Introduction

Raging against the (GA4 BigQuery) Machine

Raging against the (GA4 BigQuery) Machine

As an Internet child of the 90s, I still use the back button. And I’ve seen people who use the back button—a lot. Go to the homepage. Go to the About Us page. Hit the back button. Go to the careers page. Look at the job description. Hit the back button. Look at another. Hit the back button. 

That is why I’ve been mourning the loss of Unique Pageviews and Events in Google Analytics 4. Unique pageviews have solved the problem of accidental double-tagging, poorly implemented SPAs, and more. 

And don’t tell me using Sessions is a sufficient substitute. It’s not. 

But wait! Just do it in BigQuery

Ha ha, no. 

Yes, I know how to do it in BigQuery. But that’s me – I’m knee-deep in this stuff every day. I’ve had clients who want to log into GA4 and know how many unique pageviews there were for a specific period. And they can’t. 

The ongoing joke in Measure Slack for almost any question in the GA4 channel is “Do it in BigQuery.” It’s a joke because we shouldn’t be expecting casual GA users to build crap in BQ. And sure, it’s excellent for us data analytics professionals, but let’s be honest – most Google Analytics users (before the forced migration to GA4, that is) are casual users. 

I’m living large in GA4. I love my BigQuery connections. My SQL queries. My CTEs, my Left Joins. I love knowing I can transform almost any GA4 data into anything I want. 

But Google will lose adoption and analytics users when they ask everyday folks to become data engineers. We all know that GA is a drop in the bucket, revenue-wise, for Google. So maybe they don’t care. 

But I have clients who set up BigQuery late in the game. Or their IT security team won’t let them set it up. Or they have 360 and wonder what they are paying for if the interface can’t get them the absolute minimum of what they had before. Or they only had the funds to hire me for the GA4 technical implementation and not to engineer an entire data warehouse for them. “Just do it in BigQuery” doesn’t cut it. 

We don’t need to create total parity with Universal Analytics. But the total loss of ease and basic metrics has made my job as a consultant much harder.