How to create a metric definition

A metric definition turns raw transaction data into a business value you can reuse, such as net revenue, number of deals or deposits from a region. Commission schemes and targets then use that value. Defining a metric once, and testing it well, means every scheme and target that relies on it counts the same thing in the same way.

Last updated: 31 Oct 2026

Before you start

Have these ready:

A written description of the value and what it will be used for
The source column that holds the value
The agreed way to combine values (the aggregation)
Any filters the value needs, for example only closed deals
A few sample records: some that should be counted, some that shouldn't, and some right on a boundary

Step by step

Go to Settings > Metric Definitions and look at how existing metrics are named.
Create a new metric definition.
Enter a Metric Name, a Code and a description. Use uppercase letters and underscores for the code, for example NET_REVENUE. In the description, say exactly what the metric includes.
Choose an aggregation: Sum, Count, Average, Min, Max or Date Difference.
Select the Value Column that holds the number.
Turn on sign mapping only if your agreed business rule needs it.
Add filters by choosing a field, an operator and a value.
If you use Import Deal Rules, review every condition that is brought in.
Save the metric definition.
Test it with matching, non-matching and boundary records before you use it in a scheme or target.

Choosing the right aggregation

Sum adds the values, for example total revenue.
Count counts the matching records, for example number of deals.
Average, Min and Max give the mean, smallest or largest value.
Date Difference measures the time between two dates.

Example (illustrative only): three closed deals worth $4,000, $6,000 and $10,000. Sum gives $20,000, Count gives 3, Average gives about $6,667, Min gives $4,000 and Max gives $10,000. Work out your own sample by hand first, then check that the metric returns the same answer.

Test filters one at a time

Filters that are too narrow leave out valid records. Filters that are too broad let in records that shouldn't count. When a metric uses more than one filter, test each filter on its own first, then test them together and confirm the result matches your hand-calculated sample.

Check before you edit a shared metric

One metric can feed several schemes and targets. Before you change it, check which schemes and targets use it, so a fix for one doesn't change results somewhere else. Keep the small hand-calculated sample as evidence for whoever approves the change.

Common problems and quick checks

The metric returns no value: check the selected column, the period, the source records, the filters and the sign mapping.
It includes unexpected records: review every filter and any imported deal rule. Test each one separately.
The code is rejected: use only uppercase letters and underscores.
The total is wrong: compare the aggregation with your hand-calculated sample, using the same period and filters.

Bring a sample of your own data to a demo.

Book a Demo

9:41