Home » Guide to Google Analytics 4 for Fashion

Guide to Google Analytics 4 for Fashion

2 factors have influenced and pushed for GA4 for example the privacy policy. Also the tech players like Apple are progressively deprecating the third party cookies.

Google Analytics help fashion business answer some fundamental questions in a very quick manner:

  • Which users segments are more profitable for the business? 
  • Which marketing campaigns are more effective?
  • Which content is helping you achieve your business goals
  • Where are your users coming from, and where on your site or app are they are leaving

Let’s imagine that you want to run a promotional campaign that entails a discount on a selection of products in your store. Your manager asks you how it’s going to impact on the channel P&L, what do you answer? Google Analytics can help analyse the performance of previous campaigns in terms of both traffic generated, orders, average order value and total revenue generated. And if you want you can also upload the returns data and evaluate the net results of the promotional campaign, several weeks after it’s complete, when all orders will have included the returns.

Let’s imagine that you need to choose between a performance campaign and an advertising campaign based on impressions on popular fashion magazine. How do you calculate the potential return on investment?

Google Analytics has always used first party cookies however there has been a reduction of the data collected, some are activate after the consent. Some are limited to comply with the laws.

Third party cookies are often used by marketing tools to create to create audiences to target, for example similar audiences, in Google Ads and Google Marketing Platform GMP.

  1. Google has invented the one tag that is not just for analytics but also for campaigns and the whole marketing platform. That’s a first party tag to analyse the data, not activation
  2. Created the consent mode, above the GMP to gather the consent, this sends the data to the marketing platform. If the user doesn’t accept the cookies, the platform sends some anonymised data that allow the Marketing platform to understand and measure the results of the campaigns.
  3. Google Analytics 4 is integrated with the previous 2 steps and it’s at the heart of the marketing platform, with first party data available. We can use the GA4 platform to import CRM data and retail data using GA4 interface or Big Query. Big Query is a sort of data warehouse. Sync with big query is free with some limitations in the free version.

Google has also communicated that the historical data will not be available after 6 months after the sunset of Google Analytics Universal, so we need to export the old data from GA Universal and and re-use of the old GA data.

For example big query is already integrate with ga360 but for the free account there is no possibility to mass export the whole database.

How to migrate to GA4?

All new implementation pass by tag manager either google’s one or other provider. 

Check the data is consistent with old GA Universal

When should we activate GA4?

As soon as possible.

It will be the only tool to track, and it starts from scratch.

The interface has changed and therefore we need to learn how to read the data.

What are the evolutions of GA4 vs Google Analytics Universal

GA4 is built on 3 pillars

  1. Compliance with European and American privacy policy;
  2. Focus on users and events rather than pages and sessions, cookies are still present;
  3. Machine learning already active and present in the current version of GA4;

Privacy features

GA4 Privacy features

IP Anonymisation, according to the Privacy authority, it wasn’t enough in the previous version.

One of the privacy features implemented in GA4 is the fact that GA4 doesn’t store users data.

GA4 tracks the IP addresses of the users in order to recognise the geo ip location and to assign the user to a geo report but it doesn’t save the IP of the users in the GA4 database. By not saving the IP address of the user it’s not possible to recognise the single user but only to analyse the engagement data. For example it’s possible to analyse the purchases of users from a particular city, but it’s not possible to know the exact address of these customers.

Il French Privacy Authority Cnil has stated that GA4 should be server side, to avoid the ip. The italian garante doesn’t provide guidelines as it’s mere control role.

Data retention in Google Analytics 4

Less data is retained however the data is available in the database also for longer period. However some information about the users will not be available for longer periods e.g. custom data to comply with privacy laws

Data delation in GA4

In GA4 it’s possible to delete some user data e.g. sensitive data that has been collected. You can use this feature to clean data that corrupts the analysis. GA4 is very effective in deleting data. The new data model is much more flexible because it’s linked to events rather than user sessions.

GA4 tracks users’ actions rather than pages

When users browse your website they complete actions like view a page or download a file. This is a significant difference in the way we need to read and build our reports. The actions taken by the users are defined as engagement.

Cross device tracking

GA4 collects data from Web and Native apps. Google Analytics 4 leverages the Google Graph ID to track users across different devices and platforms.

Google Analytics 4 can use four different methods to unify them into a single cross-device user journey:

  • User-ID
  • Google signals
  • Device ID
  • Modeling

Analytics creates a single user journey from all the data associated with the same identity.

Google Analytics 4 still places a cookie, while GA universal has a user ID to track cross device, this is not so simple in a segment such as fashion e-commerce, because the users only log in at the end of the customer journey and in some cases users don’t log in at all.

GA4 uses the graph ID di google, based on google signals, where Google puts and uses the proprietary graphs ID based on Google Login data to analyse brands performance on GA4.

Lately there are new release features like enhanced conversions, that allow you to pass to the media tag the email of the users. Through the email Google is able to associate the user behaviour from GA4 to Google Graph signals.

In GA4 everything is an event

Because of the granularity of the events or actions tracked by GA4, the data is more customisable and malleable or manipulable. The data is more granular.

This can require more awareness on what we want to track. We need a proper tracking plan. It’s more similar to adobe analytics. It requires planning in the implementation. You can also build custom views.

Events or user actions in Google Analytics 4

GA4 Flat datamodel

The data model in GA4 is flat, there isn’t the hierarchy that we see in GA Universal: Behaviour > Content > Pages. This is because GA4 tracks users behaviour on both web sites and native mobile apps, where the distinction between pages and actions is not relevant any longer. Typically in the native apps users are engaged in doing something rather than browsing or reading.

The flat data model of GA4 allows the users of GA4 to analyse the data cross platform.

