#16 - New videos on GA4BigQuery, and lots of community updates in September

Hi there,

First, a warm welcome to all new subscribers who joined since the last newsletter.

Lots of updates to cover! A few announcements first, and then we'll dig into the plethora of updates. While BigQuery didn't have "overwhelming" updates, the community was very active.

MeasureSummit: 8th-10th October

First and foremost, there is only one day until MeasureSummit starts, with many interesting presentations on various topics, e.g. mobile app analytics, server-side tagging, BigQuery, paid advertising and data privacy (to name only a few). You can sign up for free, which allows you to watch the sessions live and also on-demand afterwards, while the event lasts—and they usually extend it to the weekend too.

It's a cramped agenda and with all those on-demand ("second stage") sessions, it's practically impossible to watch all the presentations of interest. The All-Access Pass lets you retain access to the videos, also allowing you to download the slides, the transcripts and the videos themselves.

Other announcements

A few other changes went live recently on the GA4BigQuery website.

• A new page was published titled "Videos on BigQuery" which collects previously broadcasted material on video format. For example, my MeasureSummit talk this year ("Creating a Data Warehouse in BigQuery") will heavily build on my previous talk ("The Power of Native Data Transfers in BigQuery"), which is available on this page (and also on Youtube).
• Johan and the Simmer team (Simo and Mari Ahava) released a revamped version of the GA4 BigQuery training course. We've updated the course page accordingly (and I've renamed it to "SQL course"). The Launch discount has ended, but Premium GA4BigQuery subscribers can continue to avail of a 15% discount.
• A new (free) article was released under the "GA4 | tutorials" section, diving into cohort analysis using GA4 data in BigQuery.

Updates on the BigQuery platform

While August came with many UI changes, September updates are more infrastructural and "backend-y" in nature.

• BigQuery Advanced Runtime will be the default of new projects, which will make queries faster.
• Soft failover is generally available for managed disaster recovery. Unlike hard failover, this method ensures complete data consistency before switching regions (source).
• SQL code completion is now available for BigQuery projects. Based on Gemini. Looks like it is free to use, but I haven't tested it yet.
• Gemini CLI: you can explore your data by asking natural language questions in the command line interface.

There are some frontend-y updates as well:

• a Saved Query can now be opened in Google Sheets. It's not a life-altering feature, but can save time if you need to set up Connected Sheets (i.e. when you want to explore and process a BigQuery table in Google Sheets).

• The "Visualization" tab in Query Results allow visualizing geographic results. While this probably won't be used much in digital analytics per se, it's worth being aware of the data viz capabilities becoming more and more powerful in BigQuery itself.

• Another frontend related "heads-up": there are rumors of the Job History starting to disappear for some people (as part of an AB test or experiment; originally it's usually a bar on the bottom).

GA4Dataform Premium release

You might already be familiar with GA4Dataform which jump-starts analytics projects by unnesting and preprocessing GA4 data. The Premium version offers a lot more capabilities like intraday support, cross-session user_id stitching, cost monitoring, and many more. Most of you have received coupons with various discount levels for the premium version.

BigQuery related content in September from the community

• Although we've published a cohort analysis article here (based on a LinkedIn article by Enio Kurtesi), here is another one on customer retention, also with code and detailed walkthrough.
• A piece of code to determine which Looker Studio reports are hitting your BigQuery tables. Accompanying LinkedIn post here.
• A post about the best way to count returning users in GA4 BigQuery data.
• An analysis of lost keywords using Search Console data, by Marco Giordano. Also, how to calculate the contribution of anonymized queries to clicks, or create a breakdown of pages in terms of their overall traffic contribution.
• The smart new trick of WITH expression to reuse calculated columns in the same query, not having to write CTE.
• Dataform now supports schema changes.
• Code to compare On-Demand and Editions pricing on your GCP jobs.
• Attibution modelling approach for other columns rather than traffic source (e.g. which product brought in the customer).
• Comparing different table backup types (clone, copy and snapshot) from BigQuery cost perspective.
• A fail-safe functionality in BigQuery that is worth being aware of: an additional last resort for saving your data from oblivion when you have passed the Time Travel duration, but you're desperate to recover it nonetheless. Same author (Matt Dixon) posted about the rich selection of job monitoring visualizations in BigQuery.
• Using n8n to get aggregated GA4 data into BigQuery, without the complexity of building it yourself or paying for expensive tools.
• A short study about sessions misrepresented as "Direct" traffic, despite the GCLID being present in the URL. An example where looking at GA4 data in BigQuery makes it possible to spot the problem, while it would remain hidden in GA4 itself. Note that the behavior shown in the article is also partially due to the session_traffic_source_last_click elements not updating with parameters received mid-session, as discussed in this article.
• A tutorial on using AI.GENERATE functions.
• Event parameter data checker tool to accelerate GA4 data development in BigQuery.
• Developing a semantic search capability on BigQuery data.
• If you'd like to dip your toes into data science, Google has published a short ebook with data science use cases.

Other relevant blog posts from the community

• GA4 enabled the use of ecommerce and item related dimensions as secondary dimensions across all reports.
• Several new tag templates approved in GTM. Also a new server-side GTM template to generate constant values with different data types.
• Treemaps now available in Looker Studio. More tips by experts.
• There was a "GCLID scare" around the new Apple iOS 26 version, which was later debunked by Simo Ahava and a few other people.
• Podcast on MCP servers in digital analytics.
• A free, Google Sheets Add-on called Addocu that automates documentation by connecting to key Google platforms such as GA4, GTM, and Looker Studio to create a centralized, accurate view of your setup in seconds, thanks to the power of metadata.
• Website tracking tracker by Stape.
• Tutorial to get your subdomains to show up as separate traffic sources - by Julius Fedorovicius.
• Event parameters for Google Ads and Floodlight tags—by Simo Ahava and Google documentation; replacing a workaround ("Improved Google Marketing Tags using gtag") earlier developed by Giovanni Ortolani Barbosa.
• GTM Template to automatically map and transforms GA4 e-commerce parameters to the correct format for Meta, ensuring consistent and accurate data flow.
• Market Basket analysis using Google Sheets Apps script.

Some fun stuff

• Google uses Google Tag Manager to track user activity on the Google Cloud Platform.
• Not "fun" as such, but a nice and useful semi-phylosophical article by Timo Dechau about the end of digital analytics.
• Another "deep" article by Juliana Jackson about the importance of data activation in her newsletter.

That's it for now. Thanks for reading and happy querying—and if I missed any key updates from September, feel free to add them in the comments!

Best regards,
Balazs

1,509 words