We worked on aggregating the analysis results into a more useful format for benchmarking purposes. SIG aggregates results by calculating the relative system code weight for each metric value (e.g. sum of LOC of methods with complexity = 5 / total system lines of code), so we decided to do the same. We can now use this data to generate nice histograms that show the relative weight in each metric value. (see attachment, note that clicking on the bars lists all methods with that cyclomatic complexity).
An example benchmarking report. 