TECH **Power BI Endorsement (Promoted vs Certified) & Data Lineage: Zero-Fluff Study Guide**

Power BI Endorsement (Promoted vs Certified) & Data Lineage: Zero-Fluff Study Guide

For engineers, analysts, and certification candidates who need to ship production-ready Power BI solutions—fast.

1. What This Is & Why It Matters

You’re handed a Power BI workspace with 50+ datasets, reports, and dashboards. Some are trusted by executives for board meetings; others are experimental prototypes. How do you instantly know which assets are production-ready? That’s where endorsement (Promoted vs Certified) comes in.

Meanwhile, your CFO asks: “Where does this ‘Revenue’ number in the Q3 report actually come from?” You dig through layers of Power Query, DAX, and SQL—only to realize the source is a manually edited Excel file. Data lineage is your GPS for tracing data from report to source, so you can answer that question in seconds.

Why this matters in production:
- Endorsement prevents "shadow analytics" (users wasting time on untrusted reports) and ensures governance.
- Data lineage reduces audit risk, speeds up debugging, and enables impact analysis (e.g., "If we change this SQL view, which reports break?").
- Ignoring this? You’ll drown in duplicate work, compliance violations, and "why does this number differ?" meetings.

Real-world scenario:
You’re a Power BI admin at a healthcare company. A critical patient-safety dashboard is using an outdated dataset. Without endorsement, users might rely on it unknowingly. Without lineage, you can’t trace the data back to its source (a deprecated SQL stored procedure). Result: Regulatory fines and a frantic weekend of manual validation.

2. Core Concepts & Components

Endorsement

• ? Promoted
• Definition: A dataset/report marked as "ready for production use" by a workspace admin or member.
• Production insight: Promoted assets appear in the "Recommended" section of Power BI’s "Get Data" experience, but aren’t guaranteed to be accurate—just vetted by a human.

• Example: A sales analyst marks their "Monthly Pipeline" report as Promoted after internal review.

• ? Certified

• Definition: A dataset/report that has passed formal validation (e.g., data accuracy, performance, security) and is endorsed by a central governance team.
• Production insight: Certified assets are the gold standard—they appear in Power BI’s "Certified" section and can be used for regulatory reporting.

• Example: The finance team certifies the "Quarterly Earnings" dataset after auditing its DAX measures.

• ? Endorsement vs. Sensitivity Labels

• Endorsement = Trust (is this asset reliable?).
• Sensitivity Labels = Security (who can access this asset?).
• Trap: A dataset can be Certified (trusted) but still have a "Public" sensitivity label (unsecure). Always check both.

Data Lineage

• ? Lineage View
• Definition: A visual map in Power BI Service showing dependencies between datasets, reports, dashboards, and data sources.
• Production insight: Lineage is read-only—you can’t edit it, but you can use it to trace errors or plan changes.

• Example: Clicking on a report in Lineage View shows all upstream datasets and downstream dashboards.

• ? Impact Analysis

• Definition: A feature that predicts which reports/dashboards will break if you modify a dataset.
• Production insight: Run impact analysis before changing a dataset to avoid breaking downstream reports.

• Example: You’re about to rename a column in a dataset. Impact analysis shows 12 reports will break—so you update them first.

• ? Dataflows

• Definition: Reusable ETL processes in Power BI that centralize data transformation and improve lineage.
• Production insight: Dataflows break direct dataset-to-source dependencies, making lineage cleaner and governance easier.

• Example: Instead of 10 datasets connecting to the same SQL table, they all connect to a single dataflow.

• ? Shared Datasets

• Definition: Datasets published to a workspace and reused across multiple reports.
• Production insight: Shared datasets simplify lineage (one dataset feeds many reports) but increase blast radius (a change affects all downstream reports).
• Example: A "Customer Master" dataset is shared across 20 reports. If you delete it, all 20 reports break.

3. Step-by-Step Hands-On: Endorse a Dataset & Trace Lineage

Prerequisites

• Power BI Pro or Premium license.
• Admin or Member permissions in a workspace.
• A dataset and report published to Power BI Service.

Task 1: Promote a Dataset

1. Open Power BI Service (app.powerbi.com) and navigate to your workspace.
2. Click "Datasets" in the left nav.
3. Hover over your dataset (e.g., "Sales Analytics") and click the three dots (⋮) → Settings.
4. Under "Endorsement," select "Promoted" → Apply.
5. Verification: The dataset now shows a blue "Promoted" badge in the workspace.

