This post is intended to detail the entire set up process of getting your data from GA360 into BigQuery and implementing the Looker Block on top. If followed correctly, set-up shouldn’t take more than 15-30 minutes. I’ve added some detail around the common errors at the bottom, along with troubleshooting and resolution tips
1) Set up the BQ Export:
Follow the directions here. The instructions are very detailed and easy to follow. If you encounter an error, ask the customer to file a support ticket through the GA console and cc you on responses.The Support team is usually pretty responsive. If you don’t receive an answer within 24 hours, please email dillon@looker and firstname.lastname@example.org asap and we’ll escalate for you.
Note that it takes 24 hours for the data to populate in BQ once you enable the report. You’ll get an initial backfill of the last 13 months or 10 billion page “hits” (“hits” are the number of assets that load on a page, not to be confused with an “impression” or “pageview”).
If you hit an error, or are unable to populate text fields, you’re likely running into a permissioning issue. Please make sure that the Looker customer contact has:
(A) OWNER status on the BQ project. This can be found in the IAM panel of the project in console.cloud.google.com
(B) EDIT permissions to the Analytics Property (i.e. the Google Analytics webpage being analyzed)
The below pic doesn’t have the appropriate permissions, but shows where permissions can be granted within the GA console (not to be confused with the GCP Console referenced in part A).
2) Connect BQ to Looker:
Follow the steps in the Looker documentation
3) Implement the Block:
The Block is well architected, thorough, and very easy to implement. As of now (Looker v 4.22), there is no way to programmatically add Blocks to a project, so a copy + paste of the 2 view files, 1 model file, and 3 dashboard files will likely be the easiest. You can find the LookML files in the dedicated github repo. There are two points of customization worth calling out (these are also noted in the README of the Block):
- (A) You’ll need to replace the schema name on line 19 of the “ga_customize” view
- (B) Replace the model names in the dashboards if you named your model anything other than “ga_connectors360” (use the global search + replace functionality)
After that, test the dashboards and explore. You should be good to go!
Troubleshooting Common Errors
BigQuery permissions / Getting a 403 Error - the customer / prospect must go into their console, and ensure that the service account they created in step 2 has BQ User AND BQ Data Editor privelages. Also, be sure that service account has permissions for the correct project.
Once the service account is created, navigate to the IAM panel. Once here, click on “PERMISSIONS” in the top right.
Upon opening the permissions panel, check the box of the service account you just created. MAKE SURE THE SERVICE ACCOUNT HAS EDITOR ACCESS for the project you’re in. If you don’t see “Editor” in the panel below, add the service account using the “add members” option at the top.
Next, navigate to your BQ console and select a dataset. Then “share dataset”, and you’ll get the following screen. Make sure “Project Editors” have “can edit” permissions.