GA4 & GTM advice Measure what matters

A place for everything and everything in its place

A good place to start when considering your migration to GA4 is with the account structure. This is step 1 of the process of getting setup.

Account structures only need to be as complex as your organisational needs. Not complicated.

Before you begin

I found several examples in the GA4 documentation on account structure recommendations. I mostly agree with the approaches provided in the examples but I would do things slightly differently.

I highly recommend that you ask yourself (and your stakeholders) some fundamental questions before you decide on a sensible account structure for your GA4 setup. Note that I didnโ€™t say ideal setup, because your business is not static, and your setup will evolve over time. What we are trying to do, is not create the perfect account structure, but create something that addresses your current needs and hopefully grows with you.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

๐Ÿ‘‰ Who will be using the data? One team or many? Itโ€™s a good practice to group these users logically.

๐Ÿ‘‰ How many data streams? Web, app or both?

I have captured my approach in the flowchart below which steps you through the same questions.

GA4 Account Structures

Examples

To help bring the decision making to life, Iโ€™ve worked trough a few examples below. This should cover many of the common use cases you are likely going to encounter.

Not-for-profit

Requirements

In this example we look at a not for profit. They are based in one location and primarily raise funds in their local area. There is just one marketing and fundraising team.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

One location.

๐Ÿ‘‰ Who will be using the data? One team or many? It's a good practice to group these users logically.

One team.

๐Ÿ‘‰ How many data streams? Web, app or both?

One website.

Solution

โœ” Create one account, with one property for the web data stream.

Financial services group with several brands

Requirements

In this example we look at a large financial services group with multiple brands each with its own website. Individual teams look after each brand and site. There is no natural customer interaction between brands in the portfolio. The data from GA4 needs to be viewed by the respective marketing teams.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

One location.

๐Ÿ‘‰ Who will be using the data? One team or many? It's good practice to group these users logically.

Multiple teams. Each responsible for their own brand and website.

๐Ÿ‘‰ How many data streams? Web, app or both?

One website per brand

Solution

โœ” Create one account

โœ” One property for each website and the corresponding web data stream associated with it.

โ— In this case, I donโ€™t see a need for a property with data from all the individual brands feeding into it. I call this the โ€œFrakensteinโ€™s propertyโ€. Every analystโ€™s worst nightmare.

โœ” Instead, if such an aggregate report is needed with benchmarking that compares performance between brands, I would do this in Data Studio instead of in GA4

Educational institution

Requirements

In this example we look at a large educational institution with several faculties each with their own unique needs. There is only one website catering to both domestic and international students. The data from GA4 needs to be viewed holistically by the respective marketing teams (domestic and international) and by the different faculties.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

One location.

๐Ÿ‘‰ Who will be using the data? One team or many? It's good practice to group these users logically.

Multiple teams.

  • Domestic marketing team
  • International marketing team
  • Faculties

๐Ÿ‘‰ How many data streams? Web, app or both?

One website.

Solution

โœ” Create one account

โœ” One property for the entire website

โœ” One property for the domestic marketing team

โœ” One property for the international marketing team

โœ” One property for each faculty

Health & wellbeing app

Requirements

In this example we look at a health & wellbeing app with a website for marketing the app. There is just one marketing team. Customers will naturally move from the website to the app but the primary location for customer interaction is the app. The data from GA4 needs to be viewed by the marketing team to understand in app behaviour and website performance.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

One location.

๐Ÿ‘‰ Who will be using the data? One team or many? It's good practice to group these users logically.

Single team

๐Ÿ‘‰ How many data streams? Web, app or both?

  • One website
  • One app
Solution

โœ” Create one account

โœ” One property for the web data stream

โœ” One property for the app data stream

โœ” One property that combines data from the app and web data streams

Multi-national QSR brand

Requirements

In this example we look at a multi-national QSR brand operating in multiple geographies. Each market has its own website and local app. Each market also has its own marketing team. Customers can purchase their favourite meals directly from the website which is ecommerce enabled or from the app.

๐Ÿ‘‰ Where does the data reside? In one or many places for legal and/or business reasons?

Data from each country needs to reside locally.

๐Ÿ‘‰ Who will be using the data? One team or many? It's good practice to group these users logically.

Individual teams in each country. There will be a need to roll up reporting across multiple markets.

๐Ÿ‘‰ How many data streams? Web, app or both?

  • One website
  • One app
Solution

โœ” Create one account per market

โœ” One property for the web data stream for each account

โœ” One property for the app data stream for each account

โœ” One property that combines data from the app and web data streams for each account

โœ” Use Data Studio or some other BI tool to roll up reporting and compare performance between markets. Personally, I am not a big fan of aggregating country level metrics into a roll up view within GA4. I can't see it improving analysis but only smoothening the numbers by lumping metrics together.

If you couldnโ€™t find your particular use case above, or have any questions at all, DM me on LinkedIn and Iโ€™ll be glad to help.

I hope you found this helpful, and thank you very much for reading!

Does it spark joy?

GA4 is a major upgrade to Google Analytics. Instead of trying to compare it with the previous version, I think it is a good idea to forget what you know, and start afresh.

In the words of the modern sage Marie Kondo:

โ€œThe best way to find out what we really need is to get rid of what we donโ€™t.โ€

Find this liberating, you will. Trust me.

It all comes in a stream

All the data in GA4 arrives via a data stream.

At the time of writing this post, GA4 defined a data stream as a flow of data from your app or website. You can find data streams in the Admin section under Property.

There are 3 streams available out of the box:

  • Web for websites
  • iOS for iOS apps
  • Android for Android apps

However, in theory, a stream could be any customer touchpoint which gives us some insight into the power and flexibility of this new approach.

GA4 Data Streams

You can connect one or more streams to a GA4 property.

Iโ€™ve shown you how to connect a data stream in this short video below. It assumes you have a GA4 property already created. If not, donโ€™t worry, Iโ€™ve got another short video walking you through that as well.

Create a data stream in GA4

Create a GA4 Property

This is another one of the key shifts in the GA4 architecture. Google have gotten rid of Views. The Property is the smallest unit within the GA4 architecture.

Account > Property > Stream(s)

If you are wondering how many properties you ought to create, I find a useful way of thinking about the solution is to ask another question;

Who is going to be looking at the data and how would they use it?

Iโ€™m not trying to be mysterious here so let me break down how I would approach a scenario.

Case: Waves of Wellness

Waves of Wellness is a mental health surf therapy charity, committed to changing lives by delivering for-purpose, innovative support programs for people experiencing mental health challenges.

Iโ€™ve listed the key attributes about the organisation that I incorporated when developing a potential solution for their GA4 setup.

  • Single market (AU)
  • Single website
  • Web only

Solution (Web Only)

  • One account, with one property linked to the web data stream

If Waves of Wellness had a mobile app, I would approach it as follows.

Solution (Web + App)

In addition to all the property I created for web, I would create a separate property for app traffic and another property that combines app and web.

One account with

  • One property linked to the web data stream
  • One property linked to the app data stream
  • One property linked to the web and app data streams

This is a basic solution and does not factor in any data filters you might want to apply such as internal IP filtering. I will cover that in another post.

The reason I do things this way is linked to how I would use the reporting interface in the future.

  • Sometimes I might want to look at just data from the web without it getting muddled by app data.
  • Other times I would want to look at things wholistically.

Itโ€™s easier, and takes less effort to do this up front. So keep things in their place.

I hope you found this helpful, and thank you very much for reading!

If you want a hand with your GA4 setup, feel free to reach out to me via LinkedIn.