Semantic Models

[binitafulpagare]

Hello everyone,

As Microsoft Fabric projects become larger, managing semantic models efficiently becomes increasingly important.

I'd like to know how experienced teams approach semantic model design.

Do you create one centralized model for the organization, or multiple domain-specific models? How do you balance performance, maintainability, and business requirements?

Any design recommendations or lessons learned would be greatly appreciated.

Thank you!

[v-abhinavmu · Accepted solution]

Hi @binitafulpagare,

Thanks for reaching out to the Microsoft Fabric Community forum.

It depends on your organization's size, data domains, and reporting requirements. However, many organizations follow a domain-driven approach rather than relying on one large semantic model.

Here are some general recommendations:

• Use domain-specific semantic models (for example, Sales, Finance, HR, and Operations) to make models easier to maintain, secure, and optimize.
• Avoid creating a single, monolithic semantic model unless your business requirements justify it, as large models can become more difficult to maintain, govern, and optimize over time.
• Promote consistency by using shared dimensions (such as Date, Customer, and Product) and standardized business definitions across models.
• Reuse existing semantic models whenever possible instead of creating duplicate models for similar reporting scenarios.
• Implement Row-Level Security (RLS) and Object-Level Security (OLS) where appropriate to support different user groups without duplicating models.
• Optimize performance by following a star schema, keeping only the required tables and columns, and minimizing unnecessary calculations.
• Establish good governance practices by using clear naming conventions, documentation, version control, Git integration, and deployment pipelines as your Fabric environment grows.

In practice, many organizations adopt a hybrid approach:

• Enterprise-certified semantic models provide trusted, reusable models for common business reporting.
• Domain-specific semantic models address specialized departmental requirements while maintaining consistent business definitions across the organization.

This approach often provides a good balance between scalability, maintainability, performance, governance, and business flexibility.

For additional guidance, Microsoft recommends designing semantic models using a star schema, as it improves both performance and usability:

Understand star schema and the importance for Power BI - Microsoft Learn

Understand star schema and the importance for Power BI - Power BI | Microsoft Learn

I hope this helps. Please feel free to reach out if you have any further questions.
Thank you.

[v-abhinavmu]

Hi @binitafulpagare,

Just following up to see if you had a chance to review the information shared. If you have any additional questions or need further clarification, please feel free to let us know, we'll be happy to assist.

Thank you.

[v-abhinavmu]

Hi @binitafulpagare,

I wanted to check if you had the opportunity to review the information provided. Please feel free to contact us if you have any further questions. 

Thank you.

[v-abhinavmu · Accepted solution]

Hi @binitafulpagare,

Thanks for reaching out to the Microsoft Fabric Community forum.

It depends on your organization's size, data domains, and reporting requirements. However, many organizations follow a domain-driven approach rather than relying on one large semantic model.

Here are some general recommendations:

• Use domain-specific semantic models (for example, Sales, Finance, HR, and Operations) to make models easier to maintain, secure, and optimize.
• Avoid creating a single, monolithic semantic model unless your business requirements justify it, as large models can become more difficult to maintain, govern, and optimize over time.
• Promote consistency by using shared dimensions (such as Date, Customer, and Product) and standardized business definitions across models.
• Reuse existing semantic models whenever possible instead of creating duplicate models for similar reporting scenarios.
• Implement Row-Level Security (RLS) and Object-Level Security (OLS) where appropriate to support different user groups without duplicating models.
• Optimize performance by following a star schema, keeping only the required tables and columns, and minimizing unnecessary calculations.
• Establish good governance practices by using clear naming conventions, documentation, version control, Git integration, and deployment pipelines as your Fabric environment grows.

In practice, many organizations adopt a hybrid approach:

• Enterprise-certified semantic models provide trusted, reusable models for common business reporting.
• Domain-specific semantic models address specialized departmental requirements while maintaining consistent business definitions across the organization.

This approach often provides a good balance between scalability, maintainability, performance, governance, and business flexibility.

For additional guidance, Microsoft recommends designing semantic models using a star schema, as it improves both performance and usability:

Understand star schema and the importance for Power BI - Microsoft Learn

Understand star schema and the importance for Power BI - Power BI | Microsoft Learn

I hope this helps. Please feel free to reach out if you have any further questions.
Thank you.

[binitafulpagare]

Hi @v-abhinavmu,

Thank you for the detailed explanation and for sharing the Microsoft Learn resource.

The hybrid approach you described—using enterprise-certified semantic models alongside domain-specific models—makes a lot of sense for balancing governance and flexibility. I also found your points about reusing semantic models, following a star schema, and implementing RLS/OLS particularly helpful.

I have one follow-up question based on real-world implementations: As an organization grows, how do teams decide when a domain-specific semantic model should be merged into an enterprise-certified model, or when it should remain independent? I'm curious whether this decision is typically driven by dataset size, user adoption, governance requirements, or another factor.

I'd also be interested to hear how other community members handle this as their Microsoft Fabric environments continue to scale.

Thank you again for your guidance!

[v-abhinavmu · Accepted solution]

Hi @binitafulpagare,

You're welcome, I'm glad you found the information helpful.

There isn't a prescribed threshold for merging a domain-specific semantic model into an enterprise-certified model. Instead, certification is intended for semantic models that meet your organization's quality standards and are considered reliable, authoritative and ready for use across the organization. Models that are valuable for sharing but haven't met those governance standards can be promoted instead.

The decision is typically based on your organization's governance and certification criteria, rather than on factors such as dataset size alone.

Additionally, you may find the following Microsoft Learn articles helpful:

Endorse your content - Power BI | Microsoft Learn

Promote and Certify Power BI Content with Endorsement - Power BI | Microsoft Learn

Thank you.