I'm back to a condition where I can't consider methodology (I'm on a bus heading to a glacier) but my reply was to your comment about how laying out the data this way prevents all in cohort comparison. It is still trivial, you just pivot on the label. If that's not what you went, sorry.
