There is only one year left before Universal Analytics goes away. So what is going to happen to our data? What is the best migration plan?
There is no doubt that these burning questions have been raised by everyone who is using this tool. With Google deciding to sunset Universal Analytics, many people are wondering what the next step is.
🚨 Note: We’ve recently ran a survey in which 120 users shared their opinions on GA4.
Master the basics with our FREE GA4 Course for Beginners
This decision definitely caused a stir in the digital marketing community. Doubts have arisen over the full-time usage of GA4 as well as its viability.
You might also be asking yourself if this is a good solution and why you have read so many negative comments about it.
🚨 Note: If you need to make your quick analytics health check, have a look at our GA4 Audit guide.
We heard you and decided to address this dramatic change by talking to industry professionals and asking their opinion on it.
In particular, we asked 14 experts how they are currently using Google Analytics 4 and what are some of the biggest advantages and pitfalls they have found so far.
Should you switch to Google Analytics 4? Read on and find out!
So we have asked our experts the following questions:
Are you already using GA4 (with clients) as the main tool? What are the biggest advantages and pitfalls you have found so far?
Undoubtedly, migration is a complex process and should be planned carefully. Therefore, dual tagging is still a common trend.
Hussain Mehmood from MarketLytics: “The main benefits have been a flexible data model, advanced funnel reports and exports to BigQuery”
I’ve been using GA4 as the primary tool for only a few clients so far.
The main benefits have been a flexible data model, advanced funnel reports, and exports to BigQuery.
Pitfalls are primarily around confusion about dimensions and reporting interface that need to be better understood before being useful.
Dana DiTomaso from Kick Point: “We are already using GA4 with clients but not always as the main tool depending on a few factors”
We are already using GA4 with clients but not always as the main tool depending on a few factors.
If they are brand new and don’t have overly complex analytics requirements, then GA4 is their primary analytics source, with reports in Google Data Studio.
If they have years of data in GA Universal, we are typically still using that for primary analysis, but the flexibility of events in GA4 means that sometimes we have a hybrid approach.
For example, if we want to capture more parameters in an event than the Universal category/action/label hierarchy would allow for, we’ll use GA4 to report on that event specifically.
That being said, all data is still presented in Google Data Studio to clients, so they don’t need to be flipping between two versions of GA.
Yehoshua Coren from Analytics Ninja: ”We aren’t using the interface at all, just BigQuery”
For the most part, no, we aren’t using GA4 with clients as the main tool. We have one client for whom GA4 is the primary tool for data collection, but we aren’t using the interface at all, just BigQuery.
Some of the reporting capabilities within Explorations are pretty good, but the product isn’t ready yet to replace the very comfortable reporting that most clients are used to in Universal Analytics.
Advantages:
- BigQuery. Though the 1M events per day limit for non-GA360 clients is pretty limiting
- Funnel Exploration is done nicely
- Ability to have both App and Web data modelled consistently in a single location is a benefit
- More robust segmentation engine compared to Universal Analytics
Pitfalls:
- Reporting interface crashes pretty often (for larger data sets)
- Identity resolution is sub-par (no visitor stitching, logged in users ‘change’ primary IDs mid-session)
- Differences between attribution and sessionization compared to Universal Analytics make reporting differences very confusing to end users
- Standard reports are much less useful than Universal Analytics’
- Pathing report doesn’t let users choose their own nodes
- VERY strict and asinine data collection limits. Such as character limits on user properties and event parameters. Even the ‘page_location’ dimension is truncated after 300 characters
- GA4 javascript automatically batches hits, which can cause payloads to fail if they are above 16kb in size (which is really quite small)
Fred Pike from Northwoods: “I am absolutely sold on GA4 as the future”
I am still double-tagging. I’ve not moved any client to GA4 100% yet.
I’ve worked on two large-ish website relaunches recently and have done the side-by-side mapping of the user interactions we want to track.
In GA3, of course, I’m still having to stuff two or sometimes three fields into the event label or, sometimes, the event action.
YouTube tracking is a perfect example, where the event label usually has the video title, the YT URL, and, sometimes, the page the video was on.
Put that GA3 field-stuffing next to GA4 and it becomes immediately obvious how great the GA4 data model is.
I love the event-driven approach and love love love having enough parameters to handle all the data points I want to track. (At least so far!) So I am absolutely sold on GA4 as the future.
I am not a fan of the UI and I hate the limitations in the Explore report writer. But I think the data model, and knowing you’ll do much of your analysis outside of GA4, makes it viable.
