UNDERSTANDING GOOGLE ANALYTICS 4 (GA4) REPORTS WITH DATA-DRIVEN PROPERTY SETTINGS FOR YOUR AFFILIATE MARKETING ANALYSIS

In the past months, several comments have been made about Google’s new analytics product GA4 and its effect on Affiliate Marketing performance analysis. The general notion is that affiliate is devalued in GA4 and attribution is favoured for Google’s channels.

It doesn’t help anyone when conflicting and incorrect information is presented on large platforms.

Ishtvan Torpoi and Julian Roberto Henrichs decided to write this article to help advertisers and marketers understand what the data is telling them when trying to analyse affiliate marketing performance in GA4 with data-driven property settings.

By the end of this article, you will leave with clarity on this topic and actionable insights on how the reports can be used to your advantage.

We read the following statements online:

“Instead of using the Traffic Acquisition report, the All Channels report under Advertising is the one you should use.”

“DDA compares different conversion paths to infer the value of the channels within that path. To do so, DDA requires at least 400 conversions per unique path over 30 days.”

We’ve investigated these claims because we did not believe there was evidence backing them. So we put it to the test.

Are you ready for some investigative debunking? Let’s get on with it.

GOOD OLD UNIVERSAL ANALYTICS (UA)

Ask anyone in the affiliate industry and they will say they miss UA. Before GA4, we absolutely loved the Traffic Acquisition report in UA.

Here’s why:

UA primarily uses a last non-direct click model for attributing conversions in its Traffic Acquisition report, where the last known non-direct source gets the credit for the conversion.

As per Google’s documentation, there are two methods by which a session ends:

Time-based expiration:

• After 30 minutes of inactivity
• At midnight

Campaign change:

• If a user arrives via one campaign, leaves, and then comes back via a different campaign.

Reference: UA sessions

The Traffic Acquisition report in UA focuses on how users are brought to the site and which sessions led to a conversion.

And that’s the key difference here for this report; UA counts a new click in an existing session as a new session. GA4 does not and leaves the session open longer, regardless if other channels generated new clicks within the same session after it has already been initiated.

On top of that, as Google states: “By default, a session ends or times out after 30 minutes of user inactivity. There is no limit to how long a session can last.”

Read more here: Sessions

As explained by Google for the UA vs GA4 session differences: “UA counts only one key event per session, for the same goal. GA4 usually counts multiple key events per session, for the same key event. To reduce key event count differences between your UA property and corresponding GA4 property, update your GA4 key event counting method setting to Once per session.”

Reference: Session differences

Affiliate Marketing Specialists and Consultants used the Traffic Acquisition report to track the last click channels responsible for closing the conversion.

Good to know for further reference: Universal Analytics uses session-based reporting, whereas GA4 uses event-based reporting.

THAT SAME REPORT IN GA4

In Google Analytics 4 (GA4), when the property is set to use Data-Driven Attribution (DDA), the primary attribution model that powers the conversion and attribution reports indeed becomes DDA. However, the Traffic Acquisition report, particularly the "Session source/medium" section, operates differently.

Regardless of the attribution model selected for the property (including DDA), the Traffic Acquisition report in GA4 typically attributes sessions based on the last click (or last non-direct click). This means that for this specific report, the session source and medium reflect the last interaction that led to the session, not necessarily the touchpoint that would be credited with a conversion under the DDA model.

Confirmed by Google: “The session counts in your Google Analytics 4 property may be lower than the session counts in your Universal Analytics property. This is because Google Analytics 4 doesn't create a new session when the campaign source changes mid session, while Universal Analytics does create a new session under that circumstance.”

Reference: Difference with UA.

Additional information about the Traffic Acquisition report and session-scoped dimensions:

“Session-scoped dimensions show you where both new and returning users are coming from when they start new sessions.”

“Google Analytics surfaces information about session-scoped dimensions in the Traffic acquisition report.”

Purpose:?

The report shows you which channel (source/medium) users came from for a specific session, and if combined with an event dimension, it can show you which channel initiated a session where the user completed a purchase or transaction event.