Type of events in GA4

In GA4 there are several events pre-configured and other events that you can implement via Google Tag Manager.

For example the scroll is tracked out of the box (OOTB).
File downloads and video views are OOTB, and also links to external website useful for magazines and press.

Custom Dimensions in GA4

The use of custom dimensions in e-commerce is very important but is not yet available in GA4 Nov 2022.

We can set custom dimensions for users logged in, ID, generic such as template page and device, seasons, country. The product ones are not available yet e.g. discount.

Active users in GA4 is the default users metric

The active users in GA4 is the default user tracking metric. An active user is generated in GA4 when GA4 records one of the following the first-visit, first open or engagement_time_msec. The idea of active users comes from the native apps users where it’s very important to distinguish between registered users and users who are actually engaging with the app.

Because in GA4 these two worlds, the website and the apps, are merged this become a critical metric – KPI.

Universal Analytics highlights Total Users (shown as Users) in most reports, whereas GA4 focuses on Active Users (also shown as Users). So, while the term Users appears the same, the calculation for this metric is different between UA and GA4 since UA is using Total Users and GA4 is using Active Users

Page view vs Unique page views

In GA Universal you had Unique Page Views within the session, Unique Pageviews has been removed entirely Google Analytics 4, the metric doesn’t exist. GA4 provides two similar metrics instead:

  • Views represents a view of a page on your site, and is identicle to the UA Pageview you’re familiar with
  • Users represents the number of users that visited a specific page

In GA4 all unique events are related to users. Moving the focus of the tracking from the sessions to users the unique metrics are based on users.

So if a user see the pages twice it’s only one unique user and 2 page views.

The sessions in GA4 still last 30 minutes but UTM tracking is different

The sessions still last 30 minutes.

If I start a session from organic and after 15 minutes I come back from an ad, GA4 maintains the same session, while GA universal counted twice. This is much more similar to Adobe.

Generally the session in adobe are less the session in GA universal.

How is the bounce rate calculated in GA4?

Initially the bounce rate made sense because there was only one tracking solution e.g. page view. after a while sites have developed several tracking metrics e.g. scroll.

Initially Google said it wanted to stop it with GA4 but then decided to add it back because of the need of professionals.

The bounce rate wasn’t tied to the time on site, so it was possible for a user to spend a minute on a single page and get the bounce.

GA4 introduces the timeframe variable in the tracking.

Engaged session

An engaged session is a session that lasts longer than 10 seconds, has a conversion event, or has at least 2 pageviews or screenviews.

The bounce rate in GA4 is calculated as the difference between the total sessions and the engaged sessions. In other words the users that didn’t engage.

How is an engaged session calculated? See the FAQs below.

Engaged session

a session is an engaged session when

  1. a session that lasts longer than 10 seconds,
  2. has a conversion event,
  3. or has at least 2 pageviews or screenviews

official definition from Google.

Tracking scrolls, download and other events

In order to make sure you are tracking scrolls and other events you need to go in the Admin section of GA4 and enable “enhanced measurement”

Enable Enhanced Measurement in the Admin section of GA4 property

The introduction of Artificial intelligence in GA4

Analytics Intelligence uses machine learning and conditions to help you understand and act on your data.

How does GA4 artificial intelligence work?

GA4 Intelligence learns what are the typical user behaviours on your website and when it sees relevant differences it shows them to you. See the examples below.

GA4 AI showing an spike in users from Thailand
GA4 AI showing an spike in users from Thailand
GA4 Artificial Intelligence, has noticed a spike in page scrolls
GA4 insights showing a spike in page scrolls that is significantly different from the expected behaviour

Analytics Intelligence provides two types of insights:

  • Automated insights: Analytics Intelligence detects unusual changes or emerging trends in your data and notifies you automatically, on the Insights dashboard, within the Analytics platform.
  • Custom insights: You create conditions that detect changes in your data that are important to you. When the conditions are triggered, you see the insights on the Insights dashboard, and you can optionally receive email alerts. You can create up to 50 custom insights per property.

GA4 AI predictive capabilities

Based on the purchase behaviour of the previous customers, GA4 is able to calculate how many users are likely to convert in the near future.

In order to be able to make these types of calculation GA4, like any other learning machine, needs a sufficient amount of data.

These groups of users are called predictive audiences

A predictive audience is an audience with at least one condition based on a predictive metric. For example, you could build an audience for ‘likely 7-day purchasers’ that includes users who are likely to make a purchase in the next 7 days.

Predictive metric: purchase probability metric

Can the data layer structure be re-used for GA4?

Yes, the data layer structure can be reused. If you have implemented it on Google Tag Manager you’ll need to convert the GA Universal Tags to the GA4 Tags.

Should we keep the two versions of GA tracking in parallel?

Yes, you should keep them in parallel and confront the data, in particular the tracking of the sessions it’s different from GA Universal to GA4.

Does GA4 it track the IP address?

GA4 tracks the IP to recognise the geo ip location but it doesn’t save it. It’s not saved in GA4 database

The traditional KPIs and metrics of e-commerce are available in GA4?

Some of them are available some others aren’t, but there are additional metrics in GA4.
There are some differences: the active users is a new metric.
The total users are the same, in GA4 there could be less because GA4 enriches the user data data aggregation based on several elements like Google Graph ID. In other words GA4 it’s more accurate in recognizing the unique users.

What is an engaged session in GA4?

An engaged session is a session that lasts longer than 10 seconds, has a conversion event, or has at least 2 pageviews or screenviews.

How does GA4 artificial intelligence work?

Google GA4 intelligence learns the most frequent patterns of usage of your website i.e. the macro-trends and highlights any users behaviour that present anomalies (micro-trends).

Leave a Comment

Your email address will not be published. Required fields are marked *