Great idea. Reports are definitely not research units, and tagging individual “wisdom nuggets” will create a library of insights. A decade ago, my then IBM colleague Paul Englefield and I developed a “usability error classification” that had a similar tagging approach with a somewhat different purposes. Tagging research findings by the impact of the error as most likely place in the design process where it originated, we were able to compare research ‘apples and oranges’ across methods within one study as well as across studies. Tagging is a great mechanism to bring together research findings from multiple studies. Good that your team does it now!