My immediate thought - create an elasticity model of how changes in surrogate cause changes in north star metric by doing causal analysis

How do others think about this?

Linkedin paper: https://arxiv.org/ftp/arxiv/papers/2106/2106.01421.pdf

Athey/Chetty/Imbens/Kang paper: https://arxiv.org/abs/1603.09326v2