Yes, the issue is sporadic, affecting only some transactions but not others, which has made it challenging for us and PayPal to troubleshoot.
We and a few clients did some testing that -- at least as far as we could tell from our limited test-sample sizes at the times we tested -- seemed to suggest occurrence of the issue might correlate with the following conditions:
- Buyer chose to fund the payment from their PayPal account balance, rather than from a linked card or bank account;
...and/or... - Buyer clicked a link on the PayPal checkout site to view available shipping methods, but then neglected to manually select the sole shipping method available there*, yet PayPal let them go ahead and finish checkout anyway.
*Since we can already offer a choice of shipping methods (if the seller sets those up) in our cart before checkout, we only transmit the buyer's sole selected shipping method and rate to PayPal checkout, so there's only ever one method/rate available there.
That said, we don't know if those tentative findings would remain consistent across a larger volume of transactions, or if those tests just happened to coincide with PayPal tweaking things at their end that may have caused those results only temporarily, as they remain cagey about discussing any particulars whatsoever. We have at least reported those findings to PayPal in our MTS ticket about the issue.
Thank you for contacting PayPal Merchant Technical Support. We recently responded to your case. This is a courtesy email to inform you that your case will be closing shortly. However, you can keep the case open by replying with any new details or create a new ticket in the future.
Case Number: 10224177
Subject: Guest checkout deducts shipping cost for all guest buyers
Date Created: February 17, 2022
Received the following reply from PayPal Global Technical Support regarding my inquiry about this problem:
"Our Engineers are currently working on temporary measures to address the issue. The problem is twofold between eJunkie, and PayPal related to the way our Integration has changed over time and eJunkie not adapting to those changes as intended. I believe the plan is to roll back some of those changes, but I am still wanting on a confirmation on that situation."
As I read it, this blames e-Junkie.
My response so far has been to no longer use PayPal with the e-Junkie cart.
Interesting, this is the first we're hearing anything about "eJunkie not adapting to those changes as intended", assuming that's not just a buck-passing excuse. They certainly never informed us we needed to change or update anything, and our PayPal integration is still fully compliant with their current documentation for it.
@mbw - We'll followup on our own MTS ticket to inquire further about that with PayPal; what's your MTS case #, so we can reference that in our followup?
The case number is 10200744. I opened it on February 7, 2022
Here's the full email:
Dear M B Walker,
Our Engineers are currently working on temporary measures to address the issue. The problem is twofold between eJunkie, and PayPal related to the way our Integration has changed over time and eJunkie not adapting to those changes as intended. I believe the plan is to roll back some of those changes, but I am still wanting on a confirmation on that situation.
I will update you as soon as I know more. I apologize for any inconvenience.
Thank you,
Brian
Global Technical Support
PayPal, Inc.
Here's how PayPal MTS responded to my followup seeking clarification re: "adapting to changes as intended":
Our engineers have figured out the root cause of this issue that is caused by our system. They are currently working on a resolution to this issue. And I was not informed to have you make any changes in your integration. So you don't need to change anything at this time. If I have any further updates on the matter will let you know.
I followed up on our own MTS ticket to verify that this bug is indeed fixed. They replied Friday evening to clarify:
We have made the fix, but it is going out in the next release cycle which should be within the next week. We will restart the experiment and monitor after the release has finished to make sure things are working properly. Will let you know once it is done.