I am also conscious of the 50 custom dimensions limitation (120 in the paid version), so I’ve started using three generic parameters: type, sub_type, text.
These are broadly applicable to a large number of GA4 events and, as long as you’re using them similarly across all instances, you don’t have to create a bunch of custom dimensions.
Peter O’Neill from ZHS Orchards: “GA4 configuration options are mixed, some very powerful and some missing”
We are collecting data in GA4 but not using it yet.
I think the data collection is fine, although it would be good to have nested variables (I don’t think this is available, not that close to it). The logic matches what I used for data layer structure so no issues there.
Accessing the data via BigQuery is fine, for the companies with people that can do this.
GA4 configuration options are mixed, some very powerful and some missing. Crazy that event names don’t allow spaces, the tool was created by a developer for themselves with no awareness of the actual users.
The basic UI is poor and that is why I wouldn’t recommend companies migrate over yet. The advanced analysis tool is good in theory but still needs work to get to the level of sitecatalyst discover from 10 years ago.
As per https://www.zhsorchards.com/blog/weekly-workshop/workshop-quadfurcation-of-digital-analytics-tools/, GA4 succeeds on level 4 but fails on levels 1, 2 & 3. This would currently need to be built in Data Studio (to keep as a free option) which is great for #1, possible for #2 but not right for #3.
Until the non-analysts in an organisation can use a digital analytics tool, it is not good enough. GA4 currently fails.
Miroslav Varga from Escape Ltd: “GA4 is Google’s solution and you have to take it seriously, no matter about your personal attitude”
GA4 is Google’s solution and you have to take it seriously, no matter about your personal attitude. But our clients are not very excited to switch to GA4.
They treat it as a beta tool and don’t like to invest too much time-effort-money towards the migration. By default, we install both properties, but they don’t use GA4 as much as they use GA3.
There is also the trouble with GDPR and problems of transferring PII data outside the EU (DPA’s in Austria, Germany, and France already have verdicts in that sense). GA4 is still to come.
Stockton Fisher from Greater Than Marketing: “I am very excited about the future of Google Analytics”
We are not using GA4 as the main reporting tool quite yet for any clients. Although we have it fully set up, it hasn’t taken its place as the main source of data.
However, there are many things it does do really well and we do use it for. For example, the explorations are a quick and easy way to find some specific answers. The BigQuery export is a welcome integration that allows us to create custom tables combining data from GA and other sources.
Overall, I like the new data models and how it uses events for everything, I think that is much more flexible. However, it is more complex to get started with so that’s the biggest con for new people trying to break into the platform.
I love where it’s going, I am very excited about the future of Google Analytics. There are just a few more functionalities they need to build still to make it #1 for me.
Mike Rhodes from WebSavvy: “I can understand why Google needs to make the switch, but I wish they’d find a way to allow businesses to view historic data for a lot longer than seems likely”
We’re not yet using it with the majority of clients as their main analytics tool. But we have installed it in parallel with UA for clients that were willing to test it early & start gathering data.
The main problems have been the learning curve needed – both by us & the client – to understand the new reports & make the mental switch to the new event-based model.
Clients inevitably want to see ‘the new version of’ a particular report which hasn’t always been easy for us to create.
As such we haven’t yet used GA4 data in our GDS reports, those are still mostly driven by the Performance platforms with some UA data as needed.
I can understand why Google needs to make the switch, but I wish they’d find a way to allow businesses to view historic data for a lot longer than seems likely.
But we’re excited to see the predictive metrics & the fact that less data will be omitted (although I’d love more transparency about ‘modelled conversions’!)
All in all, it’s a chance we’ll embrace!
Some companies have started early migration to quickly adjust to the changes.
Zorin Radovancevic from Escape Ltd: “What most people forget though is the additional reporting suite in GA4 found under Explore which is much more powerful than any reporting feature inside the old Google Analytics Universal”
In some cases, GA4 is the primary tool for analytics – mostly for clients which are APP first in their approach as they are already used to the Firebase model and reporting.
GA4 is in most cases present as it was advised by Google from the early announcement in a dual setup. As the migration was quite simple and mostly relies on a good dataLayer setup and backward compatibility for some features of the old GA, we started early on some projects and have a lengthy process of updating reports and awaiting additional feature parity updates to fully migrate.
The biggest advantage(s) would fall into 3 separate categories, IMHO, there are other factors to be considered but these are relatively less impactful:
1) RAW data access with the BQ export which gives us the ability easily enrich the data set for further use cases and a very flexible event model with the increased schema for custom dimensions and metrics even in the free version.
