People focus on coverage but I think that has diminishing returns. There is a lot of 'art' to building clean reusable code which you can't easily standardize b

People focus on coverage but I think that has diminishing returns. There is a lot of 'art' to building clean reusable code which you can't easily standardize but what you can do is pay attention to things like: * Build times * Testing times * Test Flakiness * High quality fast unit and integration tests * High quality fakes * Fast, high fidelity non-prod environments * Time from commit to production, number of cherry picks/rollbacks * Good Coverage, but focus where it matters. If any of these go out of whack it needs to be a P0-P1 to fix because it really is debt that grows to become unmanageable (not to mention repelling to the best people). In most companies, even big tech, this is difficult because non-coders set priorities and allocate the resources. Even once-upon-a-time engineers who make it up the org chart find themselves preoccupied with other things and invariably sacrifice these things for features, compliance, conformity, toolchain complexity/sophistication, security, outage/incident scar tissue, etc. (And maybe rightly so as there are few successful businesses that manage this consistently)