The real problem here isn't Excel. It's the assumption that a spreadsheet is a stable grid, when the moment your data starts moving, that grid is a story that keeps rewriting itself. The user's frustration is legitimate: they're doing the right things, learning Power Query, testing formulas, trying to build a repeatable comparison. But the tool keeps punishing them because they're comparing rows when they should be comparing identities.
What they need is a way to anchor each engineering report to something that doesn't shift when a row above it disappears. That means using a unique identifier, a report number, a document code, something that travels with the data instead of relying on its position. Once you have that, the comparison becomes straightforward: match on the identifier, then compare the values that actually matter. The IF function and conditional formatting they tried aren't wrong in principle, they're just too fragile when the row order is part of the data. And Power Query flags entire rows because it's comparing by position by default, not by meaning. That's not a failure of effort, it's a mismatch between the tool's default behavior and the reality of shifting data.
This is where the opportunity lies, not just for this user but for anyone who's hit this wall. The fix is to stop thinking in terms of row-level changes and start thinking in terms of record-level changes. That shift feels small, but it changes everything. You're no longer asking "what's different in this row?" You're asking "what changed for this specific report?" That's a more precise question, and it leads to a more useful answer. The team gets a clean log of what actually changed, not a list of false positives that buries the real updates under noise.
The deeper takeaway is that spreadsheet tools are evolving, but the thinking has to evolve first. The user is already on the right path by learning Power Query. The next step is learning to merge on a key, not on a row number. That's not a technical hurdle, it's a conceptual one. Once you see the difference, the solution becomes obvious: give each report a stable identity, then compare on that. The rows can move, the sheet can grow and shrink, and your comparison still works. That's the kind of practical insight that turns a frustrating task into a manageable one. And it's available right now, in the tools they already have.