Analytics Roundup – Updates from July 2024

Recent articles and items of interest, mostly related to the Google analytics stack (GA4, GTM, Looker (Studio), BigQuery, CoLab), from Two Octobers’ Head of Analytics.

Google Chrome reverses course on third-party cookies, again…

It is no real surprise that Google has once again changed their mind  on blocking third-party cookies in Chrome, given that they have previously set and missed several deadlines to do so. 

To some extent, I see this as “whatever” news. Third-party cookies are blocked in Safari and Firefox by default, so they are not very accurate anyway, and Google’s waffling doesn’t change the fact that we should:

• Use GA4 or another reputable analytics platform to measure performance, none of which rely on third-party cookies.
• Own our own data. The writing is on the wall. So much so that we can’t even see the wall anymore.

If you want to understand the nuances of cookies and browsers, I found this article really helpful: How Different Browsers Handle First-Party and Third-Party Cookies

In related news, Google recently announced “first-party mode” for Google tags. This is a feature that allows you to handle every aspect of tracking on your own domain. If your dev or privacy folks have been tightening up your Content Security Policy (CSP), this is something you should be aware of. If you don’t know what I’m talking about, you can probably ignore it for now.

If you do want to understand it better, check out this article by Simo Ahava: First-Party Mode For Google Tags.

Looker Studio soon to release an Excel connector

This is great news for the many organizations that use Microsoft Office, but the big let down for me is that it won’t support Office 365 sheets in the cloud. Looker Studio guru Mehdi Oudjida has tried it out and includes a video walkthrough in his writeup.

BigQuery is becoming a lot more than an analytics data warehouse

Google BigQuery has been releasing new features at an astonishing rate this year. There are far too many to mention (nine in July alone!), but I’ll highlight a few. 

• Insights – this feature is seriously cool. BigQuery uses AI to identify data that might be of interest, and write queries to explore the insight. Below is an example of what it came up with when I ran it on a GA4 dataset.
  
• Table explorer – this feature needs some work. I was thinking I would get descriptive statistics comparable to what’s standard fare for pandas or Excel, but so far it just seems to be value counts. I expect more is coming.

The overall picture that’s shaping up is that Google is moving BigQuery towards being a lot more than a place to store data. They have released several features this year that significantly streamline data exploration and visualization. They also do a lot to reduce friction for people who are just getting started in analytics, or only get to do it part time (like me). I’ll highlight a few of my favorites:

• Data Canvas provides an AI-assisted drag-and-drop workspace to build queries and charts.
• You can also now explore query results in a CoLab notebook with a single click (and you can export a Data Canvas to a notebook).
• And, like every other major coding environment, they added integrated AI code assistance.

GA4 & BigQuery

Speaking of BigQuery, there are several useful additions to the GA4 schema in the BigQuery export.

• session_traffic_source_last_click – this one is jam-packed with Google Ads data and very useful for attributing conversions and revenue to specific campaigns and ad groups. It also works in conjunction with UTM tags, so you can get detailed attribution from Meta, LinkedIn, TikTok, etc. Help on configuring UTM tags.
• Google also added manual_creative_format, manual_marketing_tactic, manual_source_platform fields to the collected_traffic_source record. In contrast to session_traffic_source_last_click, these values do not persist – i.e. you will need to construct your own user or session model to tie them to a conversion or other behavior.

If you are wondering why BigQuery metrics don’t match GA4, Nick Iyengar produced a nice summary:
GA4 and BigQuery: why might data not match?

A very friendly introduction to regular expressions

My number one recommendation for anyone who wants to up their game in analytics is to get comfortable with regular expressions. If they intimidate you, Benjamin Mangold has a great intro:
Using Regular Expressions (Regex) in Google Analytics 4 (GA4)

Content we’ve published

• Five Reasons to Set Up the GA4 BigQuery Export, by yours truly

Articles/videos that made me smarter

• Make Metrics Matter, Kate Minogue
  Explores the human side of how metrics actually work (and don’t work) in an organization. Very insightful and well written.
• How to Challenge Your Own Analysis So Others Won’t, Torsten Walbaum
  This article is longer than it needs to be, but I really agree with the main point and I loved the detailed examples of using ChatGPT to do sanity checks.
• Navigating consent mode in the GA4 BigQuery export, Johan van de Werken
  I had a fuzzy idea of how this worked. Now it’s less fuzzy.