I have to start this post by giving everyone a full confession. I have been a unapologetic and enthusiastic Google Marketing Platform lover since I first started using Google Analytics in 2009. I even wrote a book about Google BigQuery back in 2017. I remembered how much flexibility was at my fingertips and how the GA tag structure allowed me to measure sites and features like never before. Add to the fact that tool was free for a number of organizations and it was a slam dunk to pick GA as a solution. This is one of the main reason why GA is installed on nearly 53% of websites (per W3techs estimates on 10M+ websites). In 2012 Google unveiled Universal Analytics. While this update removed some beloved features, the additions definitely outweighed the losses in terms of extending reporting for new devices mainly mobile tracking.
Google Analytics 4
Cut to 2020 and Google introduced Google Analytics 4 including a brand new interface and tagging structure. While most people were expecting new and exciting features taking into account all of the changes in the internet landscape over the last 8 years, what we got was a bit…different. I don’t want to harp on everything I dislike and like about the new tool (the former would possibly reach book rather than post link), I’d like to outline the changes that most excite and make me want to tear hair out. Let’s start with the dislikes. Disclaimer: While I have been contemplating making this post for a while, this post is largely influenced by Simo Ahava’s “Dear Google Analytics 4” post on his blog simoahava.com. If you are interested in marketing analytics, ecommerce tracking, GA or GTM, I strongly suggest you read his stuff.
Dislike #1: 14-Month Reporting Cutoff
While Google claim that data will be warehoused in GA4 for 24 months prior to roll-off (pretty damning limitation in itself), I have not found a report in GA4 that allows for lookback further than 14 months (non GA360 properties). I’ve been asked in about every role I’ve had as analyst about the effect of seasonality on traffic, campaign performance, transactions, etc. This change makes answering that question much harder as I would not have enough data to make a strong determination save for using BigQuery or the Data API (possible added cost, access and expertise needed).
Like #1: The New “Counting Method”
UA included metrics such as goals, events and custom metrics. Custom metrics were not widely used, so no need to go too deep there. Goals and events were used by pretty much all organizations creating GA reports, but the logic was a little flawed in my opinion. Goals were a success metric to signify completion of a task that was desired by the organization and stems from Google’s original implementation of funnel tracking (like cart completion). Goals could only be counted once in a session even if they happened multiple times and were desired to be counted multiple times. Users were also able to specify steps in the goal process where the end of the process would be considered a conversion (like view product, add to cart, checkout, etc.). This reporting was not close to robust enough to compare to offerings from competitors which is why, I imagine, Google released Ecommerce Tracking. This is different than how events were meant to be used as events are meant to track behavior that isn’t necessarily considered a success metric (clicks on a section of a page, scrolling a page, etc.). Events were counted both once in session (unique event metric) or multiple times (total event metric). Events also allowed for 3 levels of categorization. This all made them more robust in my opinion and I’ve leaned on events to track success metrics when I’ve run out of the 15 set goals afforded in a free GA view.
The new counting method allows users to create events with up to 25 parameters. Users can also configure conversion events to be either “once per event” or “once per session”. This in essence makes goals obsolete adding a layer of standardization not in UA. This is closer to what I found using Adobe Analytics and I think its a good addition.
Dislike #2: Sparsity of Canned Reports
UA included a large number of canned reports for users to dive into their data off the bat. Not to mention a full suite of very useful Ecommerce reports. Some of the reports were probably not well used (Publisher, Custom Variables), but some reports and metrics were probably heavily used (ecommerce metrics, upcoming sunset of 4 attribution models). Many of these reports/metrics can still be created using the Explore tool, however, with some caveats, extra effort and a deep understanding of Google Tag Manager. Some of the metrics are simply no longer calculated out of the box and must be either calculated with custom metrics or calculated outside of GA4. Product detail-to-cart conversion and detail-to-buy conversion rates are 2 metrics that I’ve personally used heavily in product analytics roles in the past. That’s gone. As mentioned, some of the attribution models are also on their way out.
Like #2: BigQuery is Free! Well, Kinda.
The free version of Google Analytics 4 includes the Google Analytics BigQuery integration at only the cost of processing and storage. For UA, only paid versions of GA360 included this perk and those versions can start at prices as high as $150k a year. I’ve been in a previous role where this would have been greatly appreciated and thoroughly utilized. This is an interesting addition as I am of the opinion that organizations that utilize the BigQuery integration will be the organizations best equipped for the UA to GA4 switch. This might be considered an olive branch for some of those smaller organizations that can’t foot the GA360 bill.
Dislike #3: The Collection Limits
While UA had its share of limits, some of the limits in GA4 seem, unfortunately, much easier to hit. The issue I have is with the 500 events per day per user limit. If you make use of the Enhanced Event Measurement functionality which includes events like Scrolling, Video Events and Forms, you might end up going over that limit for some of your users. For some organizations, it might not be worth tracking Enhanced Events in the first place.
Like #3: Google’s Drive toward Privacy and Compliance
Over the years, marketers have taken more and more liberty to do some nefarious things with both analytics tagging and collected data from users. This was the reasoning for both governments and browser creators to crack down with extra privacy laws and features that can’t be circumvented. With that said, Google seems to be pushing the envelope with features such as Server-Side Tagging and First Party Tagging. In my opinion, Server-Side Tagging and First Party are great developments for both marketers and analysts as it helps to comply with the tighter regulations that are being set and they have the ability to improve data quality. Improved data quality is paramount.
Dislike #4: Strange and (Possibly) Broken Functionality
I’ve noticed some really strange behavior in GA4 in comparison to UA. Here’s some examples.
Thresholding and Cardinality: This is nothing new, especially if you were using the free version of UA as you could hit sampling thresholds, however, moving to GA360 seemed to remedy some of these thresholds. With GA4, I don’t see any specific literature mentioning that thresholds would be removed for paid customers. As well, many of the thresholds seem to be tied into Google Signals. Why would I turn Signals on if it could lead to missing data? Also, reports in GA4 have cardinality limits. If you’ve ever seen the term “(other)” in your reporting, that is this limit. Google doesn’t explicitly define what those limits are, so you could be tagging a site without the knowledge of whether or not you will run into this issue.
Explore Reports: To me, this reporting is just a little too barebones for me. In my time using Adobe Analytics, I felt like their version of custom reporting (Analysis Workspace) allowed you to do one thing I’ve wanted to do in Google for years; create calculated metrics on the fly. I want to be able to calculate Event Session Conversion rate. Also, with the loss of so many metrics for product reports, I want to create Item Detail View to Cart Conversion rate or to Buy. I want to create ROI calculations that are much more robust than what is given in the interface, but…I still can’t directly in the GA interface.
Severely Delayed Data: This might not be an issue for GA360 users (I sincerely hope it isn’t), but for the free version of GA4, there have been reports, not just from me, but from many individuals about data populating 24 – 48 hours late in GA4. I have UA tagged in parallel with GA4 and my same day reporting populates much faster in UA. I hope this is just a product of so many organizations updating their tagging at the same time and simply overloading the system.
Well, I think I’ve complained and praised enough. If you want to learn more about what books and courses get wrong about teaching analytics, please click the following link!