This actually seems like a difficult number to get exactly right. It could easily be off by thousands. How many had not made 30 days yet, had already quit, had already been fired, hadn't been cleared in the union's system or UPS's system, weren't really up to date on dues, or had other issues? An inflated number of eligible voters seems almost certain, and that could make a big difference in the turnout percentage. We weren't all that far from 50%. How can we audit whether the union counted the turnout correctly? It's very, very easy to manipulate it.