Analytics Roundup – Updates from February 2025

Not a lot of updates are happening in the Google analytics stack right now, which has given me time to think a bit more about where marketing analytics is heading. I’ve also found and referenced some really smart people and articles below. If you follow one link in this roundup, it should be Juliana Jackson’s article on this topic.

Personally, I’m most excited about marketing mix modeling (MMM) as the next stage of my analytics journey. I wrote last month about the general release of Google Meridian, and since then I’ve been working on growing my understanding of the topic and related tools and services. These are some of the resources I’m digging into right now:

• Cassandra’s YouTube channel – Cassandra is a paid SaaS app for MMM, but they have a lot of good, free content.
• PyMC learning resources – PyMC-Marketing is an open-source, bayesian python library for MMM and CLV analysis.
• Marketing Analytics With Kisholoy – Kisholoy Mukhopadhyay produces YouTube videos on a variety of marketing analytics topics, including quite a few focused on MMM.

Product updates

• Google Pipe Syntax: Modernizing SQL Without Sacrificing its Strengths, Axel Thevenot
  A new and different way to write BigQuery SQL. It is intended to be easier to read and write, particularly for multi-step queries. I’ve played around with it a bit and I do think it is more intuitive.
• Modern charts in Looker Studio, Google
  An update with various changes to Looker Studio chart features and appearance. Google teased this update last month and I was really excited, given how limited Looker Studio charts are compared to Tableau and other competitors. Turns out my excitement was a bit premature.

Workflow

• How to Migrate Your GA4 BigQuery Export to Another GCP Project, Krisztián Korpa
  If you export GA4 data to BigQuery (and you should), you are likely to want to move your data from one project to another at some point. Bookmark this handy guide for when the time comes. My personal favorite is the native SQL solution.
• Combining Ad Data using Dataform in Google Cloud Platform, Michael Howe-Ely
  Creating a cross-channel view of marketing data doesn’t have to be overly complicated, and this article provides a good starting point, though it leaves a lot of opportunities to leverage the power of Dataform. The first thing I would do is to switch the queries to perform incremental table updates. This article by Prasanna Venkatesan from measurelab summarizes various incremental approaches. I pretty much always use either the ‘Incremental Load with Rolling Delete’ or ‘Incremental Loads with Unique Keys’ methods. The former is more efficient, but the latter is necessary when unique keys can span multiple dates – records keyed to session or user IDs, for example.

Attribution

• MTA vs. MMM: Which marketing attribution model is right for you?, Benjamin Wenner, Search Engine Land
  A comparison of the strengths of multi-touch attribution (MTA) and marketing mix modeling (MMM). I like the hybrid approach he describes towards the end, though I think he overstates the effectiveness of MTA. We can’t really tie the behavior of a single user across multiple platforms, so while I agree that MTA can be tactically useful in a way that MMM is not, I’m not sure that building a custom MTA model is worth the effort for most marketers. 

Ideas

• You’re Not Switching Analytics Tools. You’re Switching How You Work, Juliana Jackson
  This is the smartest thing I’ve come across this month. You should read it. Any additional comments I make will be less than.
• Analytics and the psychological cost of constant observation, me
  Sometimes, the act of measurement affects the behavior of the thing we are measuring. That can result in hidden and not-so-hidden costs.
• From elevator music to causal thinking, Kaiser Fung
  A fun and easy-to-follow explanation of the role ‘interaction effect’ can play in causal analysis. As marketers, don’t we always want one plus one to equal more than two? This article explains how that is not an unreasonable thing to wish for.
• The art of delivering bad results, Amy Esselman, storytelling with data
  Five helpful rules for presenting bad results – IMO they are equally relevant for good results.

Privacy

• Google Consent Mode Fines: Requirements for UK & US, Will Rice, Measureminds
  A good, though fairly boilerplate overview of consent best practices. In the article, they mention their Consent Mode monitor, which audits tags on a site and reports whether they are configured for consent – very cool!