#[LBG] India - connect

1 messages ยท Page 1 of 1 (latest)

vale minnow
#

Hello. One moment

#

Have you tried this in test mode?

vernal plinth
#

No

vale minnow
#

I believe it should work. Please try it in test mode first

vernal plinth
#

OK let me try

vernal plinth
#

Even when the Express account is rejected, I am still asked the password for the associated mail during onboarding for standard

#

Should I wait a little ?

vale minnow
vernal plinth
#

Seems to work ๐Ÿ™‚

#

I'll try on live (with partners account)

#

Are you able to see my live objects if I send you one ? I have questions on rejected subscription payments

#

And payment_intents

vale minnow
#

Yes. What's the object and question?

vernal plinth
#

A few weeks ago, we had problems with Subscriptions. We saw our churn explode because of unpaid subscriptions, 90% with "insufficient funds" bank reject messages. All our metrics went down, MRR, active subscriptions, etc. And the churn exploded. The best guess we did was the SCA, because we weren't SCA compliant. We operate in EU and were credit cards objects and not payment_methods that can be authenticated by the card issuer. Nothing was indicating that it would be SCA though, from reports like this one : https://dashboard.stripe.com/authentication?startDate=2021-12-13&endDate=2022-03-13 were not indicating SCA attempts at all. We bet on the hypothesis that as the card object itself can not be authenticated, the card issuer would never try to ask, therefore sending "default" reject messages. So we went SCA compliant creating payment_methods with off_session set to true. All our new payment_methods should now be SCA compliant. The same report shared upper shows that it has a positive effect. In fact, all our metrics are green, MRR going up again, active subscriptions as well, churn going down again. All good, but this makes me think that the bank reject messages are not reliable at all !

acoustic bay
#

Hey @vernal plinth, I'm stepping in here fro @vale minnow as he needed to step away. Can you share a specific subscription that saw one of these errors?

vernal plinth
#

I have a list of unpaid Subscriptions with a payment_method that should be authenticated, is there a way to know more about this ? Some samples (pi_3KbErSLJLkeHRLt11qLSJCp2, pi_3KcVxBLJLkeHRLt10zDce5Uz, pi_3KcSXMLJLkeHRLt11yaxjAOD,pi_3KcIWoLJLkeHRLt11zpIFO5N,pi_3KbV4RLJLkeHRLt11kDfXc3v,pi_3KcyBqLJLkeHRLt10DYPxP0M)

#

Hi @acoustic bay, sorry for the long messages, I want to give as much info as I can to be efficient

acoustic bay
#

No that's great, thanks so much.

#

Taking a look at one of the PaymentIntents

vernal plinth
#

ok

acoustic bay
#

So initially, SCA definitely could have been the issue. However, I can see that you are setting up these PaymentMethods with a SetupIntent now, so you are doing everything right.

#

Insufficient Funds is one of the more common decline codes that an issuer uses.

#

Nothing is jumping out to me as being wrong from your end.

vernal plinth
#

Is there a way to dig further in this ?

acoustic bay
#

Yeah, I'd contact Support about this. They can help with decline analysis.

vernal plinth
#

We don't have a rich target that is true but I am persuaded banks are blocking for unknown reasons

#

Speaking about setupintent

acoustic bay
#

It is possible. Oftentimes the best way to get more info is actually to ask your cardholders to contact their banks as that is the only way to get the true reasoning from the issuer for why they declined.

vernal plinth
#

Ok

#

I'll contact support to know more about this, but I don't think they really helped when I asked about it last time

#

About setup intents, I noticed something. For every payment_method added, there is 2 setup intents canceled a few minutes after the one that succeeded

#

I found this by looking at fingerprints logs, some sample : oq03cc1AKCRtndU4, mzkeEqVVcRogLfHQ, xdT5LZmml0AAFFQf

#

Those fingerprints are source of paid but also unpaid subscriptions

acoustic bay
#

Can you provide the SetupIntent IDs for those 3?

#

I'll look at the canceled ones to see if I can see anything

vernal plinth
#

seti_1KbnfFLJLkeHRLt1WAGCiUBG, seti_1Kc7jDLJLkeHRLt1gdtSp9ZQ, seti_1Ka8N2LJLkeHRLt1bPtPydFl

#

That's the 3 unique succeeded ones

acoustic bay
#

Yeah so all 3 of those look to have successfully set up 3DS. This doesn't guarantee that the card won't decline for a different reason upon payment though.

#

Sorry, I'm still not seeing anything here that doesn't just look like normal bank declines to me...

vernal plinth
#

How can you tell that 3DS has been asked from the logs of the setupIntent ?

#

From this event ? evt_1KbnidLJLkeHRLt1i9GECHQz

#

"status": "requires_action" and three_d_secure_2_source keys maybe ?

acoustic bay
#

Yep exactly, if it has a status: requires_action then that indicates 3DS is required.

vernal plinth
#

About the setupIntents, why a successful seti (seti_1KbnfFLJLkeHRLt1WAGCiUBG) is always followed by two seti canceled (seti_1KbnjNLJLkeHRLt1NUJxH3WM, seti_1KbnjOLJLkeHRLt1HQDBPWbE)

acoustic bay
#

Ah okay. You are using a trial period.

#

With a trial period we provide you a SetupIntent upon Subscription creation.

#

Take a look at req_1a9sCf5FGVePoR

#

You can see we return you a pending_setup_intent

vernal plinth
#

Ok so that is normal behavior ?

#

The first seti that is successful is to save the payment_method, and the others for subscriptions, but why 2 ?

#

We only create 1 sub

acoustic bay
#

Oh interesting. Didn't notice those were from the same Sub creation.

#

Let me look more

vernal plinth
#

Yeah it's not that obvious to link everything for a same customer

#

Kind of stumbled upon it by chance

acoustic bay
#

Alright so this does look to be a bug.

#

We shouldn't be creating multiple SetupIntents here.

vernal plinth
#

How so ?

#

is that an error on our integration ?

acoustic bay
#

There should only be one pending_setup_intent created upon the Subscription creation.

#

Nope, nothing wrong on your end.

#

And no real impact to you.

#

Just extra SetupIntents on your account.

vernal plinth
#

Coming back

acoustic bay
#

We are filing internally to get this bug fixed up. But yeah, shouldn't actually have any impact.