how to prepare well before July 2023

Adrien Podolak, lead analytics at Arcane

Adrien holds a Master’s degree in management from HEC. He stayed for 2 years at Venture Capital Analyst before joining Arcane in 2019 as a data analyst, then senior data consultant. He was promoted in March 2022 to lead analytics.

“Universal Analytics will no longer process new data starting July 1, 2023. Processing of Analytics 360 properties will be extended once, until October 1, 2023.” This is the phrase that now appears at the top of Google Analytics accounts. Could you explain what this actually means for users?

Universal Analytics is the version of Google Analytics being deprecated, in favor of Google Analytics 4. Such an evolution on a tool as widely used as Google Analytics requires a transition period, during which the two versions will coexist.

We are in this period which will end on July 1, 2023 for customers subscribing to the free version of the tool, and on October 1 for its so-called “360” version, which is paid for. After these accessible historical deadlines, no additional data will be recorded on Universal Analytics, whose properties will remain for a time for reading and analyzing the data.

10 years after its launch, it is therefore the end of Universal Analytics. Why did Google decide to qualify GA4 and end UA properties? What are the major changes in Google Analytics 4?

The digital ecosystem has changed in 10 years. Although Universal Analytics is a well-known and mastered tool on our market, it no longer responds as well to the business problems of its users as it did a few years ago.

The ability to collect and cross-reference data from applications and the web has become an important issue in the context of cross-platform navigation. GA4 meets this need thanks to its interfacing with Firebase (the application services solution acquired by Google in 2014).

Another important limitation of Universal Analytics is that cross-device and cross-browser users are not inferred. GA4 partially addresses this issue by introducing new user reconciliation keys (customer ID or Google Chrome login information). Users are now better recognized and recognized via GA4 which allows, for example, a better understanding of the attribution of conversions to the right acquisition levers. The GA4 audiences activated on Google Ads will also be more specified in this new context.

Raw data from GA4 properties can also be dumped daily into BigQuery, Google’s cloud data warehouse. Only properties from Universal Analytics 360 were eligible for export before. At Arcane, we build many data projects integrating first party data such as CRM, back-office or navigation data.

Through these few examples chosen from among other novelties, we feel that GA4 is intended to get closer to the new navigation uses of Internet users and to better cover the challenges of all the parties concerned in webanalysis (traffic loads, acquisition, analysts, marketing and data managers, etc.).

Why is it important that the new properties of GA4 be implemented sufficiently ahead of this deadline (July 1, 2023)? What timing for what actions?

The supply of a GA4 property is only possible from the moment the collection device (“tracking”) is configured and the GA4 property created. The whole challenge in webanalysis is to be able to analyze data over time, over a sufficiently wide date range and allowing comparative analyzes with respect to a given previous period. Year-Over-Year (YoY) analysis is very common in this regard.

Analyzes 2 examples:

  • I started feeding my GA4 property on January 1, 2022, so I will have over a year and a half of history on July 1, 2023 and can easily transition to GA4. I can also anticipate the change management internally so that the adoption of GA4 is as natural as possible.
  • I will start funding my GA4 property on January 1, 2023. The following July 1, I will have less than a year of history, which will not allow easy tracking of performance trends by YoY leverage. This approach is less precise, requires time and energy.

We normally communicate the date of July 1, 2022 as a deadline to insist on the recovery of one year of history.

Very important point, starting to feed GA4 properties does not consist in ending the feeding of UA properties which must continue the collection to allow temporal analyses, impossible to carry out for the moment in GA4 which has only very few data. history!

What advice would you give a brand to effectively migrate to Google Analytics 4? What should you pay particular attention to?

The main pitfall would be to consider GA4 as a simple version update. UA and GA4 are 2 different component tools and this difference should guide the conduct of the entire project.

Here are the main steps that we also follow to support our customers:

  • The framework to validate the outlines of the future implementation,
  • The technical implementation,
  • The configuration of the interface according to a grid of Arcane good practices,
  • Data consistency check and gap analysis with UA data,
  • Adaptation of related devices (Media, Cloud, BI, etc.),
  • Change management for better internal adoption.

What are the specificities of GA4 to know in terms of collection and activation to get the full potential of the tool?

They are many ! GA4 collection depends on a different data model than Universal Analytics. We cannot simply copy the GA4 collection mechanism identically to that of UA, otherwise we will not exploit the potential of the tool and see certain key reports not supplied. We recommend rewriting the Universal Analytics data model to take advantage of the specificities of GA4. The transposition can for example then be carried out from the client’s Tag Management System.

For activation, we encourage our customers to test the GA4 novelties. Predictive audiences based on purchase or abandonment probabilities are a good example of new audiences to test in Google Ads.

European data protection authorities have recently questioned the use of Google Analytics. Will the use of GA4 comply with the GDPR and meet the requirements of European authorities such as the CNIL? Will special conditions be taken into account in setting up accounts?

The questioning of Google Analytics by the European data protection authorities is the major subject that has been driving the ecosystem for several months. In its various communications, the CNIL does not distinguish between the two versions of GA and it cannot be said today that the use of GA4 will comply with the GDPR and meet the requirements of European authorities such as the CNIL. .

It will be important to continue to closely monitor developments in the market to provide as many elements as possible to the legal entities of the site management companies, so that each defines its own course of action on this subject.

In particular, we are awaiting elements concerning the agreement in principle for the transfer of personal data from the European Union to the United States, announced at the end of March. This agreement would be likely to put in place a framework allowing the use of Google Analytics.

We also follow the developments of the Analytics platforms and in particular that of Google concerning GA4. For example, the announcement was made last March that GA4 will no longer store IP addresses.

Last June, the CNIL finally paved the way for the use of ATMs that would comply with the GDPR via a proxification solution, the interest and impacts of which should be precisely studied before initiating such a project.

Will this compulsory transition to Google Analytics 4 not encourage certain brands to turn to other alternatives? As lead analytics, what tools would you recommend?

Our first role as a data consultant is to bring together all the parties involved within our client companies (legal, technical and business entities) in order to explain the issues, opportunities and risks of abandoning or maintenance of GA. We advocate a collective, concerted, anticipated decision made with full knowledge of the facts.

If the choice made by our client is not to retain Google Analytics, we support them in finding the best partner to meet their challenges. We are familiar with the alternative solutions, but do not see a single answer in the market and have therefore established a matrix of criteria to help in the decision making. We integrate into our analysis notions such as cost, data storage location, integration time, collection and reporting capabilities, integrations with the client’s media and cloud ecosystem, the wealth of documentation or the support provided by the solution. The diversity of our clients’ needs makes each study specific.

If the choice made is rather to stay in the Google ecosystem, we explain during the framing phases all the parameters that can be operated from GA4, as discussed above.

Whatever solution is chosen, we all prefer our interlocutors to get moving to take the direction of GA4 or another solution. Staying on Universal Analytics today is not an option.

Leave a Comment