3D Secure challenge completion time statistics
https://checkoutcloser.com/guides/3d-secure-challenge-completion-time-statistics/
## Quick answer
3D Secure challenge completion time statistics depend on which flows finish. Evervault's January to June 2026 platform snapshot reports a global median of 25.35 seconds from challenge presentation to the final result, measured on completed flows. Selected country medians range from 20.90 seconds in the United States to 35.32 seconds in Spain (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
When authentication feels slow, the useful question is where the delay occurs and which flows the timer includes. This analysis compares the published times, distinguishes completed challenges from all initiated attempts and follows authentication through to the payment outcome. These platform observations are context for an investigation, rather than a forecast of an individual store's performance.
## Key Takeaways
The global completed-challenge median was 25.35 seconds. The platform snapshot covers customer initiated 3DS sessions from January through June 2026 (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
Spain's selected-market median was 14.42 seconds above the US value. Published medians were 35.32 and 20.90 seconds respectively, a descriptive comparison rather than a causal country effect (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
A completed challenge is not a completed payment. Stripe describes successful authentication followed by an attempt to charge the customer, with payment outcome recorded separately (Stripe documentation (https://docs.stripe.com/payments/3d-secure/authentication-flow), Nov 2026).
## What does the challenge clock actually measure?
The clock starts when the challenge is presented, rather than when the shopper first reaches checkout. It ends at the final result among completed challenge flows. It therefore describes a specific part of the journey and a selected population.
Figure description: 3D Secure challenge completion time statistics scatterplot: US challenge share 7.3% and completed median 20.9 s, India 94.25% and 22.51 s, Spain 47.94% and 35.32 s, showing exposure and duration answer different questions.
Figure 1: Similar completed-flow speed can hide very different exposure. Source: Evervault, country comparison (https://evervault.com/3ds-performance), Jan to Jun 2026. Original explanation and arithmetic from the cited primary research.
### How should a merchant read the median?
Evervault defines the time metric as the median elapsed seconds from challenge presentation to the final result, measured on completed flows. The source does not describe it as the time required for all shoppers to complete authentication (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
A median splits the observed completed flows around the middle of their time distribution. It does not identify the slow tail, the share abandoning, or how long a shopper who never returns remained stuck. A store can have a short completed-flow median while losing some slower shoppers before they produce a completed result.
This matters when evaluating an improvement. If a change causes slow challenges to disappear from the completed population, the reported median may fall without the experience improving. A valid diagnosis needs the attempt count and each terminal status alongside time distributions.
### What is outside the time measurement?
The metric excludes time spent choosing a payment method, entering card details or correcting a cart. It also does not cover the entire authorization and settlement sequence. Stripe's flow documentation explicitly separates the authentication result from the subsequent charge and final payment status (Stripe documentation (https://docs.stripe.com/payments/3d-secure/authentication-flow), Nov 2026).
An authentication screen can close successfully while the charge later fails. Conversely, an unfinished challenge may reflect a technical interruption, a shopper switching apps or a deliberate decision to stop. The elapsed-time statistic alone cannot assign those causes.
## How different were the observed market times?
The selected market medians show a meaningful time difference while remaining descriptive. They are platform observations with different underlying traffic, rather than randomized treatments assigned by country.
### What does a direct comparison show?
The United States median was 20.90 seconds, Germany 21.95 seconds, the United Kingdom 29.16 seconds and Spain 35.32 seconds in the same January to June 2026 dashboard snapshot. Spain's value is 14.42 seconds above the US value and approximately 1.69 times as large. The UK value is 8.26 seconds above the US value (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
These are differences and ratios of the source's rounded medians, calculated here. They are not differences in average checkout duration, customer patience or a causal penalty for serving a market. A mixture of issuers, authentication methods, devices and customer behavior can affect the observed times.
The CSV retains all eleven country rows displayed by the source, rather than only the four highlighted in the figure. Other published medians include 26.32 seconds for Canada, 22.58 seconds for France and 34.50 seconds for Brazil. The global median of 25.35 seconds comes from the platform's broader population, not the simple average of the displayed countries (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
### Why not calculate a regional average from these values?
An average of country medians would give each displayed country the same weight regardless of volume, and would still not recover the median or average of individual challenge times. A median is not additive. Without the underlying flows and a valid aggregation rule, a new Europe-versus-America timing statistic would be invented precision.
The dashboard also excludes some observations when data are incomplete, volumes are low or outliers heavily distort results. Its published scope should travel with the numbers. The selected markets are not a population census of every DTC checkout or every authentication provider (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
## How should a team use the timing evidence?
Use the observed time as a diagnostic reference alongside the completion funnel. A longer challenge path can merit investigation, but a global benchmark does not decide the store's timeout or prove that a particular issuer is at fault.
