I feel like manual tests are usually more like integration tests than unit tests. This makes it a bit easier to estimate how much of the functionality you're covering, rather than how much of the code.
Which seems like a more useful metric. It's just hard to estimate it for unit tests.
You'll probably not get near 100% coverage manually, it's better to try and test the most critical parts rather than the parts with most code.
Mark
I feel like manual tests are usually more like integration tests than unit tests. This makes it a bit easier to estimate how much of the functionality you're covering, rather than how much of the code.
Which seems like a more useful metric. It's just hard to estimate it for unit tests.
You'll probably not get near 100% coverage manually, it's better to try and test the most critical parts rather than the parts with most code.