Two reports, one number, and a $410K gap
Finance and operations reported different revenue for the same month, every month. Nobody could say which was wrong.
Unravelling the spreadsheet nobody wants to touch: where each number comes from, which formulas are quietly wrong, and why it doesn't match the other report.
You have a spreadsheet or report that people rely on and nobody fully understands. It might have been built by someone who has left. It might disagree with another report that is supposed to show the same thing. Either way, people have stopped trusting it, or they trust it more than they should.
I trace every number back to where it comes from: which system, which extract, which filter, which formula. Along the way I note anything that is wrong, anything that works by accident, and anything that only one person knows how to do.
A plain-language map of how the report is built, and a short list of problems ranked by how much they matter. If two reports disagree, you get a reconciliation that shows exactly where and why they split.
Nothing is changed in your systems unless you ask. Often the review is enough on its own. Sometimes it leads to automating the report or fixing the feed behind it.