How To Fix Historical PayPal Referral Problems in Google Analytics – With A Custom Attribution Model

by Ed Finn

Yes, you read the title right. If you landed here then you may have been looking for a solution to the often blogged about “Paypal Referral Exclusion Fix”. What you’ve probably found so far is a common treatment for a common problem: Default reporting shows a disproportionately large amount of revenue attributed to the source/medium ‘paypal.com / referral’ and the common solution is “Add ‘paypal.com’ to the referral exclusion list”. But, as I’m sure you’ve realized by now, this isn’t a complete solution – your historical data is still impacted by the issue.

The PayPal Referral Exclusion Problem is Two Problems, Two Solutions:

  • Problem 1: Data being generated right now is impacted by this issue
  • Problem 2: Historical conversion data is rendered useless because of the credit stolen by PayPal
    • Solution 2: Keep reading, my dude.

By the time you finish reading this post you will be able to access a report in Google Analytics that shows you which channel closed the revenue and sale during the historical period where referrals from PayPal impacted your data. If you have creativity and an analytical mind, you will realize you really only need baseline consistency in tracking to get value out of attribution, in this situation and beyond.

Custom Attribution Model Template For Historical PayPal Referral Fix

We are using the Model Comparison Tool report in Google Analytics with a Custom Attribution Model to override instances where paypal.com “stole” credit from another channel, and fallback that success to the previous online channel. Google Analytics does something similar with its Default Attribution Model, overriding instances where the closing sale came from a direct visit – which uses a model known as ‘Last Non-Direct Click’.

More specifically, we want to apply the Default Attribution rules to traffic (source / medium) that showed up as ‘paypal.com / referral’. This moves credit for that revenue and transaction count to the channel, or source/medium that was present in the preceding session..

None of that makes sense to me …

Don’t worry. Steps and a template link are below to get you started.

What Are The Steps?

1. Open Google Analytics

2. Follow this template link and save the shared attribution model. Give it a name like “Last Non-Direct Interaction w Paypal.com Referral Override” → https://analytics.google.com/analytics/web/template?uid=4XpI6w0wQFGHxQACKU0dYA

3. Go to Conversions > Attribution > Model Comparison Tool. Select Only ‘Transaction’ From the top left dropdown:

4. Open the Attribution model dropdown and select the custom attribution model:

5. Et Voila! The attribution model is ready to roll. Isolate for the date range where PayPal (or another referrer) impacted your data and apply the attribution model. That credit will now be applied across the channels that typically deserve the credit post referral exclusion.

Note: I wanted to share my screenshots proving this method, but the Google Data didn’t show this problem occurring and I can’t share client data. You’ll just have to try it on your own data!

Wow, THAT WORKS! What Else Can I Do With Custom Attribution Models?

This is a single simple use for a complex and capable attribution tool that Google gives away for free. Very few Google Analytics users have explored the full capability of custom models.

Unfortunately there are not a lot of great resources on leveraging this tool, but if you want to know more you can reach me and my team by contacting us today.

Ed Finn

Senior Practitioner, Analysis

Ed’s passion is turning data into knowledge. As a Senior Practitioner on Napkyn's Analyst Team, Ed is responsible for driving the strategic direction behind some of our largest and most complex programs.

See more posts from Ed