RIP Chris Sims.
I was enormously influenced by Chris. My own, perhaps idiosyncratic take:
His main contribution came at a time when macroeconomists had constructed the first wave of big macroeconometric models. They were constructed piece by piece, a consumption block, an investment block, and so on. Each piece looked reasonable, but when assembled together, the implied macro dynamics were all wrong.
What Chris did was to turn things around, namely argue that one had to start from the actual macro dynamics, the so called VARs, and show how, with minimal identification conditions (leading to "structural" VARs), they could be used to suggest the dynamic effects of various shocks, dynamics that the structural models had to replicate.
To say that his approach was influential would be to understate its influence. Today, a model that did not fit the VAR evidence, would be simply dismissed.