In the field picker in the data warehouse export section we do not list formula metrics unless you search for them. Then you will notice that they have a ‘blocked’ sign which indicate that you can’t export this metric. If you hover over the symbol or click on the warning triangle you will get to know why and we’ll also explain the reason for this here.
Formula vs Rules based metrics
When setting up custom metric you get to select if you want it to be a formula or Rules metric, see picture below.
There are some important differences between formula metrics and rule metrics that are key to understanding why you are not able to export the formula metrics. In short, this depends on how these metrics are calculated and the format in which your data is exported to a data warehouse.
The value of a rule metric is calculated for each row of raw data in Funnel, prior to aggregating the result and is usually adding (+) or multiplying (*) two metrics together.
The value of a formula metric is calculated after aggregating the underlying data and is usually a percentage (%) or another metric using division (/).
Why formula metrics can't be exported
When your data is exported from Funnel we don't guarantee that it will be aggregated, see How is data aggregated in Data Warehouse exports?.
This means that your exported dataset is likely to contain one row for each row of raw data in Funnel, regardless of which fields you select in your export.
Formula metrics can therefore not be exported since it makes no sense to calculate them for each row of raw data. Additionally the formula metrics usually can't be summarised (if they are a percentage for example) and will only be correct for that row and not when grouping the data later on.
Rule metrics can be safely exported since the data don't need to be aggregated beforehand.
A common use case for formula metrics is calculating rates, percentages and ratios. If you are looking to do this with the data in your data warehouse, perform the calculation in your data warehouse or visualisation tool after aggregating the data (by grouping it per day, week, channel, etc). Preferably this is done as late as possible in the pipeline in case you add more steps in the process in the future. This also ensures that you don't have to send calculated metrics anywhere (which might be treated incorrectly in whatever tool/service is used) and it's easier to make adjustments to.