How to Compare Reports from Universal Analytics with Reports from Google Analytics 4

On 1 July 2023, standard Universal Analytics (UA) properties will stop processing new hits and all Google Analytics users will have to migrate to Google Analytics 4 (GA4) to keep collecting data.

This may seem like plenty of time, but it’s highly advisable to start migrating now.

Because each service uses a different measurement model, you cannot migrate historical data from UA to GA4, and will therefore have to export all your reports if you want to keep your data from UA after the service is completely sunsetted.

By the same token, it is not possible to directly compare reports from UA with reports from GA4. Doesn’t this effectively render historical data from UA useless? No. There is a workaround. But it takes time.

In this article, we will explain some key differences between the two services, as well as what you can do to meaningfully compare data collected by UA with data collected by GA4.

Google Analytics 4 vs Universal Analytics

The biggest overall difference between UA and GA4 is the model of measurement: whereas the UA model of measurement is based on sessions and pageviews, the GA4 model of measurement is based on events and parameters. Essentially, in GA4, any type of user interaction can be captured as an event. This means that all UA hit types translate to events in GA4.

For this reason, importing historical data from UA to GA4, much less comparing reports from the two services, is unlikely to ever be possible.

Screenshot of a report in Universal Analytics.

In terms of data retention, GA4 can retain data for a maximum of 14 months, which is significantly less than what Universal Analytics allows.

For companies using Google BigQuery, this is not a problem because GA4 offers a free connection to it, unlike previous versions of Google Analytics. But companies using data warehouses like Snowflake, Redshift, or Firebolt will need an integration tool to connect their data and retain it for longer than 14 months. Dataddo, for example, can connect GA4 to any warehouse.

Screenshot of a report in Google Analytics 4.

GA4 Migration: How to Meaningfully Retain Historical Data from UA

The earlier you start using GA4, the more historical data you will have directly in the service.

But even if you start using GA4 today, at some point you’ll probably want to compare data collected by GA4 to data collected by UA before today. What then? Here is the workaround.

By running UA and GA4 tracking codes in parallel, you can get familiar with discrepancies in advance of UA’s complete sunset, and develop average ratios that let you “convert” UA data into data comparable with GA4 data.

Follow these steps:

1. Identify the UA data models (metrics and dimensions) that you want to track and try to replicate them in GA4.
   
   
2. Start sending the data from both versions to a data storage (BigQuery, Snowflake, Redshift, etc.). Make sure that each data model is materialized in a separated database table.
   
   
3. Once you have the data models from each version of Google Analytics tracked and collected in separate tables in your data storage, start calculating average ratios between equivalent GA4 and UA metrics.

So, let’s say that you want to track views of a certain landing page by date. 

As mentioned in point 2, you would build database tables with dates and pageviews for each of the two services in your data warehouse.

Next, you would calculate at regular intervals the difference between pageviews from UA and views from GA4 (the names are different but the metrics are equivalent) to determine an average ratio between the two metrics. You would then apply this average ratio to any data from UA to “convert” it and make it comparable to data from GA4.

Take the numbers below, for example.

To find your average ratio, you would first need to divide the GA4 views by the UA pageviews in each equivalent row. So:

1,708 / 1,684 = 1.0143

2,401 / 2,365 = 1.0152

2,789 / 2,795 = 0.998

Then, average these totals together to get your average ratio:

(1.0143 + 1.0152 + 0.998) / 3 = 1.009

Bingo! Your average ratio is 1 (in UA):1.009 (in GA4).

In the future, you’ll know that, when comparing pageviews in any UA report with views in any GA4 report, you’ll need to multiply the UA pageview number by 1.009.

This example is oversimplified, but the point should be clear: over time, following these steps will make any historical data from UA immediately comparable with any equivalent data you start tracking now in GA4.

Now Is the Best Time to Start

If you haven’t already, it’s advisable to start transitioning from UA to GA4 now. Not only will this give you the time you need to develop average ratios for equivalent metrics, thus helping you keep historical data from UA meaningful, it will allow you to collect as much historical data as possible within GA4 itself.

Once UA leaves us for good, you’ll be comparing apples to apples.