You can use custom events from CRM, call tracking systems, and any other online and offline sources to build a conversion funnel in your attribution model.
This will let you gather and attribute in one funnel all the efforts your business does to get conversions
What data you need for custom events attribution in OWOX BI
To attribute custom events in OWOX BI Attribution, upload the data on these events to Google BigQuery.
For this, you need a table or a view with the following structure:
| Field name | Data type | Description |
|---|---|---|
name |
STRING |
Event name (user action). |
category |
STRING |
Event category. |
time |
TIMESTAMP |
Event time. |
user_id |
STRING |
User ID of the user who initiated the event. |
client_id |
STRING |
Client ID from Google Analytics. |
transaction_id |
STRING |
Transaction identifier. |
revenue |
FLOAT |
Revenue from the event that led to the conversion. |
cost |
FLOAT |
Event cost. |
utm_source |
STRING |
Event source. If the field |
utm_medium |
STRING |
Event medium. If the field |
utm_campaign |
STRING |
Event campaign. |
utm_term |
STRING |
The search term that led to the event. |
utm_content |
STRING |
Content of the link that led to the event. |
geo_region |
STRING |
The region where the event happened. |
The fields highlighted in bold are required to fill in. The utm_term and utm_content fields are optional, so you may add them to the table if you need to consider these UTM tags. The rest of the fields must be present in the schema, but you can leave them empty.
How to upload transaction data to Google BigQuery?
How to create an attribution model based on custom events
1. On the OWOX BI main page, click Create model at the bottom of the Attribution block:
2. The attribution model creation page will appear. Here, provide access to the Google account where you keep your BigQuery project with custom events.
3. As a data source, select Custom events:
4. Select the Session timeout for each event. This setting determines how the events from your table will be assigned to sessions when the attribution model is calculated. You'll be able to change this setting anytime later.
- 24 hours: all events within one day are associated with one session. This method is applied to custom events by default. Use it if it’s not critical for you to evaluate the impact of individual custom events on transactions or other sessions.
- 30 minutes: events are grouped into a new session every 30 minutes. Use this method if there are many custom events in your funnel and you do not want to have a separate session for each of them or combine all of them into one session.
Important: When using this method, the 30-minute timeout is not calculated between the events in the funnel. Instead, a 24-hour day is broken down into 30-minute intervals. Thus, the event that happened at 10:29 and the event that happened at 10:31 will be assigned to different sessions. - New session for each event: a separate session is created for each event. Use this method if you have individual events in your funnel that can affect a transaction in an ongoing online session. For example, a phone call that helped a customer make a purchase decision.
Important: When using this method, it’s possible to get duplicate Session IDs for two or more events that have the exact same combination of UTM tags, region, and timestamp.
5. Next, select the BigQuery project, the dataset, and the table with your custom events data:
Here, note the BigQuery dataset location. It can be one of the two multi-region locations: the EU (European Union) or the US (United States). After selecting the EU as the initial data source for the attribution model, you won't be able to add to this model a source from the US location. And the other way around.
5. Click Create model.
Done. You have created an attribution model based on custom events
How to add custom events data to the existing attribution model
1. On the model's page, click Add source:
2. In the opened window, select the type of data you need: Custom events:
3. Select the Session timeout for each event. This setting determines how the events from your table will be assigned to sessions when the attribution model is calculated. You'll be able to change this setting anytime later.
- 24 hours: all events within one day are associated with one session. This method is applied to custom events by default. Use it if it’s not critical for you to evaluate the impact of individual custom events on transactions or other sessions.
- 30 minutes: events are grouped into a new session every 30 minutes. Use this method if there are many custom events in your funnel, and you do not want to have a separate session for each of them or combine all of them into one session.
Important: When using this method, the 30-minute timeout is not calculated between the events in the funnel. Instead, a 24-hour day is broken down into 30-minute intervals. Thus, the event that happened at 10:29 and the event that happened at 10:31 will be assigned to different sessions. - New session for each event: a separate session is created for each event. Use this method if you have individual events in your funnel that can affect a transaction in an ongoing online session. For example, a phone call that helped a customer make a purchase decision.
Important: When using this method, it’s possible to get duplicate Session IDs for two or more events that have the exact same combination of UTM tags, region, and timestamp.
4. Select the BigQuery project, dataset, and the data table with custom events data. OWOX BI will check the table schema fidelity. If everything is good to go, you'll get the "All data is ok" message.
If there are any issues with the data table schema, OWOX BI will point them out for you.
Note:You can connect only those data tables that are located in the same multi-region location as the rest of the source data for this attribution model. For example, if the data already connected to the attribution model are located in the EU, you won't be able to connect data located in the US, and the other Way around. You also can transfer the BigQuery data between the data set in different locations as discribed in this article.
4. Now click Save. Done. The custom events data is added to your attribution model.
Now you can adjust these events as the funnel steps.
Adjusting funnel steps with custom events
When creating an attribution model based on the custom events, the funnel steps will be created automatically according to the following logic:
- Events for which transaction_id values are known will be added to the conversion step. Limit: up to 3 most frequent conversion steps will be created.
- Events for which transaction_id = NULL will be added as intermediate steps of the funnel. Limit: up to 20 most frequent steps will be created.
You can also set the settings manually using the context menu near the funnel step. For example, to merge multiple events into one step:
1. Open the drop-down menu in the step and click Edit:
2. Here, you can change the step’s name and define which event or events will be counted in this step by adding and removing a Custom event parameter, and define the conditions by which the events are included in the step.
Learn more on how to set up funnel steps.
0 Comments