored by Donal Shoup‘s book The High Cost of Free Parking. He proposed the idea that parking utilization should be around 85% to be optimal, and that introducing parking fees to achieve this this would reduce traffic congestion, fuel waste, and time waste (thus saving more than the fees paid) while raising revenue for the city by removing free parking. For the cities that have tried it, it appears to work.

This seems close enough to the 80/20 rule (known as the Pareto principle) to help us here. We should strive to consistently operate around 80-85% of our possible capacity. This leaves us room to turbo when necessary, but also means we are never idle and not providing value. When I say these numbers I don’t mean just at work, but your total capacity as a human being. When one of my team members had a sudden loss in the family that obviously demanded many of his cycles, and that meant he had less cycles to use elsewhere for a time. Try to quantify the demands life has of you and be realistic about what that means for where you spend your cycles.

If an individual or team is consistently above this threshold then they (and leadership) need to consider two things:

1. Throttle before you burn out. This will feel impossible when the workload is at its peak with no end in sight, but it simply must be done for the good of all. Take some PTO, find ways to relax and de-stress, and see if any of the work can be better balanced among teams and individuals.
   
2. Increase capacity and/or shed workload. For a team this is obvious: if there is important work not getting done then see if you can delegate this (or less important) work or hire more people. For individuals this is a little harder. Are there improvements to the workflow or process you could make? Would training help you be more efficient? Is there other low value work eating up your cycles that you could stop doing? Can you partition the tasks differently such that more people can contribute?

America treats working yourself to the brink as a rite of passage, but as a recovering workaholic who has gone beyond all limits several times I assure you it is sub-optimal just like free parking and infinite turbo. When you run out of juice and force a shutdown you halt everything, and the damage of this scenario can be extreme.

You do not provide value to your team, your organization, your family, or yourself if your burn out.

To be a consistent high performer you must learn and respect your actual capacity. To be a good leader you must stop individuals and teams from running too hot for too long while also making sure no one is idle.

If you are turboing right now without an end in sight, consider when to start throttling. Taking care of yourself is not defeat. On the contrary, it is the most optimal thing you can do.