> Coverage as a whole I think is useful only if you have an engineering organization that doesn't see the point in tests - which is going to be its own uphill b

> Coverage as a whole I think is useful only if you have an engineering organization that doesn't see the point in tests - which is going to be its own uphill battle. I think coverage stats are always useful as they help find the edge cases that people forgot to test. A common culprit I've seen is error handling code where a bunch of tests target the happy path, but nothing tests the error logging when something breaks.