I am one of the devs in the pictures. We often have measurable improvements -- the shark hat (meant for performance improvements) is obvious. Results are shown in our performance monitoring tools.
The impact of our rainbow-hat (meant to enhance the codebase, contribute to some open-source projects we use, do what you think makes the world a better place) is harder to put on a graph. But we could count code metrics, documentation, merged features in OOS projects etc. We don't do that , though.
What we do is talking about what we did with our hat time after each sprint when choosing the new hat-wearers.