#20 - Query parameters, Global Queries and other BigQuery updates in February

Mar 25, 2026 — Balazs Vajna

Hi Folks,

Although we're approaching the end of March soon, we still have February to cover with regards to BigQuery related news and updates. So let's have a look what all happened in February on the BigQuery front.

Main BigQuery platform updates

You can now use query parameters in your SQL code, marking these parameters with "@", and setting their value in "Query settings". From a SQL development perspective, it doesn't look different to using procedural language and a DECLARE statement, but there are some benefits to this, as described in this LinkedIn post.

Using parameterized queries and setting parameters in Query Settings.

"Global queries" have arrived, i.e. it has become possible to query tables from datasets in different regions within the same query. Previously this had been forbidden (or perhaps just not possible), resulting in having to do various types of workarounds (table copies, read replicas etc) if you needed data from multiple regions.
The Youtube data transfers were sightly improved by now allowing the extraction of reach reports.
The AI.CLASSIFY function now allows multi-level classification i.e. to use multiple categories instead of only binary classification.
The "Insights" generation feature (which could e.g. be used to generate column descriptions in tables) has been extended to datasets, i.e. now you can create "dataset insights" which includes drawing up the relationship between tables within the dataset. Very useful exploration feature, especially when dealing with a dataset for the first time.

Dataset insights & relationship graph on the Google Merchant Center example.

💡

By clicking on the edges (lines) you can reveal the join (or common) fields. You can drag the nodes to tweak the visual representation if you want to optimize it for sharing with others.

The Conversational Analytics feature was slightly improved by introducing the ability to define custom glossary terms to better guide the agent.
a good few features became "generally available", but I won't list them here as they had been covered in past updates when they came out in "preview" form (e.g. undeleting datasets, which is a feature that has been around for a while).

The news came out that Legacy SQL support will be limited (and, eventually, probably retired) in BigQuery, but if you don't particularly know what this means, this will not affect you.

Community posts and updates around BigQuery

An article was published on this very platform about how you can export Campaign Manager 360 data in a much easier way than setting up the Data Transfer for the same data source. Display & Video 360 works very similarly, but I'll endeavour to cover that too in another similar article in the future.

A blog post with a deeper dive on Conversational Analytics in BigQuery (a topic that will continue to be hot for a while, I guess).
How to set a personal query limit (tied to your user profile) to avoid cost peaks even if you don't have access to the BigQuery API quota limits.
Another cost-saving tip leveraging Change History and ingestion timestamps.
A free Medium article describing how we can now define a full-fledged semantic layer in BigQuery, by defining virtual columns, join columns baked into the DDL (and using them with graph dot notations instead of explicit joins), defining measure columns with business logic hardwired into the schema itself etc. It is based on a Google research and lays out a novel way to process and analyze data.
A LinkedIn post by the Superform Labs team on user stitching — a premium GA4Dataform feature, but the post is insightful in its own right.

💡

Premium GA4BigQuery subscribers can still avail of a discount on the GA4Dataform Premium Plan. Reach out to get your coupon code.

Now that we've touched on the Dataform topic above, a long blog article came out describing in detail how to use Dataform in the context of Google Analytics data. While I think by far the easiest approach is to go to GA4Dataform.com and install it from there, this article is a great read nevertheless, to understand the principles of why and how you'd use Dataform if you were to build your own process for it.

another blog article, but this time on the DBT side: "From Raw GA4 Events to AI Insights in Slack — With DBT, BigQuery, and MAKE"
Building knowledge graphs in BigQuery — essentially using BigQuery as a graph database (as it can now support Graph Query Language). They're really shoving everything into BigQuery and it's becoming an all-purpose database. Graph analytics are particularly fascinating, perhaps partially due to the impressive visuals it involves (and looks like it's possible to pull that off too in our favourite tool).
A reminder about Visualization cells in BigQuery Studio Notebooks.
Using query priority models (interactive or batch) to better manage queues, and ultimately, query cost.
A nice AI.SEARCH use case example.
A LinkedIn post illustrating the use of clustering algorithms on GA4 data in BigQuery.

The number of BigQuery related content certainly seems to be increasing. The size of the above list is impressive, especially considering the fact that I was very busy in February and I probably missed half of those.

But let's move on to the other, non-BigQuery-but-still-interesting community updates (that I didn't miss, and which made the cut through my arbitrary filter of what is interesting, from the vast sea of digital analytics updates).

Other relevant news from the community

There are many interesting and useful webinars happening these days. One I'd def call out would be one by Stape at the very end of February, on a topic that is close to my heart. The title was: “Conversion signals & server-side tracking: how digital agencies can maximize ad performance”. It is a very important topic (sending “good” data into the advertising platforms). I'll go into it in more detail in the near future as it warrants a deep-dive. For now, you can just watch it on Youtube on the above link.

Superweek was in February — though I covered it in the previous newsletter, I thought it deserves an honorable mention here too. (Notes are still available btw.)
A Measurement Protocol tester tool, built with some Claude assistance. With a LinkedIn post that describes it in more detail.
An article about referral exclusion vs cross-domain tracking. Always good to keep ourselves anchored into the fundamentals.
A very long and comprehensive article about attribution (requires your email, but you can get it for free). Here is the LinkedIn post accompanying it for a summary / teaser.
An open-source consent management system for mobile apps (an area that's rather overlooked there), brought to us by Analytics Debugger. Released in February (hence the entry here).
Another very strong article bordering on the attribution topic: about the power of post-purchase surveys and how to implement them.

I've also decided to finish each newsletter with some funny stuff I've come across (that is somewhat anchored into the field of analytics of course). Fortunately, the community helps here too, and sometimes this can be simply delivered by pointing to just another community update that happened in the month in question. If you look back in past newsletters, you can find things like borderline provocative videos about MMM or a great Halloween-inspired data engineering horror movie mix.

This time, it's a collection of Tracking & Analytics themed Valentine's Day cards from our favorite Digital Analytics musician fellow. I'm shamelessly putting one here, but you can check out the rest in his post:

That's all — apologies to everyone whose update I missed. (I have marked gaps in my "community notes" spreadsheet in February... I just didn't have the energy to go back and fill it. This became a decent list of updates regardless; I wonder how many of you have actually made it to the end.) As always, if there are any other updates that you think should be added to this list, please mention them in the comments, and let's keep a future reference of all the relevant things that happened in February 2026.

Best regards,
Balazs