Clear History

Using this package for non-value based conversions

This package can also be used for non-value based conversions, however, the outputs may look somewhat strange.

For example, if you have a column in your events/conversions table with a 1 for a form submission, and a 0 for no submission, you can specify this field in the conversion_value macro (e.g. form_submitted) and add a clause in the conversion_clause macro (e.g. form_submitted = 1).

You can then run the attribution analysis using this column instead of a revenue column. An example output of the report table would look like this:

conversion_window_start_dateconversion_window_end_datechannelconversionsrevenuespendroas
2022-06-032022-08-01Direct699.8699.810000.00.06997
2022-06-032022-08-01Organic_Search269.5269.510000.00.02695
2022-06-032022-08-01Paid_Search_Other50.450.410000.00.00504
2022-06-032022-08-01Display_Other21.321.310000.00.00213
2022-06-032022-08-01Referral12.412.410000.00.00123
2022-06-032022-08-01Unmatched_Channel4.084.0810000.00.00040
2022-06-032022-08-01Video1.51.510000.00.00014

Note that ‘revenue’ in this table is now being calculated based on the form_submitted column in the events/conversions table, where a conversion will only ever have a value of 1. This is why the values for conversions and revenue are identical.

You will notice that the values for ROAS in this example are extremely small. In a value-based calculation of ROAS, one conversion is generally associated with a value much greater than 1 (e.g. one conversion can have a purchase value of $100, etc.). When we calculate ROAS using one conversion equal to a value of 1, the ROAS figures can look very small depending on the number of conversions and the amount spent per channel.