### What should the local review record?
Record challenge presentation, final authentication result and final payment result against the same permitted transaction identifier. Keep a visible distinction between successful, failed, abandoned and technically interrupted flows. Inspect timing among completed flows and the tail of the time distribution, then examine the relevant noncompletion statuses.
For a phone journey, test the route out to a bank application and back to the storefront. Confirm that the cart and payment state remain recoverable and that a delayed return does not create a duplicate payment attempt. Stripe describes failed authentication returning to a state where another method or a reconfirmed authentication attempt may be needed (Stripe documentation (https://docs.stripe.com/payments/3d-secure/authentication-flow), Nov 2026).
The merchant should choose a recovery policy around its integration and provider requirements. A median of 25 seconds is not evidence that a 25-second timeout is safe. A cutoff at the middle of completed observations could interrupt legitimate flows that take longer.
### Which result would challenge the apparent improvement?
Suppose completed-challenge times fall while abandonment and failed payments increase. The shorter median would be a poor standalone success claim. Equally, a longer completed-flow median may accompany more previously lost shoppers making it through. Compare the statuses and the business outcome, rather than optimize speed in isolation.
The practical decision is whether a specific change improves successful paid orders without unacceptable fraud or operational costs. This public timing snapshot cannot answer that question for a store. It supplies a clear measurement definition and a reason to preserve the full path from challenge presentation through payment completion.
## Does a similar challenge median mean a similar authentication experience?
No. The US and India have completed-challenge medians of 20.90 and 22.51 seconds, only 1.61 seconds apart, while their displayed challenge shares are 7.30% and 94.25% of initiated authentications. India therefore has about 12.91 times the challenge share in this platform snapshot. Germany's median is 21.95 seconds with a 22.43% challenge share. Similar completed-flow timing can coexist with very different exposure to the challenge itself. (Evervault, January to June 2026 platform data (https://evervault.com/3ds-performance), Jun 2026).
The dashboard separately reports overall authentication success of 98.6% in the US,81.5% in India and 94.3% in Germany. Those differences describe different traffic mixes and issuer environments. They do not establish that challenge frequency caused the success-rate differences or that changing a timeout would close them.
### Can challenge share times median seconds estimate average delay?
That product would not recover average authentication time. The timing value is a median among completed challenged flows, while challenge share concerns initiated requests. A median is not a mean, incomplete flows are outside the timing population and other authentication stages still take time. The underlying joint distribution is not supplied. A speed or cost model needs actual per-attempt durations and statuses, not a weighted average made from unmatched aggregates.
There is also an explicit source reconciliation limit. Evervault's methodology says frictionless and challenge shares sum to attempts, yet its global displayed 68.36% and 22.67% sum to 91.03%, leaving 8.97 percentage points unexplained by those two displayed categories. The US displayed shares sum to 88.03%. We do not label the residual as failed, abandoned or unclassified traffic because the public page does not resolve the discrepancy. The individually reported timing and challenge observations remain scoped as displayed. (Evervault, methodology and snapshot (https://evervault.com/3ds-performance), Jun 2026).
### Which pair of measures should a local team inspect?
Inspect challenge incidence among all initiated attempts and the duration distribution among the relevant completed flows, then keep abandonment and final payment status alongside them. A country or issuer with quick completed challenges can still expose almost every shopper to that extra step. A lower challenge incidence can coexist with longer challenges among the smaller group required to complete one.
For a proposed timeout change, verify the provider's integration rules and test delayed returns from a bank app without creating duplicate attempts. Preserve the original cart and distinguish authentication success from the eventual payment outcome. The source does not publish a tail percentile or a safe timeout, so a 25.35-second median cannot supply either.
The challenge exposure and timing table (https://checkoutcloser.com/guides/3d-secure-challenge-completion-time-statistics/data/3ds-exposure-and-time.csv) and calculation (https://checkoutcloser.com/guides/3d-secure-challenge-completion-time-statistics/analysis.py) preserve all eleven countries' displayed metrics, plus the documented classification mismatch. They support a more complete diagnostic question, not a causal ranking of markets or providers.
## Methodology and Sources
We read Evervault's complete public performance dashboard and methodology on 13 November 2026, preserving the exact source text. Its data period is 1 January to 30 June 2026; the citation date identifies that period end, not an invented publication date. We transcribed the eleven displayed country medians and retained the separately reported global median. The script calculates selected time gaps and ratios from rounded inputs. No challenge-time distribution, abandoned-flow duration, confidence interval or causal country effect is reconstructed. Stripe's current authentication-flow documentation supplies the stage distinction, not a performance estimate. Provider traffic and exclusions limit transfer to an individual merchant.
