So that won't scale beyond the individual, which might be acceptable from a Stoic perspective but I doubt that's the solution you're looking for.
Oren Ellenborgen shared my post in today's edition of Software Lead Weekly [1], and in his summary of the post he suggests using Key Failure Indicators (KFIs) to counterbalance Key Performance Indicators (KPIs). From what I understand, the idea is that you observe metrics of success but you also observe metrics of health, and take action if the former starts degrading the latter.
For example, if you want to deliver features more quickly, you might measure the time between a change is submitted for review and the time it's deployed. But if you're concerned that this could degrade quality, you could also measure the number of defects that are identified. If defects increase as delivery time decreases, this indicates the success metric is causing some undesirable behavior. Hope this helps.