Task 2: Certify a Dataset (Admin-Only)

1. Ask your Power BI admin to enable certification for your workspace (if not already enabled).
2. Admin steps:
   • Go to Admin Portal → Tenant Settings → Certification.
   • Enable certification and specify who can certify (e.g., "Finance Team").
3. As a certifier, go to the dataset settings (same as Step 1).
4. Select "Certified" → Apply.
5. Verification: The dataset now shows a green "Certified" badge and appears in the "Certified" section of Power BI’s "Get Data" experience.

Task 3: Trace Lineage for a Report

1. Open a report in Power BI Service.
2. Click the "Lineage" tab (top-right, next to "Edit").
3. Explore the lineage graph:
4. Upstream: Click a dataset to see its data sources (e.g., SQL Server, Excel).
5. Downstream: Click a report to see which dashboards use it.
6. Run Impact Analysis:
7. Click a dataset → Impact Analysis.
8. Power BI lists all reports/dashboards that depend on it.
9. Example output:
   ```
   Impacted reports (3):
   • Sales Dashboard
   • Executive Summary
   • Regional Performance ```

4. ? Production-Ready Best Practices

Endorsement

• Governance:
• Create a certification policy (e.g., "All Certified datasets must have a data dictionary").
• Use Power Automate to auto-notify dataset owners when their assets are endorsed.
• Security:
• Certified datasets should have sensitivity labels (e.g., "Confidential").
• Restrict who can certify to a small governance team (e.g., "Data Stewards").
• Cost Optimization:
• Archive unendorsed datasets after 90 days of inactivity (use Power BI REST API to automate).
• Reliability:
• Tag endorsed datasets with metadata (e.g., "Owner: [email protected]", "Refresh Schedule: Daily").

Data Lineage

• Debugging:
• Use lineage to trace errors (e.g., "Why is this number wrong?" → Follow lineage to the source).
• Export lineage as a PDF for audits (click "Export" in Lineage View).
• Change Management:
• Always run impact analysis before modifying a dataset.
• Document changes in the dataset description (e.g., "Updated 2024-05-01: Added 'Customer Segment' column").
• Performance:
• Replace direct SQL queries with dataflows to simplify lineage.
• Avoid circular dependencies (e.g., Dataset A → Dataset B → Dataset A).
• Observability:
• Monitor lineage complexity (e.g., "This dataset has 15 upstream sources—consider consolidating").
• Set up alerts for broken lineage (e.g., "Dataset X’s source table was deleted").

5. ⚠️ Common Mistakes & Traps

6. ? Exam/Certification Focus (PL-300)

Question Patterns

1. "Which endorsement level is appropriate for a dataset used in regulatory reporting?"
2. Trap: "Promoted" (wrong—Certified is required for compliance).

3. Answer: Certified.

4. "You modify a column in a dataset. Which feature shows which reports will break?"

5. Trap: "Lineage View" (shows dependencies but doesn’t predict impact).

6. Answer: Impact Analysis.

7. "How can you simplify lineage for 50 reports using the same SQL table?"

8. Trap: "Use direct connections" (creates messy lineage).
9. Answer: Create a shared dataset or dataflow.

Key Distinctions

7. ? Hands-On Challenge

Scenario:
You’re a Power BI admin. A user complains that their "Sales Dashboard" report is showing outdated data. Trace the lineage to find the root cause.

Steps:
1. Open the "Sales Dashboard" report in Power BI Service.
2. Click the Lineage tab.
3. Follow the upstream dependencies to find the dataset.
4. Check the dataset’s refresh history (Settings → Refresh History).
5. Solution: The dataset’s last refresh failed due to a SQL timeout. Fix: Increase the timeout or optimize the query.

Why it works:
Lineage shows you exactly where the data comes from, and refresh history tells you if the data is stale.

8. ? Rapid-Reference Crib Sheet

Endorsement

• Promoted: Settings → Endorsement → Promoted (blue badge).
• Certified: Settings → Endorsement → Certified (green badge, requires admin setup).
• ⚠️ Default: No endorsement (users won’t see it in "Recommended").
• ⚠️ Certification is workspace-specific (a dataset can be Certified in one workspace but not another).

Data Lineage

• Lineage View: Click "Lineage" tab in a report/dataset.
• Impact Analysis: Click a dataset → "Impact Analysis."
• Dataflows: Improve lineage by centralizing ETL.
• Shared Datasets: Simplify lineage but increase blast radius.
• ⚠️ Lineage is read-only—you can’t edit it, only use it for tracing.
• ⚠️ Lineage doesn’t show Power Query steps—only dataset-to-source connections.

9. ? Where to Go Next

1. Microsoft Docs: Endorsement in Power BI
2. Microsoft Docs: Data Lineage in Power BI
3. PL-300 Exam Guide: Endorsement & Lineage
4. Power BI Governance Whitepaper (Best practices for enterprise deployments).