Build Your First Metric

In this guide, we will walk through the process of creating a business metric from raw data. We assume you have already ingested data into a Topic (e.g., order).

Goal: Create a "Total Revenue" metric to track sales performance.

Step 1: Create a Semantic Model

Before defining a metric, we must map the raw data topic to a business entity.

1. Navigate to Metrics Platform > Semantic Model.
2. Click Create Model.
3. Basic Info:
   • Name: Sales Model
   • Source Topic: Select your datamart topic (e.g., order).
4. Define Measures:
   • Find the field representing the amount (e.g., order_amount).
   • Click Add to Measures.
   • Rename it to Revenue (Business friendly name).
5. Define Time Dimension:
   • Find the transaction date (e.g., order_date).
   • Mark it as the Time Dimension.
6. Click Save.

Why? This decouples your metrics from cryptic database column names.

Step 2: Define the Metric

Now we create the standardized KPI.

1. Navigate to Metrics Management.
2. Click New Metric.
3. Configuration:
   • Name: Total Revenue
   • Semantic Model: Select Sales Model.
   • Measure: Select Revenue.
   • Aggregation: Sum.
   • Format: Currency (e.g., $).
4. Click Save.

You now have a governed metric that can be reused across the organization.

Step 3: Visualize and Analyze

Finally, let's use the metric.

1. Navigate to Metrics Analysis.
2. Create a New Dashboard.
3. Add Card:
   • Select Total Revenue from the library.
   • The system immediately shows the total number.
4. Break Down:
   • Drag a dimension (e.g., Region or Product Type) to the Breakdown area.
   • The chart automatically updates (e.g., to a Bar Chart or Line Chart).
5. Save your dashboard.

Next Steps

• Explore advanced Metric types
• Configure Alerts for this metric