There's this popular assumption that someone's performance is just equal to one’s outcome. Especially this is relevant for higher rank managers. I think this is demoralizing.
I really like that in the finance world when rewarding portfolio managers, not only outcome is important but how much risk was taken to achieve it. There are two ways of adding value as a manager there. You can have a higher return at a given risk or reduce risk at a given return.
Maybe we can and should follow their example when assessing the performance of somebody.
I think assessing performance just by looking at the output can be a sign of being a lazy supervisor. Of course, digging deeper is a lot of work. You'd also have to understand the work of your subordinates. Which by the way is one of the most important indicators of having happy employees.
But I'd argue it is worth the effort. Especially when we evaluate short-term performance, luck and other external forces can have a major impact. Especially, as in the financial world where the market is not perfectly efficient, the IT world is far from efficient as well.
If person A has achieved outcome X, and person B's achieved 90% of X, but has taken half the risk of the person A, I'd prefer to work with the person B.
Measuring risk isn't easy.
But we can assess whether an IT manager reduces uncertainty as much as possible until it stops being cost-effective to do so.
Does he/she use Agile methodologies? What about Extreme Programming, Data Driven decisions, DevOps or even Resilience Engineering?
We have plenty of tools to reduce and manage risk(uncertainty) in our IT toolbox.
Evaluate your people in a way that ensures sustainable long-term success.