Report on a data-driven attribution model

To get an overview of performance reported by a data-driven attribution (DDA) model, create a Floodlight column that applies the model to your conversion data. Then compare the number of conversions reported by the model with a Floodlight column that applies standard last-click attribution (or other attribution models).

Here are the basic steps for DDA reporting (the details are available later in this article):

1. Create a Floodlight column that applies the model to your conversion data.
2. Create another Floodlight column that applies last-click attribution to your conversion data.
3. Optionally create a formula column that shows the percent difference between the model and last-click attribution.

4. If your model uses custom channel groupings, add the columns to a labels report to see how each channel grouping as a whole contributes to conversions. You may also want to report on specific campaigns, ad groups, keywords or product groups to see if the model reveals significantly under- or overvalued items.

   If your model uses automated channel groupings, add the columns to a campaigns, ad groups, or keywords report to how each campaign, ad group, and so on contributes to conversions.

Create a Floodlight column to report on the DDA model

1. Navigate to the advertiser that contains the DDA model, then click any tab.

   Steps for navigating to an advertiser

   1. Click the navigation bar to display navigation options.

   2. In the "Agency" list, click on the agency that contains the advertiser. You can search for an agency by name or scroll through the list.

   3. In the "Advertiser" list, click on the advertiser.

   4. Click Apply or press the Enter key.

   Search Ads 360 displays the advertiser page, which contains data for all of the advertiser’s engine accounts.

2. Click the Columns button in the toolbar above the performance summary graph.

3. In the Available columns list, click ▸Custom conversions, then click Floodlight.

4. Click + New column.

5. Name your column and select Floodlight activities.

   You can include all Floodlight activities in the advertiser, or limit the column to specific activities. Even though the DDA model may be created with a limited set of Floodlight activities, the model projects the patterns it learns from observing those activities onto all activities in the advertiser. The more similar activities are to the activities in the model, the more accurate the model's projection will be.

   For example, if the model learns from observing activities leading up to a shopping cart, then the model's projections for other activities that lead to other shopping carts are likely to be accurate.

6. Use the Metric list to select the metric you want to report on.

   The type of metric needs to match the type of Floodlight activities you selected. For example, if you select Action Floodlight activities, you need to select the Actions, Cost/action, Action conv % or other action-related metrics. If you select a transaction-related metric, the column will display 0.

7. From the Attribution model list, select the DDA model.
   If you see (Calculating) next to the model name, the model is still in its initial learning period. You'll need to wait for the learning to finish before you can select the model. Learn how to view the model's status. 

   • The Cross-environment conversions check box is selected by default in new columns to give credit to the devices used.
     Recommendation: leave the check box selected.
   • The column does not include cross-environment conversions that occurred before you created the model. Learn more about including cross-environment conversions in DDA reports.

8. Click Save.

Create a Floodlight column to report last-click attribution

Follow the steps above, but select Last click attribution instead of a DDA model.

Important: Specify the exact same set of Floodlight activities, choose the same metric, and only include cross-environment conversions if you did in the steps above. This column needs the same settings as the DDA column to accurately compare conversions with the DDA model.

Create a formula column to compare values

To show the percent difference in conversions reported by the DDA model and last-click attribution:

1. If you haven't already done so, create a Floodlight column that shows the number of conversions calculated by the model, and another Floodlight column that shows the number of conversions calculated by last-click attribution.

2. Navigate to the advertiser that contains the DDA model, then click any tab.

   Steps for navigating to an advertiser

   1. Click the navigation bar to display navigation options.

   2. In the "Agency" list, click on the agency that contains the advertiser. You can search for an agency by name or scroll through the list.

   3. In the "Advertiser" list, click on the advertiser.

   4. Click Apply or press the Enter key.

   Search Ads 360 displays the advertiser page, which contains data for all of the advertiser’s engine accounts.

3. Click the Columns button in the toolbar above the performance summary graph.

4. In the Available columns list, click Formula columns.

5. Click + New column.

6. Enter the following equation:
   To_percent(( c:"DDA column" / c:"Last-click column") -1)
   Where "DDA column" and "Last-click column" are the names of the Floodlight columns you created in step 1.

7. Click Save.

Add the columns to a labels report

If your model uses custom custom channel groupings, add Floodlight columns to a labels report to see how each channel grouping as a whole contributes to conversions:

1. Navigate to the advertiser that contains the DDA model, then click any tab.

   Steps for navigating to an advertiser

   1. Click the navigation bar to display navigation options.

   2. In the "Agency" list, click on the agency that contains the advertiser. You can search for an agency by name or scroll through the list.

   3. In the "Advertiser" list, click on the advertiser.

   4. Click Apply or press the Enter key.

   Search Ads 360 displays the advertiser page, which contains data for all of the advertiser’s engine accounts.

2. In the left navigation panel, click Labels.

3. Use filters to show only the labels you're using as custom channel groupings in your model.

   For example, if your DDA model contains the "Generic", "Brand" and "Brand + Product" channel groupings, create the following filter:
   Label equals Generic | Brand | Brand + Product

   Don't remember which channel groupings are in the model?

   1. Click Attribution in the left navigation panel.
   2. Select the check box next to the DDA model.
   3. Click View data-driven model details in the toolbar above the reporting table.
4. Add the Floodlight columns and formula column you created in the previous steps.
   When you're done, you should see something like this:

   
   
   You can also create Floodlight columns that report cost per action (instead of just the number of actions). For example:
   

5. Optionally adjust the report's time range.
   Be sure to choose a date no earlier than 60 days before you created the model. The column can only report conversions as far back as 60 days before the model's creation date.
   
   If you include today's date, you may notice a discrepancy between the data in a column with a DDA model applied and a column using the last-click attribution model. The data for today's date isn't included in a DDA model column. For example, if the time range for a report is the last 7 days, and includes today's data, the DDA model column might display 2500 actions. A column using the last-click model might display 3000 actions because 500 of the actions occurred on today's date.
6. Optionally save this reporting view so you can return to it later.

Add the columns to a campaigns, ad groups, or keywords report

To see how the DDA model reports on specific campaigns, ad groups, or keywords:

1. Navigate to the advertiser that contains the DDA model, then click any tab. Optionally scope into a specific engine account, campaign, or ad group.
   For example:

   • To see the model's attribution at the campaign level, scope to the advertiser or an engine account and click the Campaigns tab.

   • To see the model's attribution for keywords in a specific ad group, navigate to the ad group and click the Keywords tab.

2. Add the Floodlight columns and formula column you created in the previous steps.

3. Optionally adjust the report's time range.
   Be sure to choose a date no earlier than 60 days before you created the model. The column can only report conversions as far back as 60 days before the model's creation date.

4. To see reporting data for each Floodlight activity in the model, segment the report by Floodlight activity.