Traffic Acquisition for Affiliate

The report can be helpful when marketers want to know about the channels that start a session that ends in conversions. It shows converting session initiators. How many and which of your affiliates are starting the final sessions that convert?

Advantages

It is a fresh perspective to the long dominating “Last click cookie” model that advertisers used for reporting.

The report focuses on the impact of a channel that made the user visit the shop in the first place for that final session. It comes very close to a “First click” model (for the last initiated session that results in a conversion), although the consumer journey could have many more preceding sessions, without any events or transactions.

Disadvantages

However, it completely neglects channels that followed these initiating sessions. If a converting session had various channel touchpoints, it attributes the revenue to the first channel.?

Therefore, this report does not tell you which channel was responsible for the last click resulting in a conversion.

OTHER GA4 REPORT TYPES

Since GA4 is focused more on event-based attribution, its reports tell a very different story than UA did. GA4 will provide insights into the value your affiliate program provides based on different events and engagement metrics, weighted versus the other channels involved in the conversion path leading up to a conversion.?

Confirmed by Google: “Universal Analytics hit types include page hits, event hits, ecommerce hits, and social interaction hits.

In contrast, Google Analytics 4 data is event-based, with the principle that any interaction can be captured as an event. As such, Universal Analytics property hit types translate to events in a Google Analytics 4 property.”

Reference: GA4 vs UA

User Acquisition

Even if the GA4 property is set to use Data-Driven Attribution (DDA) for conversion events, the User Acquisition Report still primarily focuses on identifying the immediate source or medium that led to the new user's first session. This report follows user-scoped attribution instead of session- or event-scoped attribution.

This report helps in understanding which marketing efforts are successful at bringing in new users, rather than attributing conversion credit across multiple touchpoints.?

Confirmed by Google: “The User acquisition report is a pre-made detail report that you can use to get insights into how new users find your website or app for the first time. The report differs from the Traffic acquisition report, which focuses on where new sessions came from, regardless of whether the user is new or returning.”

Reference: User Acquisition in GA4

User Acquisition for Affiliate

By applying this logic to your affiliate performance analysis, you should get a rather realistic picture about the volume of new users your affiliate publishers delivers. Results in this report can show huge discrepancies with affiliate network data based on the type of affiliate publishers you decide to work with.

Conversion Path

The Conversion Path Report in GA4 is a powerful tool for understanding the customer journey leading to conversions. Let’s briefly discuss its benefits.

The report details each interaction a user has on their way to converting, including organic searches, paid ads, social media engagements, direct visits, and more. This enables you to see which channels are most frequently involved in conversion paths and how they interact.

By analyzing the conversion paths, you can identify which channels and touchpoints are most influential in driving conversions, highlighting the roles of different marketing activities in the customer journey. This report is key to understanding where your affiliates join the party to make a conversion happen.

The report will also show how many interactions occur before a conversion and the time it takes for a user to convert after the initial interaction. This helps in understanding the complexity of the conversion process and the duration of the sales cycle.

Last but not least, it allows you to see how different marketing channels contribute collectively to conversions, providing insights into multi-channel synergy and the cumulative effect of your marketing efforts.

Google calls it the Attribution Path Report now, might be a bit confusing if you are used to Conversion Path Report (most accounts will still show Conversion Path). But you can access it through the Advertising section in the left hand menu. Then you select Conversion Paths to get the report. Don’t forget to change the default dimension from Default Channel Grouping to source/medium.

Reference: Conversion Path

Conversion Path for Affiliate

We use this report very often to understand where our affiliates enter and leave the customer journey. Do you have a coupon heavy affiliate program? Then the odds are extremely high that they will appear as last click just before conversion. Combine that with the amount of dedicated coupon code usage, and you will find that they don’t use as many as you’d have thought.

Last Click (LC) Model Comparison

This report gives you a specific perspective on how attributing conversions to the last interaction before a conversion compares to the more distributed approach of DDA. Be aware, Google changed the name of this report in its documentation to Attribution Model Report though the GA4 interface still shows Model Comparison at the time of writing.

