Assume you have total legal spending as a percentage of revenue for hundreds of legal departments.
Assume further that you gather demographic and career data about the general counsel of those departments. Directories such as the Aspen Directory of Corporate Counsel will provide quite a bit of this background, supplemented by Martindale-Hubble, LinkedIn, and company websites.
With some digging, it would be possible to know the current general counsel’s age, whether they were promoted from within, formerly a partner at a law firm, or a general counsel previously as well as such data as their tenure as general counsel and in the law department and (perhaps) their gender.
In this natural experiment, it would be easy to calculate correlations between those personal facts and key benchmarks such as total legal spending as a percentage of revenue.
Not that correlation proves causation. But I would not be surprised to find that particular metric declines after the first year or two of a general counsel being in the position but might rise after a decade. The pause before a decline means the general counsel has to invest some time to learn the company and shape the department. The increase after a decade implies some rigidity in thinking, some givens that are not examined.