W3C

Web Payments Working Group

15 February 2024

Attendees

Present
Anne Pouillard (Worldline), Bastien Latge (EMVCo), David Benoit, Doug Fisher (Visa), Fahad Saleem (Mastercard), Gerhard Oosthuizen (Entersekt), Grégoire Leleux (Worldline), Ian Jacobs (W3C), Jean Luc Di Manno (FIME), Jeff Owenson (Discover), Juliana Cafik (Microsoft), Kenneth Diaz (Modirum), Melissa Sebastian (Modirum), Michael Horne (American Express), Nako Siskov (Netcetera), Nick Burris (Google), Praveena Subrahmanyam (Airbnb), Rolf Lindemann (NokNok Labs), Rouslan Solomakhin (Google), Sameer Tare (Mastercard), Stephen McGruer (Google), Steve Cole (MAG), Tomasz Blachowicz (Mastercard), Tony England (Visa)
Regrets
NickTR
Chair
Praveena
Scribe
Ian

Meeting minutes

SPC user journey improvements

Stephen McGruer slides

smcgruer_[EST]: Thanks to our designer David! This is draft, we hope to understand underlying needs through this exercise.

[Stephen walks through slides]

Stephen: In the slide we refer to "issuer"; is that too card-specific? Is there are more general term we could use?
… we have not yet figured out a way to put the RP origin in the dialog. The concern is that it might not make sense to the user.
… in many cases the RP origins are opaque.
… and sometimes the service is provided by an entity who is not the bank, for example.

Tomasz: Thanks for sharing this. Definitely an improvement.
… good that there is a change in the language. I have doubts about whether the end user will understand who the issuer is.
… I think "Issuer" is not the write term (and many not translate well into other languages)
… in terms of showing the RP origin, the chances the user will understand are even slimmer.
… but maybe a domain name can go into the sentence.
… I agree that "credentials verified" is not the best term. Perhaps "Authentication completed"

Gerhard: Context would be valuable. Could we have a display context. LIke "This is a card payment" or "This is a bank payment."

"This is an open banking payment"

This might help us solve the generic problem.

Stephen: I'm not opposed to having context information provided (in a controlled fashion).

dougf: +1 to the direction of the new design
… in terms of important additions, one of the things that we've found in our pilot is people not being clear who is authenticating them.
… it will need more discussion how to communicate who is authenticate.

IJ: Is it a deeper challenge understanding what's going on?

dougf: The question is when you create a credential, with what entity and how it will be used. Can the user understand why an entity is able to do an authentication. Understanding the process is difficult.
… in our pilot we found perhaps an equal split between people thinking its the issuer or merchant or browser.
… the browser could help, even if we don't know how exactly yet.
… the successful authentication message is MOST important at credential creation time.
… during authentication the success was clear to people; so we found more need at creation time.

Tony: Is there an opportunity to pull in a single icon (e.g., bank + network together).
… the customer might recognize particular imagery.

smcgruer_[EST]: The card icon is up to the caller. It's a good note to us that this string might be long. It interests me that there might be 2 icons combined into one.

Tony: That's the case in the token space.
… from a customer experience perspective they just want to know they are using the correct product.

Fahad: Are these icon fields (going to be) optional?

smcgruer_[EST]: This is an open question. What do we need to add to make this work?
… Gerhard's proposal is to provide a "context". We were thinking that kind of bundle.
… but there are combinatorial complexity issues.

Fahad: The new UX makes a lot of sense for a 3DS flow.
… but in Tony's use case (which is also pertinent), we could skip issuer and have a composite icon

smcgruer_[EST]: That scenario should be easy to support.
… I'd love to see if people have example UIs of other payment flows. What are people showing users today?

