There's a smarter way to handle this, and the frustration in that Reddit post proves exactly why traditional conditional formatting has been holding users back for years. FearlessResearcher48 is trying to do something perfectly reasonable: color each cell based on its own value against a fixed scale of 0% to 100%, not against the other cells in the selection. The tool they're using simply cannot deliver that logic without a workaround that feels like a puzzle.
The problem is baked into the design of legacy spreadsheet software. When you apply a gradient color scale to a range, it normalizes the values within that range. So the highest number in your selection always becomes green, and the lowest always becomes red, regardless of whether that "low" is actually 78%. That's useful for comparing cells to each other, but useless when you want each cell to stand on its own against an absolute benchmark. And formatting each cell individually? That forces the tool to treat each one as its own one-cell universe, so every value, even 78%, becomes the maximum and turns green. Neither approach gives you a simple rule like "if this cell is below 40%, make it red; if it's above 80%, make it green."
What this user really needs is a conditional formatting engine that understands independent thresholds, not relative comparisons. An AI-native spreadsheet can do this naturally because it treats each cell's logic as a separate condition, not as a group statistic. You define the rule once: red for values under 40%, yellow for 40% to 80%, green for above 80%. Then you apply it to all four cells, and each one evaluates itself independently against those fixed breakpoints. No workarounds, no frustration, no guessing why a 78% turned red.
This isn't about replacing spreadsheets. It's about making them work the way you already think. The user's instinct to set a clear, absolute scale is exactly right. The tool just needs to catch up. So here's the concrete point: if you find yourself fighting conditional formatting to do something that feels obvious, the tool is the problem, not your logic. Explore a solution that lets you define rules by meaning instead of by range.