Suggestion for the accumulated limits on duplicate files, would it be possible to only apply that to the first record found for that user? On our site, multiple files is almost a guarantee (with a good number of customers "forgetting" they own a certain product and re-purchasing it later by mistake).. so needless to say, going through thousands of customers isn't exactly a simple task, although in theory there could be a query written that would do user-by-user comparisons. Possibly break it up into x number of compares per cron job or something.

Just a thought.