2) The process of (re)evaluating Analytics efforts in terms of what do we even need to port to a new solution is a good exercise and with the soon to be deprecated universal it just speeds up the process.
3) Machine learning in the backend exposed as reporting utilities (likelihood and anomalies) with the addition of Consent mode which even in beta allows for additional conversion and behaviour modelling in the future more regulated digital presence.
If we take into account the disadvantages I would emphasize that the product itself is unfortunately in its early stage in some of the most important aspects and needs more love from Google, especially in these areas:
- Documentation – it just needs to be brought to a decent level. For instance Measurement protocol (alpha/beta) and parameter reference is quite thin and in more mature implementations and use cases it is quite essential.
- The dreaded baked-in reporting – GA is known, comfortable, easy at least for the experienced users and even many competitors/alternatives have taken a similar UI approach (Piwik, Matomo, Yandex …) – so the UI hurdle is the main pain point which will take a lot of time to compensate.
What most people forget though is the additional reporting suite in GA4 found under Explore which is much more powerful than any reporting feature inside the old Google Analytics Universal (unless you had experience Analysis in GA360) which may also be a very good entry point to GA4 reporting – hard but rewarding.
Aside from the UI, there is a big misunderstanding with the basic concepts of dimensions and metrics – naming changed, use cases are a bit frightening at first such as event vs hit based and this is what one would expect when migrating from GA to Adobe Analytics but not in this case when upgrading from GA3 to 4. And to sum up, the biggest disadvantage is the initial reporting setup where it just feels empty compared to the old prebuilt interface – for the majority of users (All Pages, Landing Pages, eCommerce reports categorized based on metrics set and scope … ) all of the current standard reports are not useful and need to be rebuilt/built from scratch, unfortunately. - Feature parity with Google Analytics Universal – still not there and this should be the sole focus of further product development.
I have no doubt that GA4 will be accepted soon by the wider community but I also feel that GA will potentially lose a part of the customer base which needs a less complex preset tool which GA4 is not and will not be at least in the short term.
What differentiates Google from the other vendors is its built-in ability to quickly integrate with the entire Google stack and this will be a very heavy argument in any alternative finding process.
Moreover, the GA community will recover once enough How to articles become available for the current lacking easy-to-use reporting and implementation concepts.
Fortunately enough we have started with GA4 in its infancy and we have seen how it progressed so my view is optimistic.
Such significant changes cast doubts on the credibility of Google Analytics. As a result, different alternative tracking tools might be taken into consideration.
Gerry White from Rise at Seven: “The question for many businesses is are they prepared to rethink their metrics and re-configure something when the older version of analytics had been tuned and the teams had been trained?”
Google is sunsetting GA3 far too quickly, this isn’t a GA4 bashing thing it is simply that far too many larger businesses will need longer to transition, especially as the metrics are not one to one.
GA4 has a very different way of looking at the world and it is far more focused on users rather than sessions, it has taken me a little bit of time to understand what and where some of the data is, often a report you rely on in GA3 is simply not there but can be rebuilt.
The question for many businesses is are they prepared to rethink their metrics and re-configure something when the older version of analytics had been tuned and the teams had been trained?
The other question that is starting to pop up now is that this sudden, forced change is starting to make us question if Google is the right company to trust? Should we be investing into something like SnowPlow?
Sadly this isn’t as easy as it sounds for companies below enterprise level unless we start to see more analysts develop the ability to maintain and offer alternatives as solutions.
When it comes to GA4 migration, another important issue must be addressed: product knowledge gaps and the quickest ways to bridge them.
It is crucial to understand how the data is going to be analyzed and visualized and if this can be leveraged on your own without any external support.
Paul Koks from Online Metrics: “For clients mainly using the GA3 reporting UI for reporting/analysis, GA4 is really a pain – training of internal teams is important/required”
Here are some thoughts/observations:
For the majority of clients, GA3 is still the main tool in use. We have set up GA3 and GA4 in parallel and do a lot of extra customizations in the next few months and some probably later this year.
GA4 is still under heavy development (I feel) and not yet up to par with GA3. Should be much better at the end of 2022. Another challenge is that things are changing often (i.e. interface, list of standard dimensions, channel definitions) which requires (new) changes in the setup. For clients mainly using the GA3 reporting UI for reporting/analysis, GA4 is really a pain – training of internal teams is important/required.
