GA4: 3+ Years Later
Cavaliero, D.
Ah yes, GA4 -- the analytics community favorite meme/beating child & software responsible for 50% of all male digital marketers male pattern baldness the past 5 years.
Rewind ~ 36 months ago
Universal Analytics was shut down. Most companies/organizations still hadn't read the 23742131 emails/notifications or banner messages Google put in front of their faces about the fact their data was going to go away. This wasn't Google's fault -- people were warned...
Except...
Implementors were still lacking the customization and JS API level controls needed to make the transition easy.
The Google provided GTM GA4 tag templates were clunky. The platforms reporting capabilities were (largely) unhelpful or non-existent. Features like UA's Custom Task and a sensible way to push data to N+1 properties still hadn't been addressed.
To make matters worse, Google shifted the entire GA paradigm to an event based system -- which required a certain level of analytics "know-how" to implement in a way that wasn't absolute chaos.
However, Google's primary analytics users were less "analytics engineer", and more "digital marketer flying by the seat of their pants". Google effectively built a Saturn V 🚀 with a half written instruction manual, for an audience who (for the most part) just received their drivers license.
As you could imagine (hell you may have also experienced it yourself) the transition was... not exactly smooth.
So, how is it now?
Honestly, a LOT better. It still has its warts -- the UI still has some problems, but reporting capabilities have improved from some much needed expansion of filtering capabilities (e.g. filtering by Event Name and the inclusion of "some" enhanced measurement event dimensions).
Major Nitpicks
Reporting
- Not allowing users to easily copy values via
Ctrl + C/Cmd + Cfrom the standard/explore report table cells. You have to right click and "copy" from the context menu. (why?!) - Not allowing users to copy values via ANY means from explore report table cells is... a choice.
- Not allowing users to resize explore report columns. (why?!)
- Making explore reports "read-only" (especially the date range) when shared in a property must have been thought up by a literal madman. (I will accept no other explanation).
- Truncating the values in explore report columns and making certain dimensions (like
Page Location) useless as a result. - Enhanced measurement event dimensions are available sporadically in some dimension pick lists, but missing from others. Standard dimension support and consistency needs to be prioritized.
Implementation
- Lack of support for
gtaggroups and routing capabilities to give implementors the ability to mirror data to other properties without cloning every tag in a GTM container. (The capabilities EXIST and WORK if you usegtagoutside of GTM, they just made it inaccessible in GTM for some reason). - Lack of ability to add more than 1
Event Settingsvariable to a givenEventtag makes it less useful than it could be. - No custom task feature (or equivalent) is still a big 👎.
Minor Nitpicks
Reporting
- Custom reports added to the aggregate report library will 🪄magically✨ lose their set report descriptions after any minor edit to the report unrelated to the description.
Configuration
- Google's own "recommended" events lack consistency in how they are named. Some are
object_actionothers areaction_objectsome are justaction(likeclick-- which should have been namedoutbound_clickin the first place). - Allowing event names into the dataset that don't pass Google's own validation when creating an event via the "create event" feature is... a choice.
- Requiring key events to be toggled on/off at the
eventlevel and not a special custom dimension that is able to be set at collection manually.
You'll see a common trend amongst the above -- most of it is UI/UX related. Google seems more concerned about forcing the use of Material UI design paradigms than making the actual product useful/useable in some cases. Some of it is just a pure lack of accessibility best practices (ability to copy via keyboard shortcuts). In certain cases, its just a perfect storm of "looking good -- yet feeling bad" which is exactly the opposite of what any UX professional would be aiming for.
To give Google credit, it has gotten much better. However, given Google's size and resources, I would expect improvements to be made at a more rapid pace (e.g. it shouldn't take 3+ years to fix an accessibility issue).
Is it worth using?
Yes it's a solid tool/platform.
However... 👇
Don't short change the time/effort needed to implement it properly. Many companies think GA4 is "bad" because they rushed through the implementation process and they expected reports to be handed to them on a silver platter by the tool.
GA4 provides a lot of flexibility in its data-model. That flexibility does come with some downside -- Google isn't handing you pre-built reports like in the UA days - you need to take your building blocks and build your own reports to make it worthwhile.
Furthermore, any measurement plan worth a damn (regardless of the platform) requires planning and a proper skillset to implement/maintain. Marketing data collection isn't an "exact" science, but it isn't something you can toss to an entry level marketing person and expect to get very far either.