Semantic Models Comes First: Managing Models for Growth and Self-Service

The semantic model should be the stable foundation of your analytics environment. It should not a place for report-specific experiments, formatting measures, or quick fixes. When the core model stays clean and well-structured, teams can extend it through composite models and report-layer logic without adding technical debt.

This session presents a practical semantic model lifecycle: how to design, publish, extend, and refactor models without breaking downstream reports. We’ll compare when to use calculation groups vs. DAX UDFs, where field parameters belong, and how to use Semantic Link and the Power BI MCP server to safely refactor, rename, reorganize, and validate dependencies across environments.

You’ll learn:

What belongs in the shared semantic model vs. the report layer

How to structure measures, folders, relationships, and calc groups for clarity

How composite models enable report-specific logic without polluting the core model

When to choose UDFs vs. calculation groups for reusable logic

How Semantic Link and the MCP server support safe refactoring and lifecycle maintenance

Walk away with a framework for building once, extending safely, and preventing model clutter as your reporting ecosystem grows.

Learn how to separate core semantic-model logic from report-specific logic to keep the shared model clean and stable. Understand a practical lifecycle for designing, extending, and safely refactoring semantic models without breaking reports. Know when to use calc groups, UDFs, and field parameters so reusable logic stays organized and avoids model clutter.