Last year, Google announced that they are going to stop the data collection in Universal Analytics (UA) and instead replace it with Google Analytics 4 (GA4). A major inconvenience, some might think. We at Conversionista! see this as a big opportunity to ramp up your web analytics and reporting to the next level. So how can we use GA4 to improve the way we work with Google Analytics data?
This blog post will take you though some of the possibilities of managing your Google Analytics data using Google BigQuery.
Let’s start with the basics – the tool itself. Google Analytics 4 (GA4) has a completely new data model, a new user interface and a new Data API. With this, GA4 offers superior measurement and insight possibilities compared to its predecessor, Universal Analytics (UA). And last but not least: it includes a major feature: the Google BigQuery Export. This means that all Google Analytics accounts using the latest property type can now integrate natively with BigQuery. This is a big deal, as this connectivity was previously only available for GA360 accounts. So what does this mean and what possibilities does it unlock for your web analysis capabilities?
What is Google BigQuery?
BigQuery is a cloud-based data warehouse used to collect, save and manage large sets of data. The C! Analytics team makes extensive use of BigQuery when working with Google Analytics (GA) data, as it enables many possibilities for data analysis and handling, impossible to do in the GA user interface (UI) alone.
It provides the user with the tried-and-true Google Analytics data model, but with the additional ability to break it down on a more granular level and to customise the reporting and visualisation to the company’s specific needs using a SQL-like query language.
Improved Data Quality
The Analytics interface includes a number of tools to make it easy to perform ad-hoc analysis. However, in order to keep the interface as fast as possible, compromises have been made regarding flexibility and data quality. The data in the GA interface is already aggregated. For an analyst, this will be perceived as restricting as it limits the flexibility to analyse user level data points happening across multiple sessions and devices. When exporting the Analytics data into Google BigQuery, however, the data becomes available on event level and can be queried and aggregated in a manner customised to the specific need of the business. This becomes especially relevant for companies offering more complex products or services, as these users usually need multiple visits over a longer time period, often using several devices, before they feel ready to complete the goal.
We know that a person might add an item to their cart in one visit and complete the purchase in another. Here BigQuery comes to the rescue, as it provides the ability to analyse at user level using the raw data provided to connect the different sessions and attribute actions prior to the conversion to as far back as your data ranges. This makes you able to see and analyse pre-purchasing behaviour from users, and hence gives you a more complete and realistic view of the user journey.
Another advantage of exporting the data to BigQuery before doing an analysis, is to get access to the full dataset without the limitations of data sampling. Both Analytics Standard and Analytics 360 sample session data at the view level, after view filters have been applied. This means that data is likely to be sampled when running a custom GA report or when using a long date range. When this happens, Google extrapolates only a fraction of the data and models the rest. This means that the data shown in such reports are an approximation. Using this data as a base for analysis and business decisions can be directly harmful as actual results can differ from the sampled results. In the BigQuery export, every hit is available to query and will give you results based on the whole dataset.
Taking the Analysis to the Next Level
In the sections above, we have highlighted how your data analysis can be improved both when it comes to quality and flexibility. It’s important to know that exporting the data to BigQuery also opens up for more advanced and sophisticated possibilities to use data as a foundation for your business strategy. Once you have your Google Analytics data in BigQuery, you can combine it with data from other sources to gain a holistic view of your site’s performance.
Breaking Down Silos
Many companies work with their data in silos, where they have separate dashboards and reports for their 1st party data in e.g. Azure or AWs and for their Google and marketing channel data.
With your GA4 data in Google BigQuery, you can easily send it to another platform to enrich your CRM and ERP data and to make sure you store and visualise all relevant site and user data in one place in order to break down these silos. Or, you can work the other way around and link your internal data to the GA data in BigQuery, to get a picture of the full customer journey: all the way from the first interaction, initial acquisition to long term customer lifetime value. You will also have the possibility to combine your data with reports from Google Ads, Facebook and other marketing channels. In doing so, you get a full understanding of how your channels and online marketing strategy perform from a channel specific behaviour and get a full cost and conversion perspective, regardless of the different channel attribution models.
With all the data in one place, you can use your data warehouse as the one source of truth for customer and site performance. This comes with the possibility to use machine learning and predictive analytics to perform this kind of user journey mapping, together with cluster analysis, statistical modelling for forecasting future conversions, look at projecting return of ad spend or setting up anomaly detection systems to highlight if there are issues with the site. This can be done both in BigQuery ML, which lets you create and execute machine learning models using standard SQL queries, or connecting your data for use within the programming language Python.
We know many companies are struggling with the migration to GA4 and the implementation of a semi-new tool with a slightly different data model than they are used to. Our hope is that this blog post will open your eyes for, in our opinion, one of the main advantages of GA4 compared to UA. It will not only pave the way for a more sophisticated way of doing web analytics, but could also help you break down your data silos and get an improved way of optimising your site performance in an even more data-driven way.
We at C! are here to help you, so if you have any questions or would like some help with data analysis please feel free to contact us.