A well-known Swiss e-commerce player and long-standing customer of Webrepublic has been using digital marketing for years to convince existing and potential new customers of its product range. Thanks to its digital marketing activities, the retailer collected valuable user behavior data on various digital channels. This data is now being used across platforms. For the first time, Webrepublic is testing an innovative targeting approach to optimize Facebook ad performance based on Google data. The potential of cross-platform targeting is huge: the costs per purchase are lower, and the return on ad spend (ROAS) is higher.

Download Case Study

Project overview

Initial Position

Due to their high reach and good performance, social media are an important part of the e-commerce merchant's marketing mix. The constant development of campaigns is important here. Thanks to many years of promoting products on social media and Google Ads, the company has access to large amounts of data on user behavior, which the e-commerce retailer wants to use profitably across digital channels to reach the right target group at the lowest possible cost.

 

Goals

  • Use existing user data of Google and Facebook efficiently across different digital channels
  • Reduce cost per purchase on Facebook and increase ROAS

Measures

Webrepublic's social media team wants to exploit cross-platform targeting capabilities and evaluate their performance. For this purpose, they are testing targeting users based on a combination of the Google Click Identifier (gclid) and Facebook pixel data. Due to the many purchases already made via Facebook Ads, the e-commerce merchant's ad manager is particularly suitable for this test. Thanks to the existing data, the Facebook algorithm reacts quickly to changes, which leads to clear test results. In cooperation with the SEA and Digital Analytics departments of Webrepublic, GDPR-compliant target groups are set up on Facebook based on gclid data.

 

The targeting setup in detail

With dynamic product ads on Facebook, the e-commerce merchant usually uses retargeting to address users who have placed a product in their shopping cart on the e-commerce merchant's website but left the website without making a purchase. Furthermore, potential new customers are addressed by means of a lookalike audience. The newly applied Gclid target groups are played against the existing target groups (based on the Facebook pixel). Thus, a possible performance difference at the end of the campaign can be evaluated. The targeting setup looks as follows: 

 

Test group A: 

  • Gclid target group vs. website target group based on the Facebook pixel (excluding the Gclid and buyer audience) 

Test group B: 

  • Gclid lookalike (website visitors and buyers of the Facebook Pixel are excluded) vs. lookalike of the website buyer based on the data of the Facebook website pixel (excluding the Gclid audience and their lookalikes).

To avoid overlapping of the data points, an exclusion procedure was applied to the respective target groups. The two comparisons within test group A and test group B should provide important insights into the use of the four different target groups. Thereby the following will be shown:

  • Whether the use of the target group based on Gclid data shows a better performance than the use of the website visitors based on Facebook pixel data.
  • Whether the Gclid lookalike target group performs better than the website buyers based on Facebook pixel data.

 

 

 

 

 

ABOUT GOOGLE CLICK IDENTIFIER

The Google Click Identifier (gclid) is appended to the URL after clicking on a Google Ad. This allows advertisers to use tracking tools, such as Google Analytics, to track which ad was clicked each time the site was visited. 

Results

Within test group A, it was found that the target group performed better based on the Gclid data: the target group based on Gclid data resulted in a higher ROAS and lower costs per purchase.

The test between the potential new customers (lookalike target groups) did not produce a winner. Although slightly lower costs per purchase were achieved with the lookalike target group based on Gclid data, the lookalike target group based on Facebook pixel data showed a slightly higher ROAS. These results show that by attracting new potential customers, the use of both target groups and therefore both data sources can be highly relevant, depending on which metric represents the campaign's KPI.

  • Reduced cost per purchase by 45%
  • Increased ROAS by 125% based on 1-day click attribution
  • A holistic approach to campaign design and target group definition is encouraged: Campaign results are considered the basis for future targeting opportunities beyond SEA and social media

The results are based on the 1-Day Click Facebook attribution window and were won onMarch 2 and 16, 2020. 

“Social media is a powerful performance channel. Such a targeting approach could be scaled across various other data sources. Gclid-Targeting is hopefully just the beginning and an inspiration on how we can exploit cross-platform targeting options on social media.”
Portrait Melissa Guentharft

Melissa Günthardt

Consultant Social Media Marketing, Webrepublic AG

SERVICES PROVIDED