## Glossary
3D Secure: A protocol used to authenticate a cardholder during certain card payments.
Challenge: An additional authentication interaction requested by an issuer.
Median: The middle of an ordered distribution, also called the fiftieth percentile.
Completed flow: A flow reaching the final result used in the source's timing definition.
Authorization: The subsequent decision about whether the card payment can proceed.
## Sources
Evervault: 3DS Performance Dashboard (https://evervault.com/3ds-performance)
Stripe: Authenticate with 3D Secure (https://docs.stripe.com/payments/3d-secure/authentication-flow)
Quick answer
3D Secure challenge completion time statistics depend on which flows finish. Evervault's January to June 2026 platform snapshot reports a global median of 25.35 seconds from challenge presentation to the final result, measured on completed flows. Selected country medians range from 20.90 seconds in the United States to 35.32 seconds in Spain (Evervault, January to June 2026 platform data, Jun 2026).
When authentication feels slow, the useful question is where the delay occurs and which flows the timer includes. This analysis compares the published times, distinguishes completed challenges from all initiated attempts and follows authentication through to the payment outcome. These platform observations are context for an investigation, rather than a forecast of an individual store's performance.
The global completed-challenge median was 25.35 seconds. The platform snapshot covers customer initiated 3DS sessions from January through June 2026 (Evervault, January to June 2026 platform data, Jun 2026).
Spain's selected-market median was 14.42 seconds above the US value. Published medians were 35.32 and 20.90 seconds respectively, a descriptive comparison rather than a causal country effect (Evervault, January to June 2026 platform data, Jun 2026).
A completed challenge is not a completed payment. Stripe describes successful authentication followed by an attempt to charge the customer, with payment outcome recorded separately (Stripe documentation, Nov 2026).
The clock starts when the challenge is presented, rather than when the shopper first reaches checkout. It ends at the final result among completed challenge flows. It therefore describes a specific part of the journey and a selected population.
Figure 1: Similar completed-flow speed can hide very different exposure. Source: Evervault, country comparison, Jan to Jun 2026. Original explanation and arithmetic from the cited primary research.
Evervault defines the time metric as the median elapsed seconds from challenge presentation to the final result, measured on completed flows. The source does not describe it as the time required for all shoppers to complete authentication (Evervault, January to June 2026 platform data, Jun 2026).
A median splits the observed completed flows around the middle of their time distribution. It does not identify the slow tail, the share abandoning, or how long a shopper who never returns remained stuck. A store can have a short completed-flow median while losing some slower shoppers before they produce a completed result.
This matters when evaluating an improvement. If a change causes slow challenges to disappear from the completed population, the reported median may fall without the experience improving. A valid diagnosis needs the attempt count and each terminal status alongside time distributions.
The metric excludes time spent choosing a payment method, entering card details or correcting a cart. It also does not cover the entire authorization and settlement sequence. Stripe's flow documentation explicitly separates the authentication result from the subsequent charge and final payment status (Stripe documentation, Nov 2026).
An authentication screen can close successfully while the charge later fails. Conversely, an unfinished challenge may reflect a technical interruption, a shopper switching apps or a deliberate decision to stop. The elapsed-time statistic alone cannot assign those causes.
The selected market medians show a meaningful time difference while remaining descriptive. They are platform observations with different underlying traffic, rather than randomized treatments assigned by country.
The United States median was 20.90 seconds, Germany 21.95 seconds, the United Kingdom 29.16 seconds and Spain 35.32 seconds in the same January to June 2026 dashboard snapshot. Spain's value is 14.42 seconds above the US value and approximately 1.69 times as large. The UK value is 8.26 seconds above the US value (Evervault, January to June 2026 platform data, Jun 2026).
These are differences and ratios of the source's rounded medians, calculated here. They are not differences in average checkout duration, customer patience or a causal penalty for serving a market. A mixture of issuers, authentication methods, devices and customer behavior can affect the observed times.
The CSV retains all eleven country rows displayed by the source, rather than only the four highlighted in the figure. Other published medians include 26.32 seconds for Canada, 22.58 seconds for France and 34.50 seconds for Brazil. The global median of 25.35 seconds comes from the platform's broader population, not the simple average of the displayed countries (Evervault, January to June 2026 platform data, Jun 2026).
