Corbado Observe · Login Methods

Every method, measured the same way

Passkey, password and social, side by side on the same production traffic: how often each method is present, how often flows with it complete, what its error rate costs and how fast it really is.

• Presence, completion, errors and speed per method
• Each KPI vs flows without that method
• P50 and P90 duration, not just averages

The numbers behind the mix

Method decisions deserve method data

Rolling out passkeys, retiring SMS, adding a social provider: every method decision changes completion, errors and speed. When you want the causal version of this comparison, Method Uplift measures the lift.

One scorecard per method

Flow presence share, completion with the method, error rate and duration: the same four KPIs for passkey, password and social, so comparisons are apples to apples.

With vs without

Every KPI is shown against flows without that method: completion +2.4pp with passkeys, error rate +18.1pp with passwords. The delta is the argument.

Speed that survives scrutiny

P50 and P90 flow duration per method. A method that is fast at the median but slow at the tail is a support queue waiting to happen.

Login Methods FAQ

Common questions about per-method KPIs

What does flow presence share mean?

How is "completion with method" different from a success rate?

Which methods are covered?

Can I break the comparison down further?

Do I have to replace my IdP or auth stack?

Settle the method debate with production numbers

Which method completes, which one errors, which one is fast: measured on your own users, not on vendor benchmarks.

• Same-population comparisons, not cohort guesses
• Completion delta with vs without each method
• Read-only on top of your existing IdP