As Google says: “Use the Attribution models report (formerly called "Model comparison") to compare how different attribution models impact the valuation of your marketing channels.”

Reference: Model Comparison reports

Here's what you learn from the Last Click Model Comparisonis report in that context:

The report highlights which channels or campaigns were the final touchpoints before a conversion, showing their direct impact on driving conversions. This is useful for understanding which efforts are effective at sealing the conversion.

It allows you to compare how conversion credit is allocated differently under Last Click Attribution versus DDA. While DDA distributes credit based on the assessed contribution of each touchpoint in the conversion path (up to 20 touchpoint in a path are assessed), Last Click gives all credit to the final touchpoint.

By comparing the two models, you can identify which channels are frequently the last in the conversion path, but also understand how these channels contribute to the conversion process earlier on or in conjunction with other channels, as reflected in DDA.

This comparison can reveal opportunities for optimizing marketing efforts. If a channel performs well under Last Click but not under DDA, it might be strong at closing sales but not as influential earlier in the customer journey, or vice versa.

LC Model Comparison for Affiliate

We use this report heavily to analyse our affiliate marketing last click performance versus other channels. It shows the same information we are used to seeing with the Universal Analytics Traffic Acquisition Report. It’s still not 100% accurate but it gets very close to the truth of what we used to look at in UA.

Data-driven (DDA) Model Comparison

This report provides insights into how the Data-Driven Attribution (DDA) model distributes credit for conversions across various customer touchpoints. Here's what it tells us:

DDA assesses all the touchpoints in the user's journey to a conversion, allocating credit based on the influence of each interaction. This report shows how much each channel contributes to conversions according to this model, providing a nuanced view of marketing effectiveness.

By comparing DDA with other attribution models (like Last Click), you can understand how the perceived value of different channels changes based on the attribution logic. Channels that seem less effective in a Last Click model might appear more valuable in DDA because of their role in the earlier or mid parts of the customer journey.

The report offers insights into which channels not only initiate customer interest, but also nurture and assist towards conversions. This can inform more balanced marketing strategies that invest in channels throughout the customer journey, not just those that convert at the end.

Understanding how different touchpoints contribute to conversions helps in optimizing marketing spend and efforts. Channels that are important at various stages of the customer journey can be identified and prioritized for investment.

Data-driven Model Comparison for Affiliate

This report will be a bit painful for a very large portion of existing affiliate marketing programs. It will show great discrepancies when directly compared to the Last Click Model Comparison report. That’s because the quality of those affiliate programs needs to be improved to start generating more valuable customer journey touchpoints.

Programs with a healthy distribution of valuable top, mid and lower funnel focused publishers will see a much higher conversion and revenue match rate in the DDA Model Comparison report versus their affiliate platform data or LC Model Comparison report.

In the same healthy affiliate program, the Data-Driven Model Comparison report should not show a discrepancy of more than 10%-15% compared to the Last Click Model Comparison report. See the below image for reference from a large Belgian retailer that Ishtvan manages.

The medium was tagged differently for a distinction between the different networks. We can see the network defined as befr-alwayson has a much lower impact on customer journey touchpoints than the other two. The main network tagged as affiliate delivers the best value in customer journey touchpoints.

OUTCOME OF OUR TEST

Now that you understand what the key report types do and how to use them for your affiliate marketing analysis, we get to the main part of our research; debunking the two claims made by industry veterans.

The following path was followed to test the two claims:

Let’s go back to the claims and see what we found.

Claim 1

“Instead of using the Traffic Acquisition report, the All Channels report under Advertising is the one you should use”

We agree that the traffic acquisition report is not suitable to assess the performance of a channel based on last click. However, the claim that the All Channels report in Advertising is the right one is not correct for data-driven property settings. It shows data based on data-driven attribution. The image below shows decimals in the conversion volume and revenue, which means data-driven attribution was applied.?

This is not the correct report to view Last Click Attribution data in a DDA property.

