> ## Documentation Index
> Fetch the complete documentation index at: https://lightdash-mintlify-cccf65ca.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Drivers

> Define which metrics influence each other in your .yml files and visualize those relationships automatically in Saved Trees.

**Drivers** are metrics that directly influence or feed into another metric. By defining drivers in your `.yml` files, you can automatically visualize how metrics relate to each other in [Saved Trees](/guides/metrics-catalog/saved-trees) — no manual drawing needed.

For example, `total_order_amount` is a driver of `total_revenue`, `average_order_size`, and `total_completed_order_amount` — because each of those metrics is influenced by the total order amount. On the canvas, this creates a tree where `total_order_amount` fans out to the metrics it drives.

### Defining drivers

Add a `drivers` property to any metric listing the metrics that flow into it. You can reference metrics from the same table by name, or from other tables using the `table.metric` syntax:

<Tabs>
  <Tab title="dbt v1.9 and earlier">
    ```yaml theme={null}
    models:
      - name: orders
        columns:
          - name: order_amount
            meta:
              metrics:
                total_revenue:
                  type: sum
                  drivers:
                    - total_order_amount
                average_order_size:
                  type: average
                  drivers:
                    - total_order_amount
                total_completed_order_amount:
                  type: sum
                  drivers:
                    - total_order_amount
    ```
  </Tab>

  <Tab title="dbt v1.10+ and Fusion">
    ```yaml theme={null}
    models:
      - name: orders
        columns:
          - name: order_amount
            config:
              meta:
                metrics:
                  total_revenue:
                    type: sum
                    drivers:
                      - total_order_amount
                  average_order_size:
                    type: average
                    drivers:
                      - total_order_amount
                  total_completed_order_amount:
                    type: sum
                    drivers:
                      - total_order_amount
    ```
  </Tab>

  <Tab title="Lightdash YAML">
    ```yaml theme={null}
    type: model
    name: orders

    dimensions:
      - name: order_amount

    metrics:
      total_revenue:
        type: sum
        sql: ${TABLE}.order_amount
        drivers:
          - total_order_amount
      average_order_size:
        type: average
        sql: ${TABLE}.order_amount
        drivers:
          - total_order_amount
      total_completed_order_amount:
        type: sum
        sql: ${TABLE}.order_amount
        drivers:
          - total_order_amount
    ```
  </Tab>
</Tabs>

This creates edges on the canvas pointing from each driver to the current metric (e.g., `total_order_amount` → `total_revenue`, `total_order_amount` → `average_order_size`, etc.).

#### Cross-table drivers

Drivers don't have to live in the same table as the metric they drive. Use the `table.metric` syntax to reference a metric from another table:

```yaml theme={null}
metrics:
  conversion_rate:
    type: number
    sql: ${signups} / ${visits}
    drivers:
      - signups              # same-table metric
      - marketing.ad_spend   # cross-table: 'table.metric_name'
```

### How drivers appear on the canvas

* **Automatic**: Driver edges appear as soon as both the driver and target metrics are on the canvas.
* **Read-only**: You cannot delete or modify driver edges from the UI — they are managed through your `.yml` configuration.

In Saved Trees, driver edges are displayed alongside manually-created edges, giving you a complete picture of both the relationships you've drawn and the ones defined in your data model.

<Info>
  Driver edges cannot be removed from a Saved Tree. If you don't want a driver edge to appear, remove one of its endpoint nodes from the canvas.
</Info>