Attributor announced that HarperCollins and Harvard Business Publishing had signed up as clients of their “FairShare Guardian” copyright protection service. I admit, once again, that my initial reaction was not good. Spending tens of thousands, perhaps hundreds of thousands of dollars a year on piracy enforcement might make sense, if piracy cost publishers paid sales that exceeded the cost of the service. But … for most markets, there is not enough data to either confirm or deny that claim. When Attributor announced its estimates of piracy’s cost in January, I observed that the report tracked the instance of piracy, not its impact. My blog post suggested sharing data first (rather than rushing to put an estimate of losses out there) would have been more helpful. Around that time, someone posting on the @attributor Twitter account offered to work with Magellan (this was in the public feed, not a direct conversation, so I am comfortable sharing it). I said “sure”. That was the last time I heard from @attributor. Reflecting on the most recent announcement, I’ve come to view Attributor as the new IBM (as in, “Nobody ever got fired for buying IBM”). Buying a monitoring service is comforting, and it lets publishers tell authors and agents they are doing something. Attributor’s impact on paid sales is about as murky as the various estimates of losses in book sales. Standing in place – making the safe decision to defend what you have – is natural, but not typically prudent. As Wil Johnson points out, the tools for peer-to-peer file sharing are evolving. Publishers who sign with Attributor are fighting the last war, without the data in hand to understand the impact of piracy on paid content sales. By focusing on enforcement, publishers also risk crowding out weak signals that may point to new, untapped markets. Paolo Coelho has used piracy as a leading indicator of interest, to the benefit of his work and his bank account. For publishers, engagement could be the more effective response to the instance of piracy. My offer to work with Attributor stands, though. More shared data is a good thing in this area.