So far, if a company lacks advanced (SQL) skills / BQ knowledge, Google Data Studio is a good tool for reporting (and some analysis). In addition to advanced analysis/explorations.
The data model is really good/flexible in a way (advantage), but companies can’t leverage that without external support.
Despite some remarkable improvements, the new reporting model can throw GA4 viability into question.
Brian Clifton from Verified Data: “To mitigate the issues, we are encouraging clients to use Data Studio as their reporting interface instead”
For implementation, the new data model is a huge improvement, particularly for enterprise users. For example, it provides a lot more flexibility with every data point now allowing for up to 25 associated parameters.
In addition, removing the free limit of 10M hits/month is a big plus, though it does mean considerably more planning and thought is required for the data structure from the get-go. That’s because if more than 50 custom dimensions are required (still a very generous allowance), you need to move to the paid/360 product.
Having said that, encouraging a well-thought-out plan for your data structure is a good thing, which is lacking with the “catch-all” approach of Universal Analytics.
However, I have to say the user experience of finding data and assessing reports is very poor in GA4. It’s as if the product has been built “by developers for developers”, with little consideration given to how existing users assess their website and marketing performance.
To mitigate the issues, we are encouraging clients to use Data Studio as their reporting interface instead. Maybe it is Google’s intention to focus on data collection and less on reporting, though it is an odd way to go about it.
Some experts tend to believe that these are not pitfalls yet new challenges we would have to face, as new technologies are being developed, and it is fundamental to make a mental shift as well.
Bryan Lamb from InfoTrust: “The more advanced our client’s UA architecture and usage is, the more likely they are to encounter a need for something that does not exist yet in GA4”
We do have clients using GA4, besides some of the predictive metrics and automated insights that Google has deployed in GA4 we’ve found that clients are liking the following in GA:
- Pathing and Funnels reporting is better
- Combining dimensions and metrics can be a bit easier (clients sometimes struggled with mixing scoped dimensions in reporting resulting in odd or unexpected results)
- The simplified data model has made reporting easier for some clients
- The ability to build a left nav that is tailored to an organization’s needs instead of a standard reporting suite. We’ve had challenges in the past because there are reports available in UA that either don’t meet business needs or are for features that are not used causing confusion and frustrations for end users
- eCommerce metrics are no longer session bound. If I view a product and add it to my cart but then stop for lunch and come back and buy in the afternoon reporting for that would show abandonment in the morning session and conversion in the afternoon in UA so EE reporting was a bit awkward. In GA4, with the removal of session-based data, I can now see a more accurate picture of eCommerce behavior using Funnel Reports
- Elapsed time is greatly improved
- Real-time debugging is much easier
- Consent mode. This is one that is currently of more value for our European clients but with the changing landscape for consent and privacy, GA4 will be better positioned with more granular data controls to manage how data is collected and used
- At the level we work at with clients the enhanced measurement is not really a plus because they all have more advanced customized GA tracking but for smaller organizations with basic tracking needs and little to no development resources I could see the enhanced measurement being a big plus for them
I would not call these pitfalls but more challenges some clients are facing in the transition to GA4:
- The deprecation of sessions. Almost all of our clients have built their KPIs around Sessions so moving to User in GA4 is challenging. Especially for clients that were using Landing Page, Exit Page, eCommerce Conversion Rate and other Session scoped Dimensions and Metrics
- Sessions and Users are calculated differently. Hit/Event scoped is really the only place you can expect parity between UA and GA4 so we’ve been encouraging our clients to dual tag so they can start understanding what the difference is for these numbers
- The mental shift from Category, Action, and Label to Event Name and Parameters. We’ve been having to do a lot of assisting and educating around this to help clients start to make the shift. A lot of our clients depend on a generic event dataLayer push method in GTM with Category, Action, and Label as parameters. We’ve had to assist a number of clients to determine how to bridge that gap to a GA4 friendly Event Name and Parameters to minimize the need to involve developers to update their dataLayer in the near term. (We don’t have any clients that don’t use a TMS but I imagine this becomes even more difficult if they use hard-coded UA. Even more so if they are still using analytics.js and have not moved to gtag.)
- Pending Features. While the product itself is out of beta there are still a number of product features that have not been released yet. The more advanced our client’s UA architecture and usage is, the more likely they are to encounter a need for something that does not exist yet in GA4. Google is continuing to release additional features so we’re confident the product will get there but in the short term can be a blocker for some clients working to transition to GA4 as their source of truth
- No views. We’ve had some clients find challenges in moving to a GA structure that doesn’t allow for the filtering and manipulation that UA provided with views
I would say one general thing to note is that we almost exclusively work with larger enterprises with more advanced analytics needs, so our client experiences are probably going to be different than the general population.