Ian: Check out Open Banking Customer Experience Guidelines

Gerhard: Also Brazil open banking has UX guidance

tomasz: It would be good for the RP to indicate who is doing the authentication.
… the sentence in the UX could say "citibank has requested to authenticate you to confirm your purchase."
… this is the valuable context.
… this could be done generically (not just 3DS)
… eg., the cardholder could be authenticating to the merchant (or their service)
… so "Store.com wants to authenticate you"

smcgruer_[EST]: We are open an interested in putting the origin of the RP. That is the more trusted string.

smcgruer_[EST]: The credential is exactly for the RPID.

Juliana: There are guidelines for 3DS with clear language definitions and so on
… that consumers are familiar with, and the closer we stick to that, the better.
… that would include difference between authentication and authorization.

<Gerhard> question: Could the Common Name of the URL in the HTTPS certificate be used?

smcgruer_[EST]: We rely on origin because other strings might be lies.

Ian: Maybe show both: the string and the origin

[Moving on to the fallback screen]

smcgruer_[EST]: Reminder - this fallback screen is important for privacy protection.
… but nobody likes this screen.

smcgruer_[EST]: Ian presented some ideas a few months ago. We run with that here -- when no credentials are found, the dialog looks the same. And "Continue" sends a signal to the caller that further authentication is required.

Gerhard: Yes, I agree. :) The button intention could be "Verify" rather than "Continue"
… in my slides (for next meeting) I propose some additional language

smcgruer_[EST]: Also in this proposal there are three signals through the API (instead of 2 signals in the current approach).

The signals are: (1) continue with FIDO (2) cancel (3) try another way

smcgruer_[EST]: we are also looking at an idea where, if you choose to authenticate another way, what should happen with the SPC dialog? Should it say "Verifying" and stick around, for example?
… this is an interesting idea for me, but I think this will be challenging if 3DS is happening in an iframe and therefore appears behind the SPC dialog

smcgruer_[EST]: I like this idea but it may not work and it may be preferable for the browser to "just get out of the way"

smcgruer_[EST]: We have some questions (in the deck).

* "Store" field. Is it always a store? Is "merchant" more common? should there be an enumeration?

smcgruer_[EST]: Regarding the needs for icons to be larger. Is this a regulatory requirement? is this for "glanceability"?

smcgruer_[EST]: We don't want to tie too closely to card payments. I like Gerhard's idea of providing context

Gerhard: Is Netflix a merchant? Is a gym a merchant? There is no universal word.
… some systems say "pay" "charge" "subscribe"
… enumeration might help
… merchant might like to choose how they would like to be called.
… regarding "icons" they are not really icons, they are "visual representations of cards"
… for recognizability.

smcgruer_[EST]: Autofill has customized card art and we've seen users appreciate the detailed identification of cards. SPC should support that level of detail.
… we are more interested in the size and positioning of card beyond the card art

dougf: The card art increases trust. Fraudsters are not perceived to have that kind of information. It needs also to be large enough for accessibility reasons.

IJ: What about hover or click to enlarge?

smcgruer_[EST]: That idea was not well-received.

smcgruer_[EST]: I agree for custom card art. I'm mostly focused on custom bank / network logo.

tomsasz: The reason for the issuer logo (in 3DS land) is to let the user identify the card issuer. This helps identify which card.

tomasz: It's valuable to bring SPC closer to 3DS screen (for 3DS flows)

dougf: We don't have a specific UI for credential creation; we might potentially want good practices.
… but when there is continuity of experience, there's greater association in terms of who is authenticating.

smcgruer_[EST]: Maybe creation screen could add confidence if done right.

Gregoire: One thing bothers me a bit, the details on the transaction. There's an increasing need for precise information for some types of transaction (e.g., recurrence)
… I am afraid that this could be a limiting factor for markets where precision is required

smcgruer_[EST]: I agree. I am worried that browsers may not be able to keep up with all regulatory requirements ...
… having said that, I am more optimistic about some spaces where EMVCo has already done work on this!
… therefore I'm more optimistic we can find a way to address this without taking arbitrary strings as input.

Next meeting

29 February

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).