Currently, each code coverage provider (emma, clover, or cobertura) has a preferred code coverage granularity (line, block, statement, class, etc.) that is hard-coded into the provider implementation. This preferred granularity is used throughout the plugin where it is not feasible to take statistics for all of the available granularities into account. For example, when summarizing the build's code coverage into a single percentage or charting coverage deltas by author. In cases like these, taking all of the granularities into account does not align with the purpose of the functionality or would make the information presented too verbose and confusing. It would be nice if the preferred granularity could be configured by the user for each build. This would allow users to focus on the code coverage statistic that they find most useful or critical. In particular, the EMMA provider could benefit from this feature, because line coverage statistics (the original preferred granularity for this provider) are not always available for reasons described in BCOV-23.