Measure Killer Measure Killer

These docs are still being polished — a few sections and screenshots are on the way. Spotted something off? Let us know.

Clean your Power BI model — advanced optimization

Go beyond removing unused artifacts. The Clean your model tab surfaces unnecessary items that are still technically referenced but serve no practical purpose — hidden visuals, dead pages, auto-detected relationships, and more.

Last updated · May 25, 2026

Unused vs. unnecessary

Measure Killer’s core analysis classifies artifacts as unused when they have zero references. But many artifacts are technically referenced yet serve no practical purpose:

  • A column that only appears in a hidden visual nobody sees
  • A measure that only exists in a page with zero views
  • An auto-detected relationship that Power BI created on its own

The Clean your model tab targets these cases — artifacts that maintain technical references despite having no practical utility.

How it works

After running a scan, open the Clean your model tab. It presents a curated list of improvement suggestions, each addressing a specific cleanup scenario. For most suggestions, the tool estimates the potential model size reduction so the biggest wins surface first.

Most suggestions only appear when the artifact is used in that one single situation, and nowhere else. This prevents false positives when artifacts serve multiple purposes.

Only suggestion 1 (Remove unused artifacts) can be applied automatically by Measure Killer — via Clean TMDL export or 1-click cleanup. All other suggestions are advisory: Measure Killer identifies the issue and the affected artifacts, but the fix is applied by you in Power BI Desktop, the Service, or any tool of your choice. You can export the full details of every suggestion to work through them at your own pace.

Measure Killer Clean your model tab — numbered cleanup suggestions with per-item size reduction

All suggestions

1 — Remove unused artifacts

The foundational suggestion — the same job covered in depth by Find and remove unused measures and columns. Targets artifacts (tables, columns, measures, relationships) lacking any references or depending solely on other unused artifacts. Export clean TMDL is the fastest option — apply it via the TMDL view in Power BI Desktop (or, coming soon, the Service’s web-based TMDL editor).

2 — Remove unused DAX variables from artifacts

Scans measures, calculated columns, and tables for DAX variables that are defined but never referenced. Expand the suggestion, right-click an affected artifact, and select “DAX expression with violations” to view highlighted unused variables.

3 — Disable Auto Date/Time

Auto Date/Time creates hidden date tables for every date column in your model, bloating size with no benefit if you have your own calendar table. The tool helps by:

  • Supplying generic calendar table DAX expressions
  • Tracking where auto date/time is used so you can replace references
  • Guiding you toward disabling the setting in File → Options → Data Load

4 — Remove field parameter tables that are not actually used

Field parameter tables are a special case due to their self-referential column structure. When a field parameter table exists solely for its own use and nowhere else, the tool suggests deleting both the table and any associated columns or measures used exclusively within it.

5 — Delete relationships connecting to unused tables

These relationships attach one or more model tables to tables that would otherwise be unused. Delete the relationship first, then remove the orphaned table.

6 — Remove hidden pages not used as tooltips, drillthrough or page navigation

Pages that meet all three criteria:

  • They are hidden
  • They don’t function as tooltip or drillthrough pages
  • They lack page navigation functionality

Delete these pages and remove any artifacts appearing exclusively on them.

7 — Remove hidden visuals not referenced in bookmarks

Visuals that are both hidden and unreferenced in any bookmark. Delete the visual and any artifacts used exclusively within it.

Note: Personal bookmarks created in the Power BI Service are not considered in this suggestion.

8 — Consider removing reports without recent consumption

Available in Shared Model Online modes only — Measure Killer needs to fetch report page views from the Power BI activity log, which requires the Get report page views setting to be enabled in the scan configuration. Reports showing zero opens within the past 28 days receive this suggestion. Unselect such reports, rerun the analysis, and then remove artifacts appearing exclusively within them.

9 — Remove pages with no views

Targets individual pages displaying zero views within 28 days (as opposed to suggestion 8 which targets entire reports). Delete these pages and remove associated artifacts.

10 — Verify if auto-detected relationships are necessary

When “Auto detect new relationships after data is loaded” remains enabled, Power BI automatically creates relationships. This suggestion flags them for verification. Probably the hardest suggestion to apply — verifying if you need an active relationship is usually not straightforward. Use Table Relations to understand the impact before removing.

11 — Remove artifacts used only in filters without active filters

Some columns appear solely in visual, page, or report-level filters where no active filter query exists (appearing as “is (All)” in Power BI) and no bookmark activation occurs. Remove the filter first, then delete associated artifacts.

12 — Verify columns used only as dynamic M parameters

Columns appearing only as dynamic Power Query parameters in DirectQuery tables and nowhere else. Unbind such columns by opening Model view, selecting the column, and clearing the “Bind to parameter” dropdown.

13 — Verify unused key columns in import mode tables

For Import tables containing key columns that lack other uses. Unmark them as keys, verify that relationships and DAX continue functioning, and delete them if appropriate. DirectQuery and Hybrid tables are excluded from this suggestion.

14 — Verify columns only used as table row labels

Columns marked as Default Label but unused elsewhere. Unmark them, verify that visuals and measures continue working, then delete if necessary.

15 — Consider excluding the semantic model’s default report

In shared (golden) model architectures with multiple connected reports, the dataset’s automatically generated report often contains artifacts nobody uses. Exclude that report from the analysis, rerun the tool, and remove artifacts appearing only there.

16 — Disable isAvailableInMdx for columns not used in Analyze in Excel

Columns marked with isAvailableInMdx = true but unused in Analyze in Excel. Disabling this property reduces storage. Modify through the TMDL view in Power BI Desktop or the Service.