The correct report to view true Last Click Attribution would be the Last Click Model Comparison report. See below. The full conversion was attributed correctly based on the last known channel before the conversion took place.

Claim 2

“DDA compares different conversion paths to infer the value of the channels within that path. To do so, DDA requires at least 400 conversions per unique path over 30 days”

We do not see this in our client’s GA4 accounts, and Google’s documentation does not support this claim either. See below.

It clearly states that for data driven attribution to be applied in the account, that a minimum of 400 conversions per conversion type are required for the model to function.?

Not 400 conversions per unique path as claimed.

Furthermore, it is claimed that approx. 98 % of all conversions with affiliate touchpoints would be excluded as per this methodology.?

We do not see this in our GA4 data for any of our clients. We can also confirm with our tested conversion path that the affiliate conversion does not get excluded from DDA modelling even though we do not have 400 conversions with this conversion path in the last 30 days.

Session initiated: google / cpc

https://www.pharmamarket.be/be_nl/?gclid=Cj0KCQjw5cOwBhCiARIsAJ5njuaF1pkMKP5ojb_uVQ1nqDTr4SxuHGAqaGnhAxRWVfitinAQuztItUYaAltpEALw_wcB

• add to cart from the homepage

In the same session, new click: affiliate

https://www.pharmamarket.be/be_nl/utm_source=daisycon_Daisycon.com+%28test%29&utm_medium=affiliate_test&utm_campaign=test

• open cart and complete purchase

A unique conversion path with only 2 touchpoints and 1 conversion (or 2 conversion if you count events as conversions).

The same conversion in the data-driven attribution model comparison report.

In theory, if the affiliate program consists of a well balanced publisher mix, it can be positively represented in the dynamic attribution, even being attributed more conversions in dynamic attribution vs. in the last click model.

So the claim that GA4 misrepresents Affiliate conversions in its dynamic attribution model is unfounded.

Also, the same affiliate network with Claim 2 criticized that Google favours its own Google channels (Display, Video and Search) in the Dynamic Attribution Model, as they are “high volume channels”.

We did not find evidence to support this claim. Volume is not the only factor in GA4’s dynamic attribution model.? The algorithm rather looks at the value of each channel by registering each touchpoint of the consumer journey with the events triggered during each touchpoint. Affiliate’s representation should be just “fine” in a well-balanced program as we also saw in our test.

KEY TAKEAWAYS AND ACTIONABLE STEPS

Align with clients and advertisers what the reports show and how the value of a touchpoint should be determined.

Use first-party tracking to minimise data discrepancies and fraud. Data discrepancy can also be caused by using third-party cookies, and fraud is minimised by using the more secure first-party tracking.

Discuss attribution models and affiliate payouts: first click, last click, dynamic, U shape, etc.

• The cookie methodology of the affiliate tracking should be as close to the GA4 data as possible if you look at the Last Click Model Comparison report. Most networks don’t track other channel touchpoints but the LC Model Comparison should show a much fairer last click attribution picture based on what you are used to with Universal Analytics.
• If you notice a lot of conversions in the affiliate platform but much less in any other reports in GA4, it’s time to evaluate the added value of your publisher base. Them being the last click in LC Model Comparison does not mean that they add a lot of value in the entire customer journey, and the other reports will paint that picture for you.
• Look at the attribution model regularly to see outliers and pay publishers accordingly. If affiliates outperform other channels in the DDA based reports, pay bonuses for affiliates via the affiliate network. If publishers perform worse in DDA based reports, you should adjust commission payouts accordingly based on their value. You might also want to remove the ones that add little value to the customer journeys.

Best report combination to use for your affiliate performance analysis

1. Conversion Path

Analyse your publisher's involvement in the customer journey.

2. User Acquisition

Analyse the amount of new users your publishers bring in.

3. Traffic Acquisition

Analyse how many sessions your publishers start which end up in a conversion.

4. Model Comparison

Analyse the difference between last click and data-driven attribution for your publishers' conversions.