How Can Incrementality Testing Aid GA Reporting?
The permanent move away from UA adds another layer of complexity to marketers’ ability to measure return on investment (ROI) and the effectiveness of their campaigns. But, while this is further disruption that marketers could do without, it’s not all bad news.
Topics
![](https://martechvibe.com/wp-content/uploads/2024/04/How-Can-Incrementality-Testing-Aid-GA-Reporting.jpg)
It’s hardly breaking news to hear that marketers are having to contend with one of the most challenging years they’ve had to face, thanks to the deprecation of third-party cookies in Google’s Chrome browser. What many may not realise, however, is that marketers are having to overcome another significant Google-based hurdle: the transition from Google Analytics 3 (GA3) to Google Analytics 4 (GA4).
GA3 – also known as Universal Analytics (UA) – has been slowly phased out since July 2023, when standard UA properties stopped processing new data. Fast forward 12 months and, on July 1st 2024, Google will force the remainder of the approximately 28.1 million websites that use the platform to migrate to GA4.
The permanent move away from UA adds another layer of complexity to marketers’ ability to measure return on investment (ROI) and the effectiveness of their campaigns. But, while this is further disruption that marketers could do without, it’s not all bad news.
Google’s big switch
Being forced away from UA to GA4 will deliver the biggest impact to marketers around measurement and attribution.
Having previously used a measurement model based on sessions and pageviews, GA4 flips the script by focusing on events and parameters. Meanwhile, the GA4 platform introduces a “data-driven attribution” option alongside the trusty “last click”.
Data-driven attribution uses machine learning to analyse how much of an impact each touchpoint is likely to have had in driving conversions. While this model can be seen as a black box, coupled with the changes to how sessions are defined, it has the potential to greatly reduce attribution fraud.
However, this creates an Analytics ecosystem that no longer relies on the pre-packaged reporting style seen in UA, instead shifting to a more complex, customisable “Explore” framework. Such a shift may delight data analytics teams, but presents a steep learning curve for marketers, particularly impacting those at smaller companies.
Thankfully, there is a solution to the problem – namely through the use of incrementality testing alongside standard GA reporting.
Also Read: The Future with Google Analytics 4
Testing for success
The advertisers that are impacted the least by the changes will be those that don’t try to overcomplicate things too much. They’re advised to stick to their current attribution model, especially in the case of the tried-and-true “last click” approach. And this model should be paired with methodologically correct incrementality testing with their current vendors.
Incrementality testing involves running an ad campaign in a similar way to a medical trial. The target audience is divided into two groups – one group is exposed to the specific activity being measured, while the other acts as a control group and doesn’t encounter this activity.
With this type of testing, it’s important that both groups have similar exposure to all other marketing activities, aside from the measured activity. This enables marketers to accurately assess how much of an impact the specific activity has had, making it particularly useful for measuring the effectiveness of lower-funnel campaigns.
For example, luxury retailer Neiman Marcus wanted to increase incremental ROAS (iROAS) while maximising the scale of transaction events. RTB House was able to show the impact of its retargeting solution by utilising an incrementality test to measure lift and the overall impact of the campaign that was run.
Incrementality testing was able to prove that the campaign generated a 65% higher iROAS than the target, a 26% net uplift in incremental sales revenue, and a 20% increase in conversion rate.
This approach creates a clearer picture of whether the vendor has been successful in meeting the agreed-upon KPIs. And enables the advertiser to establish whether certain tactics deliver truly incremental results, or just transfer conversions from one channel to another.
Advertisers can get started with incrementality testing by first reaching out to all of their current vendors and asking about their ability to carry out incrementality testing. The best vendors will conduct their incrementality tests using a blind random split with precise measurement.