Get rid of PII
There are currently 2 methods that you can use to delete unwanted data such as Personally Identifiable Information (PII) in Google Analytics:
1. Data Deletion Requests
In GA Property admin, you should see this option:
To create a data deletion request:
You specify (a) the GA Property ID, (b) the Start Date of deletion, (c) the End Date of deletion, and (d) one or more dimensions that you want deleted.
Currently you can specify these individual GA dimensions:
- All dimensions
- URL
- Page Title
- Event Label
- Event Category
- Event Action
- Custom Dimension (all)
- User Id
Notes:
- This deletion method is date-centric.
- If you think of GA data like a spreadsheet, no rows are actually being deleted. Instead, the selected dimensions will have their value changed to - (dash character). Which means you may see inflation in some metrics such as when looking at the Pageviews for URL -.
- When you select Custom Dimension to delete, ALL Custom Dimensions will be deleted. You can’t select individual Custom Dimension indexes to be deleted.
- Previously aggregated reports are NOT affected by this deletion.
Reference: https://support.google.com/analytics/answer/9450800
2. User Deletion API
This method deletes GA users and all their data. You can write Google API scripts to delete multiple users or you can delete a single user via the GA User Explorer report:
Notes:
- This deletion method is user-centric. You cannot specify the date range of deletion.
- All specified user data will be deleted. In the spreadsheet analogy, all associated rows are deleted. Which means you may see deflation in some metrics.
- Previously aggregated reports are NOT affected by this deletion.
Reference: https://developers.google.com/analytics/devguides/config/userdeletion/v3
Summary
When deleting data in Google Analytics, you have to decide if date-centric or user-centric will impact your data the least while allowing you to remove unwanted PII data. Every use case is different.
Generally, the date-centric Data Deletion Request method should have the least impact while not requiring much technical expertise.
Sparkline aims to provide data accuracy, comprehension and consolidation, and most importantly, tangible insights for businesses. Get in touch if you’d like to learn more.
