The user who posted this frustration has done everything right. Custom sort lists, column sorting in the Data Model, worksheet protection, they've tried the standard playbook. And still, the moment a slicer removes a column with no data, the entire heading order collapses. This isn't a user error. It's a limitation of the tool itself.
Pivot tables were designed for static, predictable data sets. They assume your columns will always be present and in the order you defined. But real-world data doesn't work that way. Week bands shift. Categories drop out. Slicers are supposed to give you control, yet they actively break the structure you worked to build. The user's instinct to "lock" columns in place is exactly right, and it's a feature that traditional spreadsheets simply cannot deliver reliably. The workarounds are exhausting, and they all fail under the same conditions.
What this user needs is a fundamentally different approach: a data model that remembers column order as a property of the data, not a fragile manual setting. In an AI-native spreadsheet, column headings aren't just labels you paste in, they're semantic markers that the system understands. When a slicer removes a week band, the column doesn't vanish and reorder the rest. It stays in place, empty but present, because the structure is defined at the model level, not the display level. This isn't about adding another menu option or a checkbox. It's about rethinking who owns the layout: the user, not the filtering engine.
The practical takeaway is simple. Stop fighting a tool that treats column order as an afterthought. Look for solutions that let you define your table structure once, then trust it to persist through any filter or slicer. The user's request to fix sorting before locking columns is understandable, but the real fix addresses both at once. When your data model respects your intended order as a rule rather than a suggestion, sorting becomes automatic and locking becomes unnecessary. That's the shift worth exploring.