Measure Killer · Use Cases

Find the unused.
Trace the lineage.
Reclaim your tenant.

Governance, lineage and clean-up for Power BI and Microsoft Fabric — from a single .pbix file to a tenant-wide scan.

Built by Brunner BI · Zug, Switzerland · 2026

At a glance

What Measure Killer does.

Measure Killer is the governance and clean-up tool the Power BI community has been quietly relying on. It traces every column, measure and report-level reference inside your models — and across your tenant — so you can clean up safely, audit access, and rein in capacity costs without breaking what works.

900,000+

times Measure Killer has been used in the last 12 months — across more than 160 countries and every category of Power BI / Fabric customer.

160+

countries reached worldwide

1,000+

sessions from our most active user (Ireland)

Three ways to use it

Free

Measure Killer Free

Desktop tool for Power BI cleanup — free forever, personal & commercial use. No data leaves your machine. Run it on any .pbix file to find unused columns, measures and tables.

Enterprise

MK Enterprise

Tenant-wide scans across Power BI Service and Microsoft Fabric. Lineage from data source to report visual, access auditing, governance, and everything from Free. Free trial available on request.

Automation

MK Automation

Scheduled, automated scans inside your own Microsoft Fabric tenant. Brunner BI sets it up and ships 30+ sample notebooks; you own the schedules and the runs. Includes a company-wide desktop license.

Pillar 1 · Paid

Tenant Analysis.
One pane of glass for every Power BI & Fabric item.

For every workspace, semantic model, report, dataflow and notebook in your tenant — Measure Killer shows the metadata, lineage, access and usage your governance team has been begging for. Read-only, runs against the admin APIs, no data ever moves.

Tenant summary: every workspace, item and user
End-to-end lineage from data sources to reports
Access & permissions audit, incl. AD / Entra groups
Refresh, best practices and KPIs for every semantic model
Inventory Dataflows, Apps, Reports and Fabric Notebooks
Search every DAX, M and notebook line in seconds

Where to start

Run a tenant scan with a Fabric admin account. Within minutes you get a full inventory of every workspace and item, ready to filter by owner, capacity, refresh status or last-activity date — and the lineage and access views start populating from the same scan.

Pillar 1 · Inside Tenant Analysis

Eight views, one scan.

A single tenant scan populates every one of these views — no extra configuration, no separate jobs. Most governance teams discover something on day one that nobody knew existed.

Tenant summary

Every workspace, item and user across Power BI and Fabric — with owners, creation date and last activity, all in one table.

End-to-end data lineage

Trace every dependency from data source to report, across workspaces. Search the lineage view for a specific table or column.

Access & permissions

Workspace and per-item access, with AD / Entra groups expanded so you actually see who has access to what.

Semantic models

Refresh history, schedule, average duration, tunable best-practice rules (choose which to run, adjust thresholds), KPI inventory and storage mode for every model.

Dataflows · Gen1 + Gen2

Inventory every Dataflow with its M code, refresh stats and number of queries. Search across all M expressions.

Apps & audiences

All Apps — including organizational Apps — with audiences, 28-day consumption, contacts and access details.

Reports · performance & usage

Views, page views, average load time, certified vs custom visuals, report-level measures and report subscriptions — rolled up per report.

Fabric Notebooks

Locations, access (for admins) and lines of code per notebook. Search every notebook for packages or keywords.

Export to JSON for AI or downstream tools

Every slice of the scan — DAX, M, lineage, refresh events, RLS, best-practice violations, page-level consumption with load times — as raw JSON files. Feed them to a frontier AI for prioritised optimisation suggestions, or pipe them into a governance dashboard. Capacity cost lands in v2.9.5.

AI-ready governance

Get your Power BI ready for AI.

The single biggest hit to AI accuracy is a model bloated with columns and measures nobody actually uses. Every extra field is one more thing Copilot in Power BI, Microsoft Fabric Data Agent and Power BI Q&A have to choose between — and the more dead weight in the model, the more often the agent picks the wrong one. Measure Killer cleans the model before the AI sees it.

Better accuracy

Sharper field set, better answers

Every unused column and measure is one more option the agent has to choose between — and one more way to get the wrong answer. Strip them out and Copilot reasons over only the fields people actually use. The result is noticeably more accurate AI — one of the highest-leverage gains you can deliver.

Audit usage

Find what's dead, down to the page

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

Free up capacity

Refresh optimisation

Every needless refresh burns Premium / F-SKU capacity that Copilot, Microsoft Fabric Data Agent and your other AI workloads are competing for. Measure Killer exposes refresh schedules, durations and history across semantic models, dataflows and notebooks — right-size them and reclaim capacity for AI.

Trustworthy answers

Audit relationships & RLS

Ambiguous joins produce wrong answers — sometimes silently. Missing row-level security exposes data that shouldn't be exposed. Measure Killer audits relationships end-to-end and inventories every model's RLS roles, security filters and member assignments.

The compounding payoff

Clean the model once with Measure Killer, reap the benefits everywhere — Copilot in Power BI, Microsoft Fabric Data Agent, custom AI agents and whatever you bolt on next. AI-readiness isn't a one-off project; it's what a well-governed semantic model looks like in 2026.

Pillar 2 · Free desktop

Used / Unused.
Column-level lineage for impact analysis + clean-up.

For any column, measure or table — know exactly what it powers before you change or remove it. Then clean up confidently. Free for personal and commercial use; no time limit, no telemetry, no data leaves your machine.

