Metrics

Metrics are specific calculations that need to be exactly right every time. While instructions give Qluent the background context, metrics define how to calculate your key numbers.

When a user asks a question that matches a metric, Qluent uses your exact SQL—ensuring the calculation is always correct. A verification badge is displayed so users know the answer follows your business definitions.

Use case

Some calculations need to be consistent across all reports:

• Revenue must exclude refunds and use correct tax treatment
• Customer lifetime value has a specific formula your team agreed on
• Conversion rates must match your other reporting tools

Metrics ensure these calculations stay the same, even when Qluent adapts them for different filters or groupings.

How it works

1. You define a metric with a name and SQL query
2. When a user asks a related question, Qluent finds the matching metric
3. Qluent adapts your SQL (adding filters, groupings) while keeping the core calculation
4. Users see the verified badge

Example

Metric:

Name: Total Revenue

SELECT SUM(order_total - refund_amount) as revenue
FROM orders
WHERE status = 'completed'

User asks: "What's our revenue by product category?"

Qluent adapts:

SELECT category, SUM(order_total - refund_amount) as revenue
FROM orders
WHERE status = 'completed'
GROUP BY category

The grouping is added, but the revenue calculation (order_total minus refunds, only completed orders) stays exactly the same.

Creating metrics

Add a short name and the SQL code that answers it. Qluent adapts the SQL to different questions, so you don't need to provide all variations.

Tips

• Keep SQL simple - Let Qluent handle variations
• One metric per concept - Don't combine multiple KPIs in one metric
• Use business-friendly names - "Monthly Revenue" not sum_order_total_monthly
• Add comments for complex logic - Helps Qluent understand the intent

Example metrics

• Total Revenue (excluding refunds and cancellations)
• Customer Lifetime Value (LTV)
• Conversion Rate
• Monthly Active Users (MAU)
• Average Order Value

Validation

Qluent validates metrics against your data model:

• All referenced tables must exist and be enabled
• All referenced columns must exist and be enabled
• SQL must be valid

If validation fails, the metric won't be used. If your data model changes (tables renamed, columns removed), revalidate your metrics.

Metrics vs validated queries

Metrics	Validated Queries
Define specific calculations	Teach query patterns
Used directly when matched	Influence how new queries are written
For KPIs that must be exact	For showing how to join tables, filter data, etc.

Use metrics when a calculation must be exactly right. Use validated queries to teach Qluent how to write queries for your data.