1 month from now will officially mark 1 year since Universal Analytics’ successor, GA4, was released. But for any early adopters it might feel like you’ve spent over a year unpacking a host of new changes to the service.

At Lynchpin we know first-hand that there are some people still holding off on the transition to GA4. So, in this blog we will be outlining some of our key learnings over the past few months to help make your current or future adjustment to the service as pain-free as possible.


1. Maximise your data retention period, particularly for custom reporting

You might notice that straight out of the box, the default data retention period in GA4 is just 2 months… That’s right – If users forget to adjust this setting, they will indeed find themselves unable to retrieve any data exceeding 2 months in age from Analysis Hub (now named ‘Explore’) for custom report-building.

To adjust: select Admin > Data Settings > Data Retention and adjust to ‘14 months’ (GA4’s maximum setting).

For most GA users however, 14 months is still an insufficient time period for any historical analysis. In which case, we recommend activating BigQuery (or a similar warehouse solution) to store and preserve your data for future analysis and year on year reporting. BigQuery also opens up for you a new world of analysis possibilities.

2. Approach the ‘Reset user data on new activity’ toggle with some caution

You may be thinking: “Shouldn’t my 2 or 14-month deletion date roll forward each time a new user visits my site if ‘Reset user data on new activity’ is turned on?”

…In theory, yes. However, it is worth noting that from our experience, on at least one occasion, Explore prevented the selection of data exceeding 14 months in age despite this feature being turned on.

Whether this is an isolated incident or a bug that Google will be looking to fix in the upcoming months, we strongly suggest that users exhibit some extra caution by activating BigQuery (or the like) as a safety net to avoid potential data losses.

3. The importance of custom reporting

Most users will be relieved to know that standard reports are not affected by the data retention considerations mentioned in point 1, so retrieving data that exceeds even 14 months in age without the use of BigQuery will be completely fine.

However, those familiar with UA may notice that standard reporting options in GA4 are limited in comparison to what they’ve become accustomed to.
Therefore, to get the most out of your data in GA4, we say that custom reporting is going to be of increased necessity to understand and implement well – further backing the importance of our previous points about data retention and mitigating potential data losses.

4. Improvements to analysis techniques

While most analysis techniques haven’t undergone any massive changes, Explore now offers users an incrementally better experience with features users have come to know well, including Path Analysis and Funnel Analysis.

The only considerable change to functionality comes from the introduction of a host of handy ‘User lifetime’ metrics now available by visiting: User lifetime > Metrics > selecting the + button.

From there, users will be taken to a brand-new list of tools that are organised and labelled to help users (experienced or new) to get up and running successfully with metrics such as: Churn probability, LTV, Predicted revenue, Purchase probability, and more.

It’s worth noting that it is important to understand how lifetime and predicted revenue metrics are defined and any limitations based on the available data in GA. Alignment may be needed with offline data and LTV calculations that sit elsewhere in the business. Depending on your business and the available data, you may get a usable churn probability, but for some, it will be pretty meaningless without the benefit of other data sources and more sophisticated modelling.

5. Flexibility with event tracking

Many users will notice that GA4 sees a huge shift in Event tracking. Previously, in UA, users could track an Event as: Categories, Actions and Labels (three pieces of data).
However, in GA4, users have an Event Name and up to 25 parameters (additional pieces of data), offering far more flexibility. And as mentioned in point 1, users will be able to see even more in BigQuery, opening the door for more extensive analysis.

6. Some features worth keeping an eye on…

a. Anomaly detection (via Line graph visualisations)

When inserting a line graph into a report, users will now have a view of a simple negative or positive percentage between ‘Actual’ and ‘Expected’ data points (the accuracy of the ‘Expected’ figure should be based on training times set by the user).

Overall, this feature should help most users save some time spotting data points that may require further analysis. However, what may be flagged as an ‘anomaly’ by the tool is arguably a bit generous in the eyes of most analysts.

In one of our tests, a simple dip in website visitors experienced each weekend was consistently flagged red, deeming it an anomaly.

While highlighting this pattern in user behaviour is helpful to some extent, having the tool be stricter about what it categorises as an ‘anomaly’ would be a welcome step towards unlocking some extra value for most users. Therefore, it’s important to always have an analyst who can understand the level of anomaly detection being applied and who can adapt or build their own approach that is relevant for the business.

Like point 2, hopefully this feature is something that Google will be looking to refine algorithmically over the upcoming months.

b. Detailed event streams (via App-instance ID breakdowns)

New to GA4 is a highly granular report of event information. By right clicking a row and selecting ‘View users’, a User Explorer report is then created for that event. By selecting the ‘App-instance ID’ of any of the users listed, you will be taken to a view where you now have a total summary of that particular user, including their: ‘First seen’ date, location, device type, a breakdown of their event stream, and more.

While the opportunity here may be hard to define for most users – especially those dealing with larger volumes of data – For some, this new view may help drill down to a level of insight that may be necessary and offers increased flexibility in terms of analysis capabilities.


For anyone looking to make leap to GA4 (or anyone who has recently upgraded), our overarching advice would be to not underestimate the value of a helping hand. Our testing processes and live projects over the past 11 months have uncovered some vital insight into the best practices for implementation, analytics set up, reporting and more – the details of which we’d be happy to discuss with you personally regarding your unique circumstances here.

And for any of our readers looking to learn more about GA4 before considering or planning an upgrade, we recommend taking a look at our blog from earlier this year where we discussed in-depth: what changed, is GA4 right for you, migrating from UA, and much more.