Matteo Zambon from Tag Manager Italia: “In order to use GA4 at its full potential and get results, our advice is to work hard on the switch of mentality and approach”
Biggest Advantages
In Tag Manager Italia we implement and manage GA4 along with Google Universal Analytics with our clients since October 2020, when the E-commerce tracking in GA4 was officially released.
From our experience with consultancy projects, the advantages of using GA4 are many, both for the companies that can make in-depth data analyses to reach their marketing objectives and for the Digital Analyst that can manage with more ease and accuracy several tracking tasks.
Here are the main GA4 advantages from our point of view:
- Thanks to a fully customizable Data model it is possible to manage the user tracking by adopting your own data structure, without the need of relying on the standard data structure given by Universal Analytics
- Ease of implementation: in GA4 everything is well defined as event nomenclature and parameters. This way, the implementation of push commands in the Google Tag Manager dataLayer is easier for the development team
- Audience and Conversion management: in GA4 barely everyone can easily create audiences based on events and parameters. Moreover, conversions simply rely on deciding which events correspond to which conversions
- In GA4 we can now count on more detailed and realistic dimensions and metrics that definitively give more actionable insights (e.g.: ga_session_id, ga_session_number, engagement_time_msec, user_engagement, etc..)
- Even the E-commerce tracking is more complete and specific thanks to tons of new events and parameters for better analysis, planning, and optimizing digital marketing campaigns
- The Explore Report offers the possibility to make in-depth and more granular data analysis. In particular, the Funnel Exploration Report allows us to use the Show Elapsed Time among the various funnel steps
- The GA4 DebugView is the feature we were all waiting for, to verify that the tracking implementation is “error-free”
- In order to use GA4 at full power, it’s mandatory to create a measurement plan as the first thing first in your digital analytics strategy. Implementing digital marketing and digital analytics strategies without a measurement plan is a big mistake because you risk losing track of the purpose of the tracking events you implemented. Furthermore, without a measurement plan, the possibility of making tracking errors and making wrong decisions becomes higher
GA4 is definitely a good product with great potential, but there are still a lot of things to improve.
Pitfalls
Here are the main GA4 pitfalls and a shortlist of things that should be fixed:
- Custom Channel Grouping: in the current state it is impossible to create your own channel grouping. (This feature is in the roadmap)
- Thresholding: if Google Signal is applied, even in standard reports you will have problems seeing the real metrics. In order to overcome this problem, you have to use BigQuery. GA4 refers to “unsampling” data, but this applies only to standard reports. In any case, the “unsampling” data are bound by Thresholding
- Reports are not user-friendly (including the Explore section). For example, you can’t resize the columns of the reports. Sometimes the size of the report is too invasive and the screen space seems to be never enough to have a full report overview
- The Realtime Report – in its current state – doesn’t allow to have a clear and immediate “picture” of the situation (at least, not like the “old” Universal Analytics)
- The 24h wait to see data is quite annoying, but we have to get used to it
- Data Filters based on IP addresses: from my point of view, this is a completely useless function. It is more convenient to manage filters with other parameters and not with the IP address
- The GA4 data stream can be linked to multiple domains, but a Search Console account can only be linked to one property. This is a limit that can be overcome only by using the paid version of Google Analytics 360
Other nice features that are in the official roadmap will arrive soon.
Here are the most awaited ones:
- Session scope e product scope
- Conversion metric
In conclusion, GA4 is a completely different “tool” from Google Universal Analytics you used to work with.
So, in order to use GA4 at its full potential and get results our advice is to work hard on the switch of mentality and approach.
Summary
To sum up, the sunset of Universal Analytics has prepared some significant challenges for us. New changes are awaiting us in the nearest future, and we should be ready to face them with solid knowledge and newly gained skills.
We hope this knowledge-share post has shed some light on the current situation and inspired you to look for the optimal GA4 solution.
If you’ve decided to upgrade to Google Analytics 4, check out our handy guide and learn how!
Let us know if you have already migrated to GA4 or if you keep using dual tagging. What are the biggest advantages and pitfalls you personally have experienced so far?
Check out our comparison of Google Search Console vs Google Analytics 4 and find out where lies their strength!
Join our discussion with experts and share your thoughts in the comment section – we really want to know more about your own experience with GA4.
Master the basics with our FREE GA4 Course for Beginners
