Internet identity version 2, bad UX or only my issue?

Developers internet-identity
1

Hi everyone!

The recent Internet Identity upgrade, while introducing a great OpenID integration, has a significant UX problem that negatively impacts integration.

The Problem: A Confusing Extra Step

Imagine you’ve successfully logged into a website using Google. You close your browser and come back later. You expect to be signed in automatically or, at worst, to just click “Log in with Google” again.

But with the new II, this is what happens:

  1. First Login (as a new user): You see a nice, simple button to “Log in with Google.” You click it, do the Google login, and you’re in. Example - first time login

  2. Next Login (as a returning user): The system has saved a “login state” in your browser. The next time you visit, it doesn’t show you the “Log in with Google” button. Instead, it shows a new screen asking you to select an “account.” Example 1 - google login selected , Example 2 - stateful login

  3. The Confusion: As a user, you don’t know what this “account” is. It doesn’t show your email address or the Google logo. You just see a strange, generic option. This creates doubt and makes you wonder if you’re on the right page or if you’ve done something wrong. The experience is no longer seamless.

This happens because the system is designed with a “stateful” user journey in mind, where it assumes you want to manage multiple “accounts.” This might be a good feature for advanced ICP users, but it’s a major roadblock for the average person who just wants to log in with their familiar social media account.

The Solution: A Seamless, Stateless Login Flow

A better design would prioritize the user’s intent and remove the unnecessary “account” selection step. The login should be stateless, meaning it doesn’t rely on information saved in your browser from a previous visit to determine the next step.

Here is the proposed solution for a better user experience:

  1. The Initial Screen: The first screen should always show the user a simple, clear list of all available login methods: “Log in with Google,” “Log in with a Passkey,” etc. This is the login method-first approach that other providers like Web3Auth use. Example

  2. Smart Recognition (Optional): The system could subtly use local storage to remember the last login method used, but it shouldn’t force a user action based on it. For example, it could highlight “Log in with Google” as a recommended option or even show a button that says “Continue with Google” to speed things up.

  3. No Forced “Account” Selection: The “account” selection screen should not appear in the primary login flow. Instead, it should be a separate, advanced feature for users who want to manage their identity, accessible from a settings page after they have successfully logged in. This keeps the core login experience clean and simple for everyone.

By making the login flow stateless and removing the forced account selection, you create an experience that feels natural and trustworthy. It aligns with what people expect from modern websites and removes a key barrier to adoption for users unfamiliar with the Internet Computer ecosystem.

4 Likes
2

Thanks for the feedback @baolongt !

We saw a similar feedback in our usability test that was performed during August.

We are planning on improving this experience and I will pass your feedback to the designer.

Thanks!

5 Likes
3

User feedback is a highly valuable asset. Shouldn’t you be organizing some reward programs for this? Don’t wait until there are no users left — by then, it will be too late. The key point is that your product was already like this when it first came out, for example SNS

4

Rewards program for expressing opinions? That might get very popular. :smile:

Talking seriously, as Llorenç already shared, we did usability testing and are now actively working on the findings. I’m happy to consider you as a participant when we do it again.

Thank you for the advice. My career has taught me that creating products that users love is harder than it sounds. Should you have some from the trenches experience we’d love to hear that.

5

I don’t know if you’re a new employee. To be honest, as a user and investor who has actively supported the Dfinity Foundation, I’ve participated in many projects and tests. It’s been almost five years, and now the situation is like this: the foundation as a whole lacks capability yet remains arrogant, the forum is controlled by certain people, and most of the ecosystem projects are practically dead. After such a long time of observing and experiencing this firsthand, I no longer even have the interest to try out II 2.0. So if there are still people willing to test and give feedback, you really ought to value that.

I’ve already given up struggling; all I can do now is pray that, with God’s help, ICP will have a chance to rise again.