E.G., measure the productivity of repo owners (managers/leads) by their lag in accepting pull requests. Measure the quality of submitted code by comparing how often my code is accepted compared to someone else's (w/in the organization/language/community).
From there you could build more of a productivity measurement suite, e.g. maybe connect test coverage and bug reports to specific commits and start to score a coder.
Basically, look at currently-existing data points about coders that a specific team culture believes are important to productivity. Then baseline that against a comparable group, and present reports.
Another trick is to simply measure stuff over time. (This is a bit trickier, due to what you hinted at w/ LOC... but still measures change over time, which is really useful for managers.)
The final thought I have is that it is useful to connect events over time that are related to the same code/feature/branch. If my team's workflow is design/develop/test/commit....gap....deploy to users, then bugs are temporally disconnected from development efforts. Connect things over time in an easy to use report, that has value.