Funnel has two types of custom metrics: formula custom metrics, and rule custom metrics. They operate on different levels of aggregation; rule metrics are calculated per row pre-aggregation, formula metrics operate on totalled data, post-aggregation. Read more about how formula and rule metrics are calculated.
Rule metrics work like any other metric in Data Studio so the rest of this article concerns formula metrics.
Funnel formula metrics
A formula metric is a simple arithmetic formula like for instance:
CPC = cost / clicks
Google's calculated fields
Calculated fields are almost identical to Funnel's metrics. Remember how Funnel has two types of metrics? Data Studio has them too, but the difference can be hard to see. This formula, for instance, operates post-aggregation, and has aggregation 'Auto':
SUM(cost) / SUM(clicks)
Aggregation type 'Auto' means that Data Studio will not try to aggregate these values further after the formula is applied, and cannot be changed by the user.
This formula operates on each row of data, and can be aggregated after the formula is applied:
cost * 2
The aggregation type defaults to 'Sum' but can be changed to, for example, 'Average' or 'Max':
The difference in the formula is the SUM(). This tells Data Studio to first add up all the cost values, then all the clicks values, then divide total cost with total clicks.
Without SUM() Data Studio will apply the formula to every row of data.
Note that you can't combine SUM() with non-totalled metrics, for instance
SUM(cost) / clicks
Will result in an error in the Data Studio formula editor.
How Funnel exports formulas to Data Studio
Funnel formula custom metrics are exported to Data Studio, and are available for use just like any other metric. The difference is that Funnel exports the formula, not the values; in the CPC example above, Data Studio will request cost and clicks from Funnel and apply the formula itself.
When Funnel exports formula custom metrics, we want post-aggregation, so Funnel's CPC formula:
cost / clicks
Is in Data Studio:
SUM(cost) / SUM(clicks)

