Google Analytics 4 (GA4) - an introduction
Google Analytics significantly changed on July 1, 2023. This article will address some of the most visible changes and serve as an introduction to the following support articles for the new version of analytics, Google Analytics 4 (GA-4). GA4 uses user identifiers to more accurately reflect users in the events that occur on the website. This allows cross device and cross browser de-duplication of users, or what we refer to as visitors.
Primary differences
No more views
In Google Analytics Universal Analytics (UA) we were working with views that the service desk could create and then provide access to for users who were associated with the views. Views were primarily used to separate metrics by agency site collection and individuals in those agencies had access or visibility into only those metrics they were concerned with. Sitewide (oregon.gov) metrics were available but not necessarily of concern to most Google Analytics users.
Views are no longer a part of Google Analytics in GA4. Now all site data is handled at the property level, one step more broad than views. We see this wide view of data at the Dashboard and Reports tabs of GA4.
Explorations
To allow filtering similar to views, GA4 includes Explorations. Explorations can be created to view nearly any metric needed and can be filtered to your site, a subsite or a specific page path.
Explorations are 'owned' by the user that created them, so are not visible to other users by default. The Tyler Oregon E-Government Web Team has created some basic explorations that we are happy to share at your request so that they can be copied and then worked with by individual users. This brings us to a limitation.
GA4 has a hard limit on the number of shared explorations per account and per user. For this reason, when we, or any user shares an exploration, we ask that the shared period is time-boxed fairly tightly so that we do not run into a situation where an agency web team cannot work with shared explorations at all.
There is an article with a video for each of the explorations that the service desk has created as starting points. The article will hit high points and the video will detail what is being looked at and how to map the exploration to a specific site, subsite or path.
There are some suggested starter explorations that have been built and shared so that they can be copied by agency users. The shared explorations cannot be edited, but the copies made by users can so they can be tailored to the agency needs. The articles describing each exploration detail how to tailor them.
Mixed letter cases
Another difference is that in UA, all data was 'forced' to a lowercase state so that each instance of a page path or event was combined into a single line item. In GA4 no such combining function has yet been released. Keep this in mind when considering metrics and making decisions. Explorations that consider page paths might do well to have a tab for all the common case variations of the path. For instance, for Page view metrics in the Department of Administrative Services, I might set the page path to 'Begins with' /DAS/ on one tab, and 'begins with /das/ on the second tab. Between these two, by far most of the results would show up. I can then export each set of metrics and use a formula in a spreadsheet program to combine letter case differences into a single row. Examples of this will be found in the Page views videos.
Date range stickiness
A third significant difference is when a date range is set on an exploration. In UA, the date range of 'The last 30 days' would continually roll as the days progress in the real world. In GA4, On June 30, a date range of 'The last 30 days' would run from June 1 to June 30. When I come back to the exploration on June 10, the date range will still be set to June 1 to June 30. This is true for preset date ranges and custom date ranges in the explorations.
Sampling
Something of note, but not a difference is that GA4 uses sampling of data the same way the UA did. This means that if a tremendous amount of data (in terms of computing power, not line items) exists for a given exploration, GA4 will 'sample' the data which is to take a portion of the results and extrapolate or predict the metrics for the whole range. Google has been doing this for years now and is very good at it so the sampled results may be relied on pretty heavily. However, if exact data is desired, a shorter date range, or a more precise path can reduce the overall number of results and cause Google to either take a much larger sample, or no sampling at all, considering 100% of results for the exploration being viewed. Sampling is indicated by the small triangle icon at the top right of the exploration screen. 100% data consideration is indicated by a green circle with a check mark in it in the same place.
Other articles of interest
- Google Analytics 4 Introduction
- Page views - descending by views
- Document downloads by document
- Document downloads by page
- Helpful page polls - votes
- Helpful page polls - feedback
- more will come available.
Have questions?
You can:
- Search the Knowledge Base for what you are looking for using the search feature at the top of this page.
- You can submit a question to the Oregon E-Government Service Desk
Comments
0 comments
Article is closed for comments.