An average of country medians would give each displayed country the same weight regardless of volume, and would still not recover the median or average of individual challenge times. A median is not additive. Without the underlying flows and a valid aggregation rule, a new Europe-versus-America timing statistic would be invented precision.
The dashboard also excludes some observations when data are incomplete, volumes are low or outliers heavily distort results. Its published scope should travel with the numbers. The selected markets are not a population census of every DTC checkout or every authentication provider (Evervault, January to June 2026 platform data, Jun 2026).
Use the observed time as a diagnostic reference alongside the completion funnel. A longer challenge path can merit investigation, but a global benchmark does not decide the store's timeout or prove that a particular issuer is at fault.
Record challenge presentation, final authentication result and final payment result against the same permitted transaction identifier. Keep a visible distinction between successful, failed, abandoned and technically interrupted flows. Inspect timing among completed flows and the tail of the time distribution, then examine the relevant noncompletion statuses.
For a phone journey, test the route out to a bank application and back to the storefront. Confirm that the cart and payment state remain recoverable and that a delayed return does not create a duplicate payment attempt. Stripe describes failed authentication returning to a state where another method or a reconfirmed authentication attempt may be needed (Stripe documentation, Nov 2026).
The merchant should choose a recovery policy around its integration and provider requirements. A median of 25 seconds is not evidence that a 25-second timeout is safe. A cutoff at the middle of completed observations could interrupt legitimate flows that take longer.
Suppose completed-challenge times fall while abandonment and failed payments increase. The shorter median would be a poor standalone success claim. Equally, a longer completed-flow median may accompany more previously lost shoppers making it through. Compare the statuses and the business outcome, rather than optimize speed in isolation.
The practical decision is whether a specific change improves successful paid orders without unacceptable fraud or operational costs. This public timing snapshot cannot answer that question for a store. It supplies a clear measurement definition and a reason to preserve the full path from challenge presentation through payment completion.
No. The US and India have completed-challenge medians of 20.90 and 22.51 seconds, only 1.61 seconds apart, while their displayed challenge shares are 7.30% and 94.25% of initiated authentications. India therefore has about 12.91 times the challenge share in this platform snapshot. Germany's median is 21.95 seconds with a 22.43% challenge share. Similar completed-flow timing can coexist with very different exposure to the challenge itself. (Evervault, January to June 2026 platform data, Jun 2026).
The dashboard separately reports overall authentication success of 98.6% in the US,81.5% in India and 94.3% in Germany. Those differences describe different traffic mixes and issuer environments. They do not establish that challenge frequency caused the success-rate differences or that changing a timeout would close them.
That product would not recover average authentication time. The timing value is a median among completed challenged flows, while challenge share concerns initiated requests. A median is not a mean, incomplete flows are outside the timing population and other authentication stages still take time. The underlying joint distribution is not supplied. A speed or cost model needs actual per-attempt durations and statuses, not a weighted average made from unmatched aggregates.
There is also an explicit source reconciliation limit. Evervault's methodology says frictionless and challenge shares sum to attempts, yet its global displayed 68.36% and 22.67% sum to 91.03%, leaving 8.97 percentage points unexplained by those two displayed categories. The US displayed shares sum to 88.03%. We do not label the residual as failed, abandoned or unclassified traffic because the public page does not resolve the discrepancy. The individually reported timing and challenge observations remain scoped as displayed. (Evervault, methodology and snapshot, Jun 2026).
Inspect challenge incidence among all initiated attempts and the duration distribution among the relevant completed flows, then keep abandonment and final payment status alongside them. A country or issuer with quick completed challenges can still expose almost every shopper to that extra step. A lower challenge incidence can coexist with longer challenges among the smaller group required to complete one.
For a proposed timeout change, verify the provider's integration rules and test delayed returns from a bank app without creating duplicate attempts. Preserve the original cart and distinguish authentication success from the eventual payment outcome. The source does not publish a tail percentile or a safe timeout, so a 25.35-second median cannot supply either.
The challenge exposure and timing table and calculation preserve all eleven countries' displayed metrics, plus the documented classification mismatch. They support a more complete diagnostic question, not a causal ranking of markets or providers.
We read Evervault's complete public performance dashboard and methodology on 13 November 2026, preserving the exact source text. Its data period is 1 January to 30 June 2026; the citation date identifies that period end, not an invented publication date. We transcribed the eleven displayed country medians and retained the separately reported global median. The script calculates selected time gaps and ratios from rounded inputs. No challenge-time distribution, abandoned-flow duration, confidence interval or causal country effect is reconstructed. Stripe's current authentication-flow documentation supplies the stage distinction, not a performance estimate. Provider traffic and exclusions limit transfer to an individual merchant.