
Revenue per channel in GA4
Not sure where to look when you want to see revenue by channel? Find out the different approaches here.
You look at a GA4 report and see three different “users” from three different sessions.
But your gut tells you it’s actually just one person: browsing on their phone during their commute, checking again on their work computer at lunch, and finally making a purchase from their laptop at home.
If your data can’t connect those dots, you’re not seeing the full picture.
You’re counting devices, not people.
This is precisely the problem GA4’s Reporting Identity setting is designed to solve.
It’s the engine under the bonnet that decides how Google Analytics “stitches together” all those separate sessions to create a single, unified user journey.
Choosing the right setting is crucial for accurate reporting (as is setting it up correctly fyi).
So, let’s look at how it works and which option you should choose.
Before you can pick your main setting, you need to understand the three methods—or “identity spaces”—GA4 can use to recognise a user.
This is the most accurate method by a country mile. When a user signs into your website, you can assign them a persistent, anonymous ID (a “User ID”). You then send this ID to GA4 along with all their activity.
Pros: It’s the most accurate way to track a single user across multiple devices and browsers, as long as they are signed in. It uses your own first-party data.
Cons: It’s more complex to set up (it requires developer input) and is only applicable if you have a sign-in or login area on your site.
This is the most common method.
On a website, the Device ID is simply the client ID stored in the user’s cookie. On an app, it’s the App-instance ID.
Pros: It’s on by default. You don’t have to do anything to set it up.
Cons: It’s device-specific. If the same person visits from their phone and their laptop, GA4 will count them as two separate users.
What about users who decline analytics cookies? This is where modelling comes in. GA4 takes the data from similar users who did accept cookies and models the behaviour of the users who declined.
Pros: It helps to fill the data gap left by cookie consent banners, giving you a more complete dataset.
Cons: It’s not “real” data—it’s an educated guess. Crucially, this only works if you have Consent Mode v2 implemented correctly and meet certain data prerequisites. And the more data you get, the better it becomes.
Here’s the simple breakdown:
If you have User ID set up… you should be using Blended or Observed. If you’re using Device-based, you’re effectively ignoring your most valuable user data.
If you have User ID and Consent Mode v2 set up… Blended is your best choice. It gives you the accuracy of User ID combined with the fuller dataset from modelling.
If you don’t collect User IDs (e.g., you’re a content-only site)… your choice is between Blended (which will use Device ID + Modelling) and Device-based (Device ID only). If you have Consent Mode v2, Blended is still the better option.
If you want your data to look as close to Universal Analytics as possible… you can use Device-based. But honestly, the whole point of GA4 is to move beyond this device-only view.
Here’s the best part: this setting is not permanent. You can change it at any time, and it’s retroactive.
You can find the setting here: Admin > (under Data Display) > Reporting Identity
There will be an option called ‘Show All’ to have Device-Based appear.
It won’t break your reports or permanently alter your data.
You can switch from “Device-based” to “Blended,” check your reports, and switch right back if you want to.
This makes it completely safe to test.
I had this with one client whereby the reports showed a huge spike in Direct and New Users dropped dramatically. Switching it to device based changed that which demonstrated there was an issue with the user_id information being sent.
One final thing to keep in mind. If you use an identity that includes User ID (Blended or Observed), but you have a very low volume of signed-in users, your reports might be subject to data thresholding.
This is where GA4 hides some data to protect individual user privacy. It’s not a bug, but it does mean you need a reasonable amount of user activity for these identity methods to be fully effective.
Choosing the right reporting identity is a fundamental step in getting accurate, reliable data from GA4.
For most businesses that have a login area and have implemented Consent Mode v2, the Blended setting is the most powerful option.
It provides the most complete and accurate view of how your actual users—not just their devices—are interacting with your site.
Can I change my GA4 reporting identity?
Yes! You can change it at any time under Admin > Reporting Identity. The change is retroactive, so you can test different settings without breaking anything.
What is the difference between Observed and Blended in GA4?
The only difference is modelling. Observed uses User ID and Device ID. Blended uses User ID, Device ID, and data modelling to fill in gaps from users who decline cookies.
Do I need User ID to use the Blended setting?
No. Blended is a priority system. If a User ID isn’t available for a session, it will fall back to using the Device ID. However, the Blended model is most powerful when you are able to provide User IDs for your signed-in users.

Not sure where to look when you want to see revenue by channel? Find out the different approaches here.

Confused by sessions in Google Analytics 4? Get to grips with this often used, often misunderstood metric

Learn to use the GA4 Pages and Screens report to analyse engagement, filter content, and measure specific page performance.

Author
Hello, I'm Kyle Rushton McGregor!
I’m an experienced GA4 Specialist with a demonstrated history of working with Google Tag Manager and Looker Studio. I’m an international speaker who has trained 1000s of people on all things analytics.