I'd add that even those years of retrospective analysis will be insufficient, because "hiring successfully" is fundamentally about measuring membership in a particular social group. You can never pick the right observable dimensions to measure in your analysis, it will ALWAYS be lacking. Sure, you can measure work done, tenure, etc, etc, but that does not mean a hire was successful.

You might have 2 people with opposite personality traits ace the interview then get put on a team together, now all of a sudden your whole team is worse overall. That's a massive effect but your study will just see it as noise.

The reality is your best hires will just be intelligent people who belong in the group and who push the mission forward.