Gearing up for the End of Third-Party Cookies: How to navigate the Cookie-less Future?

Preparing for the end of third-party cookies

If your site uses third-party cookies, it's time to take action as we approach their deprecation. To facilitate testing, Chrome has restricted third-party cookies for 1% of users from January 4th, 2024. Chrome plans to ramp up third-party cookie restrictions to 100% of users from Q3 2024, subject to addressing any remaining competition concerns of the UK's Competition and Markets Authority.

Although the third-party cookies are phasing out, it doesn't mean tracking users would be completely put to stop. First-party cookies will be considered more important moving forward as new marketing plans and products will be based on it.

What are first-party cookies?

While limited to a single site, first-party cookies play a crucial role by performing essential functions. They serve to remember a user's previous login status, establish accurate language preferences, and even maintain a persistent shopping cart across multiple sessions.

How do customers initially discover our brand or products? What are the key drivers that lead customers to make a purchase? Are there any patterns or trends in customer behaviour across different marketing channels?

These are a few of the questions our clients usually ask us – To try and understand a customer’s mind is very difficult but it is made easier by creating a  ‘Customer Journey’

Creating customer journeys will be crucial for the cookie-less future as most of the marketing solutions will be based on these journeys which we have already worked on such as the Custom Attribution Model, RFM, etc.

What comes to your mind when I say ‘Journey’?

Basically, when we travel from one place to another. Every journey we take has its own destination and there is a specific path that one follows to get to their destination.

What is Customer Journey?

To decode this let us consider the destination as conversion and the customer as the entity travelling. In the world of data, we have each customer who follows a path wherein there are circumstances where a customer does not complete the journey and there are instances where the customer completes a journey by getting converted either way, it is considered as a ‘Customer Journey’.

The GA4 data set is heavily organised around event and user data, supplemented with device, geo, app and traffic source data. In the Universal Analytics Schema, every row in the data set represents a single session. As the GA4 properties are event-based, every row in our data set represents an event. Every event in turn can contain multiple event parameters and corresponding values.

To show you how this works in BigQuery, we queried our nested sample set:

To create a customer journey using the following data we need to unnest the data and have an understanding of the conversions to create a custom customer journey.

NOTE: Sharing the GA4 schema to gain a comprehensive understanding of the data.

Every business is unique in itself and so creating a journey depends on the requirements of the specific client.

To simplify things we consider there are two different types of journeys a customer can have which are as follows:-

• Buyer’s Journeys

The Buyer’s Journey is the entire conversion experience from pre-purchase to post-purchase. It covers the path from the customers that are converted.

In other words, buyers don't wake up and decide to buy on a whim. They go through a process to consider, evaluate, and decide to purchase a new product or service.

To calculate the conversion path using first touch and last touch, we need to have access to the data of the users who have converted. This data should include the sequence of pages/source/medium, Campaign they visited prior to converting.

Using this data, calculate the most visited path, path length, and the number of converted users for each path. For example, if you had a user who visited page A, page B, page C, and page D before converting, you would calculate that path A-B-C-D as the most visited path, with a path length of 4, and count the user as one of the converted users.

In this scenario, we will need the events that are considered conversions so that we can bifurcate the journeys based on conversions if needed.

• Potential Journeys

Potential journeys are journeys of customers who are not converted and haven't had conversion until now.

In this scenario, we will see customer engagement on different levels and what path they have followed up till now.

NOTE: We have bifurcated journeys as it is easier to understand but if we don't specify conversion in making a customer journey then we can consider all journeys as same.

It is important to understand which dimensions and metrics are to be used while creating such journeys and if there are additional custom dimensions and metrics as per the requirement we can consider them as well.

We have provided a sample solution so you can understand how you can view your results.

Sample:

Here I have created customer journeys using –

‘bigquery-public-data.ga4_obfuscated_sample_ecommerce.events_20210131’

Dimensions & Metrics:

Event_Date

UserID

Event_Timestamp,

Event_Name

Page_Title

EventCategory

EventAction

EventLabel

This is a list of dimensions and metrics which are standard to the data in the GA4 Schema

Custom Dimensions & Metrics:

Event_Period

Conversions

Sequence_Number

These are a few of the custom dimensions and metrics we have considered for our example.

We can use this list to create the customer journeys

Here is a sample journey of a single user ‘6331004.154’, we can classify this as a buyer’s journey where in we can clearly see at the end of the journey where conversion is being completed (i.e. conversion = 1). In this scenario, we have considered conversion as ‘Page_Title = 'Checkout Confirmation'’

NOTE: I have shared with you a document which consists of both buyers as well as potential journeys, feel free to download the same to see all the journeys

Once we have these journeys we can apply - ‘Custom Attribution Model’.

Please refer to ‘Custom Attribution Model’ documentation for further details.

Additional Resources:

 [Link]