# Medallion Architecture: A complete Guide

> The Medallion Architecture organizes data into Bronze, Silver, and Gold layers for quality and governance. A guide with Databricks examples.

Source: https://www.erathos.com/en/blog/medallion-architecture-complete-guide
Em português: https://www.erathos.com/blog/medallion-architecture-complete-guide
Published: 2026-01-08
Category: Data Engineering

![Medallion Architecture: A complete Guide](https://cms-media.erathos.com/KdEVV8tSbiqntM17ipj3vM2uqU-1.png)

The **Medallion Architecture** is a modern data organization pattern widely used in **Data Warehouses**, **Data Lakes**, and **Lakehouses**. Its goal is to structure data in **progressive quality layers**, facilitating **governance, reprocessing, scalability, and analytical consumption**.

This model was popularized by Databricks and today is considered an **essential best practice** in modern data architectures.

## What is Medallion Architecture?

The **Medallion Architecture** organizes data into **three main layers**, where each layer represents a higher level of refinement and value:

- **Bronze** → raw data
- **Silver** → cleansed and trusted data
- **Gold** → business-ready data ready for consumption

> Quick summary: data enters raw, gets cleaned, and is then transformed into business metrics.

## Medallion Architecture Overview

![fbms7GB8z5tpQdccoLbOdxBF7UQ.png](https://cms-media.erathos.com/fbms7GB8z5tpQdccoLbOdxBF7UQ-1.png)

## Bronze Layer: Raw Data

![wSw0IZwLUa00otbc9no0VATFyT0.png](https://cms-media.erathos.com/wSw0IZwLUa00otbc9no0VATFyT0-1.png)

### What is the Bronze layer?

The **Bronze** layer stores data **exactly as it arrives from the source**, without significant transformations.

### Bronze layer characteristics:

- Raw data
- Structure close to the source
- May contain errors and duplicates
- High volume
- Foundation for auditing and reprocessing

**Goal:** preserve the original data as the source of truth.

## Silver Layer: Cleansed and Enriched Data

![KDJezTNfq64C6yAtBMq9XbFhPrA.png](https://cms-media.erathos.com/KDJezTNfq64C6yAtBMq9XbFhPrA-1.png)

### What is the Silver layer?

The **Silver** layer contains **clean, standardized, and trusted** data, ready for consistent analysis.

### Common transformations:

- Deduplication
- Type casting and standardization
- Validation of required fields
- Enrichment through joins

**Goal:** ensure data quality and consistency.

## Gold Layer: Business Layer

![fbms7GB8z5tpQdccoLbOdxBF7UQ.png](https://cms-media.erathos.com/fbms7GB8z5tpQdccoLbOdxBF7UQ-1.png)

### What is the Gold layer?

The **Gold** layer represents the final level of the Medallion Architecture. Here, the data is **ready for business consumption**.

### Examples of Gold data:

- Monthly revenue
- Churn rate
- LTV
- Executive metrics

**Goal:** deliver direct value for decision-making.

## How the Medallion Architecture Scales with Business Growth

One of the main differentiators of the **Medallion Architecture** is its ability to scale **without requiring deep restructurings** as the business evolves.

As new data sources emerge:

- They land directly into the **Bronze layer**
- They do not impact existing models in Silver or Gold
- They maintain isolation between ingestion and consumption

This decoupling allows data teams to evolve pipelines and business models **incrementally**, without disrupting reports or analytical products already in production.

## Medallion Architecture and Evolving Business Rules

Business rules **change constantly**: metric definitions, segmentation criteria, financial calculations.

In the Medallion Architecture:

- The original data remains preserved in Bronze
- Silver ensures structural consistency
- Gold can be reprocessed whenever a rule changes

This eliminates a common issue in traditional Data Warehouses: **recalculating metrics without losing history or causing inconsistencies**.

The result is greater trust in the numbers and less reliance on manual fixes.

## Governance and Observability in Medallion Architecture

The clear separation between Bronze, Silver, and Gold simplifies the implementation of **data governance and observability**.

Each layer can have:

- Distinct SLAs
- Specific quality rules
- Independent monitoring

Additionally, issues become **localized**:

- Ingestion errors → Bronze
- Quality issues → Silver
- Metric inconsistencies → Gold

This reduces troubleshooting time and improves the reliability of the data ecosystem.

## Common Pitfalls When Implementing Medallion Architecture

Despite its conceptual simplicity, some mistakes are frequent:

- Applying business rules in Bronze
- Skipping the Silver layer to "move faster"
- Creating multiple definitions of metrics in Gold
- Not versioning transformations
- Treating the architecture merely as a naming convention

Avoiding these pitfalls is essential for the Medallion Architecture to deliver its expected benefits.

## Benefits of the Medallion Architecture

- Better data quality
- Governance and lineage tracking
- Reliable reprocessing
- Scalability
- Clear separation of concerns

## Best Practices in Medallion Architecture

- Do not apply business rules in Bronze
- Use columnar formats
- Version schemas per layer
- Monitor quality in Silver
- Document metrics in Gold

## When to Use Medallion Architecture?

Use the Medallion Architecture when:

- There are multiple data sources
- Governance is critical
- Reprocessing is frequent
- The Data Warehouse has become a bottleneck
- BI and Analytics need to scale

## FAQ – Medallion Architecture

### What is Medallion Architecture?

It is a data organization pattern in layers (Bronze, Silver, and Gold) that improves quality and governance.

### Are "Arquitetura Medalhão" and "Medallion Architecture" the same thing?

Yes. "Arquitetura Medalhão" is simply the Portuguese translation of Medallion Architecture.

### Can I consume data directly from the Silver layer?

Yes. This is common for exploratory analysis and data science.

## Conclusion

The **Medallion Architecture** is one of the most important design patterns for modern data architectures. By structuring data into progressive layers, it builds a **reliable, scalable, and growth-ready** environment.

If you are modernizing your Data Warehouse or building a Lakehouse, adopting the Medallion Architecture is an essential step.