Column-level lineage across reports, models and dataflows
Impact analysis: see what breaks before you remove anything
Detect unused columns, measures and tables — at any scale
Shrink semantic model size and refresh times
Find implicit measures and other anti-patterns
Kill via XMLA — DAX measures and calc columns auto-backed up

Where to start

Download the free desktop app at measurekiller.com/download, open any .pbix file, and let Measure Killer analyse the model — watch the unused-column list grow alongside the size you'd reclaim. No license, no signup, no telemetry.

Pillar 2 · Inside Used / Unused

What we analyze. What we detect. What you can do.

From a single .pbix on your laptop to a tenant-wide scan, Measure Killer answers the same three questions for every model: what's in scope, what's actually being used, and how to clean it up without breaking anything.

What we analyze

Power BI reports (.pbix · .pbip)

Local files or downloaded from the Service. Visual-level usage, page-level filters, report-level measures and mobile-layout artifacts.

Service reports & shared datasets

Online thin / live reports analyzed without downloading — composite and child models, personal, Pro and Fabric workspaces. Full downstream lineage from a shared model to every report, Excel pivot and chained model.

Paginated reports

Same usage detection for paginated reports — parameters, embedded datasets and data sources included.

Excel files · Analyze in Excel

Find which columns power your Excel users — pivot tables are detected by querying activity logs.

What we detect

Unused columns, measures and tables

Across every report and model — with statuses like "Used in visuals", "Used in measures" or "Unused".

Calc groups · KPIs · field parameters

Usage tracked for the modern model objects too — calculation groups, KPIs and field parameters.

Implicit measures & anti-patterns

Spot implicit measures, unused calc groups, RLS gaps and other DAX best-practice violations.

Column-level lineage

For any column or measure, see every downstream visual and upstream source — your impact analysis, done.

What you can do

Clean TMDL

Cleaned TMDL of your model with unused objects removed — the fastest and safest way to optimize a semantic model nowadays. Apply via the TMDL view in Power BI Desktop or the Service.

DAX backup & restore

When Measure Killer removes a DAX measure or calculated column, the expression is auto-backed up so you can restore it with one click. Power Query columns are automatically backed up through an additional step in your M Code.

Save & resume an analysis

.measurekiller files save the full scan results. Exit today, pick up tomorrow — or hand the analysis to a colleague who runs the cleanup under their own permissions.

Kill via XMLA — online

Remove unused objects from semantic models directly through XMLA. No round-trip through Power BI Desktop.

Clean your model — advanced optimizations

Step-by-step suggestions to clean and optimize your model — remove unused artifacts, disable auto-datetime, prune unused field-parameter tables, drop hidden bookmark visuals — each with an estimated size reduction so the biggest wins surface first.

Performance & Cost focus

Smaller models, faster reports,
leaner capacity bills.

Most semantic models carry 60–80% of weight nobody is using. Removing the dead weight isn't optional — it's the cheapest performance optimization you'll ever do.

−78%

Up to 78% model-size reduction we see on customers' largest Power BI semantic models. Smaller models mean faster refreshes, faster report load times, and lower Premium / Fabric F-SKU compute costs.

Before

14 GB model takes 38 min to refresh; users blame the dashboard.
"We just bumped capacity again — is it the model or the queries?"
36 reports connect to this model — nobody's checked which are actually being used.
Every change in the model is a coin flip — does it break a report?

With Measure Killer

Same model trimmed to 3 GB — refreshes in 9 minutes.
Audit shows: 60% of columns never powered a visual or measure.
12 of 36 connected reports haven't been consumed — worth checking if they're still needed.
Impact analysis surfaces every downstream reference. Delete with confidence.

Use cases by role

Different teams, same scan.

Measure Killer's value differs by who's using it. Four common framings — each is a single question your team is already asking.

BI Lead / Power BI Developer

"I need to clean up this model — but if I remove the wrong column, what breaks?"

Column-level lineage + impact analysis shows every downstream reference before you delete anything. Then Measure Killer cleans it up automatically — without breaking the model. Auto-backup of removed DAX measures and calc columns means a one-click restore if you change your mind.

Dozens of hours saved vs. doing it by hand.

Data Governance

"How long do reports take to load? Who has access to what? Is row-level security set up correctly?"

Per-item access view (workspace + item level), report performance metrics, and a complete RLS inventory — every model with RLS, all roles, security filters and member assignments.

Audit-ready evidence in one export.

Fabric / Service Administrator

"The tenant grew sideways. What's actually in there, what changed last month, and are refreshes still running?"

Full tenant inventory of every workspace and item, change tracking week-over-week, refresh-event timeline for semantic models, dataflows and Fabric notebooks.

Stop discovering content by accident.

Data Engineer

"We renamed a warehouse view — where is it used in Power BI?"

End-to-end lineage from data source → semantic model → report visual. Trace any source table or view down to every Power BI and Fabric artifact that depends on it.

No more breaking dashboards you didn't know existed.

Trust & security

Designed to stay out of your way.
Built for IT review.

Your security team's first three questions, answered up front.

No data leaves your machine

Measure Killer Free runs entirely on your desktop. Your .pbix files, models, and DAX never leave your device — there is no telemetry of model contents.

Read-only tenant scans

Tenant scans use Power BI / Fabric admin APIs in read-only mode. We never modify, copy, or move your reports, models or data.

Customer-controlled storage

Enterprise scan results are stored on infrastructure you control. You configure retention, access and export policies. We do not retain copies.

Built by a Microsoft MVP

Gregor Brunner, our founder, is a Microsoft MVP for Data Platform — recognized for sustained contributions to the Power BI and Fabric community.

Full security whitepaper available at measurekiller.com/trust-center — architecture, threat model, sub-processor list and DPA on request.

Get started

Three ways in.