When building a calculated metric, you can specify the metric type and the attribution model.
To specify the metric type when building a calculated metric:
Select the gear icon next to the metric whose type you want to select.
Choose from the following options:
|Standard||These metrics are the same metrics used in standard Analytics reporting. If a formula consisted of a single standard metric, it displays identical data to its non-calculated-metric counterpart. Standard metrics are useful for creating calculated metrics specific to each individual line item. For example, [Orders] / [Visits] takes orders for that specific line item and divides it by the number of visits for that specific line item.|
|Grand total||Use Grand total for the reporting period in every line item. If a formula consisted of a single Grand total metric, it displays the same total number on every line item. Grand total metrics are useful for creating calculated metrics that compare against site total data. For example, [Orders] / [Total Visits] shows the proportion of orders against ALL visits to your site, not just the visits to the specific line item.|
Attribution IQ is how allocation models in calculated metrics are evaluated.
For a full list of non-default attribution models and lookback windows supported, see Attribution models and lookback windows.
The following example illustrates how calculated metrics with linear allocations work in reporting:
|Hit 1||Hit 2||Hit 3||Hit 4||Hit 5||Hit 6||Hit 7|
|Data Sent In||PROMO A||-||PROMO A||PROMO B||-||PROMO C||$10|
|Last Touch eVar||PROMO A||PROMO A||PROMO A||PROMO B||PROMO B||PROMO C||$10|
|First Touch eVar||PROMO A||PROMO A||PROMO A||PROMO A||PROMO A||PROMO A||$10|
|Example prop||PROMO A||-||PROMO A||PROMO B||-||PROMO C||$10|
In this example, the values A, B, and C were sent into a variable on hits 1, 3, 4, and 6 before a $10 purchase was made on hit 7. In the second row, those values persist across hits on a last touch visit basis. The third row illustrates a first-touch visit persistence. Finally, the last row illustrates how data would be recorded for a prop which does not have persistence.
There are some differences in how linear attribution works between these two tools: