Coming September 2026 · Industry first See preview

Column-level lineage from data source to Power BI visual

The ultimate governance & optimization tool for Power BI and Microsoft Fabric. Detect unused measures and columns, trace end-to-end lineage, and clean up your models with confidence.

Free download → Request a free trial → No credit card required

✓ No data leaves your machine ✓ Metadata-only interaction ✓ Built by a Microsoft MVP ✓ Swiss-engineered

Measure Killer — Semantic Model Size

Name Size Used

Sales Cube 18.4 GB

18%

Marketing Attribution 14.2 GB

22%

Customer LTV Sandbox 11.6 GB

27%

Region Sales (Legacy) 234.7 MB

47%

Q1 Forecast 89.4 MB

65%

Inventory Live 67.8 MB

75%

Executive KPIs 5.2 MB

100%

~35 GB reclaimable Clean up →

Potential savings

78%

v2.9.4 · Live

Trusted by teams at

Paid · Power BI Service

Measure Killer online mode

Optimize and clean your shared Power BI semantic models.

✓ Across all connected Power BI reports
✓ Paginated Reports
✓ Excel files
✓ Downstream / composite models (incl. changes)
✓ Consumption details per report — de-select unused ones for an even deeper clean-up

Give it a try

Request a free trial → Use our sample data → Watch the video →

Measure Killer online mode — full lineage tree from semantic model to every connected Power BI report — Online mode · Full lineage tree From the semantic model down to every connected Power BI report — including downstream and composite models. Click to enlarge. Online mode docs →

Measure Killer online mode — column-level usage details showing every visual, filter, measure and other reference for a single column or measure — Online mode · Where each field is used For any column or measure, see every visual, filter, calculation group, RLS rule and other reference — tracked across 26 categories i in total. Click to enlarge. Online scan docs →

Going tenant-wide

Tenant analysis — governance, end‑to‑end.

Same online connection, broader scope. Scan your whole tenant and answer the questions IT, security and finance keep asking — who owns what, who has access to what, what's actually being used. Tenant analysis docs →

✓ Tenant summary — workspaces, items, owners, last activity
✓ Access & permissions with AD / Entra groups expanded
✓ End-to-end lineage from data source to report, across workspaces
✓ Refresh history, schedules & best-practice violations
✓ Reports — opens, page views, load times, custom visuals and subscriptions
✓ Apps, audiences, dataflows & Fabric notebooks inventory
✓ Find duplicate semantic models & reports (similarity score)
✓ Capacity metrics — accumulate the 14-day window into long-term history you can export & analyze

Request a free trial →

Measure Killer Tenant Analysis — Semantic models tab listing every model in the tenant with used size, potential savings, refresh schedule, best-practice score and more — Tenant Analysis · Semantic models Every semantic model in your tenant — used size, potential savings, refresh schedule, best-practice score and more. Click to enlarge. Semantic models docs →

Coming September 2026 · Industry first

Column-level lineage from your database to every Power BI visual.

Trace a single column from a source table through every dataflow, semantic model and report — all the way to the individual visual that renders it. End-to-end. No gaps. Nobody else does this.

See what's coming Shipping September 2026

Get ready for AI

Get your Power BI ready for AI.

Copilot in Power BI, Microsoft Fabric Data Agent and every BI Q&A tool are only as good as the model behind them. Three high-leverage cleanups before you turn AI loose on your tenant.

Strip out noise

Less columns, measures, tables

Unused columns inflate the model the AI has to index. Unused measures get suggested as answers nobody intended. Unused tables muddy the lineage the AI is supposed to follow. Measure Killer finds and removes them — at tenant scale, with impact analysis so nothing breaks.

Archive what's dead

Reports nobody reads are noise too

Decommissioned reports and stale pages get suggested by Copilot alongside the ones your team actually uses. Measure Killer surfaces views per report and per page so you can archive the dead weight — Copilot only sees what's alive and worth your team's attention.

Best practices

Run on every model and report

Cryptic field names, undocumented measures, missing descriptions and anti-patterns all degrade AI suggestions. The best-practice analysis runs across every model and report in your tenant — fix the worst violations once and every AI tool gets better overnight.

AI as your analyst

Or hand AI the metadata and let it find the wins.

Measure Killer's tenant exports are the richest Power BI metadata you can get anywhere — every DAX expression, every M query, full report lineage, refresh events, best-practice violations, used vs. unused columns and measures, page-level consumption with average load times, and now capacity cost. Feed that to a frontier AI model and you get prioritized optimization suggestions ranked by real financial and performance impact — not generic best-practice noise.

Read the full "AI-ready Power BI" chapter in the Use Cases PDF →

99.99%

reference-detection accuracy

Remove the unused. Break nothing.

Cleanup is only safe if every reference is found. At 95% or even 98% accuracy, something will break on almost every single removal — there's always a reference hiding somewhere (a tooltip, a conditional-formatting rule, a measure inside another measure). Measure Killer's 99.99% is the best in the industry, and the difference between confidence and a quiet outage. The handful of edge cases that can still slip through are listed transparently.

See what we detect (and don't) →