Passwordless Authentication Methods for Customer Apps: Passkeys, Certificates and PKI Compared
Passwords remain the weakest part of most customer sign-in journeys. They are reused across sites, phished at scale and reset so often that support desks budget for it. Choosing among the available passwordless authentication methods is therefore one of the first decisions Schutz IT works through with clients planning a customer identity programme, because the method you pick shapes user experience, fraud exposure and back-end infrastructure for years. This guide compares the main options, explains where certificates and PKI fit in, and sets out a practical way to roll passwordless sign-in out without locking customers out.
Why Customer-Facing Apps Are Moving Away From Passwords
The case against passwords is no longer theoretical. Credential stuffing works because people reuse the same password across shopping, banking and streaming accounts, so a breach at one company becomes an attack on every other. Phishing kits now proxy real login pages in real time and capture one-time codes along with the password. On the business side, forgotten-password flows are among the highest-volume support requests, and every abandoned reset is a lost sale or a lost sign-up.
At the same time, the building blocks for passwordless sign-in have matured. Every major mobile and desktop platform now supports FIDO2 and WebAuthn, browsers sync passkeys across devices, and customer identity platforms such as Microsoft Entra External ID and Azure AD B2C can orchestrate several sign-in methods inside one branded journey. The question for most teams is no longer whether to go passwordless, but which methods to offer, to whom, and in what order.
Comparing the Main Passwordless Options
No single method suits every audience. The useful comparison is how each one performs on phishing resistance, ease of use and recovery.
Passkeys (FIDO2 / WebAuthn). A passkey is a key pair created on the user's device; the private key never leaves the authenticator and the public key is registered with your service. Sign-in is a fingerprint, face scan or device PIN. Because the credential is bound to your domain, a look-alike phishing site cannot use it. Synced passkeys roam with the user's Apple, Google or Microsoft account, which solves the old "lost device" problem for most consumers. For new customer apps this is now the default recommendation.
Email magic links and one-time codes. These remove the password but not the shared secret: whoever controls the inbox or intercepts the code can sign in. They are easy to adopt and useful as a fallback, yet they are not phishing-resistant and they push security onto the customer's email provider. Treat them as a convenience method, not a high-assurance one.
SMS one-time passcodes. SMS is familiar and works on any phone, but SIM-swap fraud and real-time phishing have weakened it considerably. Many regulators and security frameworks now discourage SMS as a primary factor. Keep it, if at all, as a recovery path for low-risk accounts.
Authenticator app push and number matching. Push approvals with number matching are a reasonable middle ground for workforce and partner portals where users already carry a managed authenticator. For mass-market consumers, asking people to install a separate app adds friction.
Certificate-based authentication. Here the user or device holds an X.509 certificate issued by a trusted certificate authority, and authentication happens through mutual TLS or a smart card. It is highly phishing-resistant and well suited to B2B portals, partner extranets, connected devices and regulated sectors. The trade-off is operational: someone has to issue, renew and revoke those certificates reliably.
Where PKI Fits Into Passwordless Sign-In
It is easy to think of passwordless as a front-end feature, but several of these methods rely on public key infrastructure behind the scenes. Passkeys use public key cryptography, and enterprise deployments often verify authenticator attestation to confirm which hardware a credential lives on. Certificate-based authentication depends entirely on a certificate authority that partners and devices trust.
For organisations issuing certificates to customers' devices, partner organisations or IoT fleets, a dedicated private PKI hierarchy is usually the right foundation. Public CAs are designed for website TLS and are not meant to issue client certificates to your partners' laptops or your smart meters. A private hierarchy lets you define your own certificate profiles, lifetimes and revocation rules, and keep the root key offline in a hardware security module.
The operational side matters as much as the design. Shorter certificate lifetimes are now the norm across the industry, which makes automated issuance and renewal essential. We covered how this shift affects security teams in an earlier post on shorter certificate lifetimes and enterprise PKI. If you are still deciding how to host the certificate authority itself, the trade-offs are set out in our comparison of cloud PKI and on-premises CAs.
Designing a Passwordless Customer Journey That Actually Converts
A passwordless rollout succeeds or fails on the journey, not the cryptography. A few design principles consistently make the difference.
- Offer passkeys at the moment of highest intent. Prompt customers to create a passkey straight after a successful sign-up or checkout, when they are already engaged, rather than on a cold settings page.
- Keep a fallback, but rank it below passkeys. Email one-time codes are a sensible fallback for devices that cannot create a passkey. Avoid making the weaker method the default button.
- Treat account recovery as a security control. Attackers target the weakest recovery path. If a passkey can be bypassed with a single SMS code, the account is only as strong as SMS. Use step-up checks and risk signals for recovery.
- Apply risk-based access. Low-risk actions such as browsing order history can rely on a remembered device; high-risk actions such as changing payout details should require a fresh, phishing-resistant check.
- Measure the funnel. Track passkey enrolment rate, sign-in success rate and recovery volume per method. These numbers tell you when it is safe to retire weaker options.
On Microsoft platforms, these journeys are built with user flows or custom policies in Azure AD B2C, or with the newer Entra External ID experience. Each supports passkeys, email one-time passcodes and federation, but the configuration choices, especially around recovery and migration of existing password users, need careful planning.
A Phased Rollout Plan for Security and Product Teams
Phase 1: Assess. Map your user segments, devices and regulatory requirements. Identify which journeys carry the most fraud or support cost. Decide which segments need certificate-based or hardware-bound credentials and which can use synced passkeys.
Phase 2: Add passwordless alongside passwords. Let existing customers enrol a passkey without removing their password. Measure adoption and fix friction points before going further. Make sure the identity platform logs which method each sign-in used.
Phase 3: Make passwordless the default. New sign-ups start with a passkey; password creation becomes an opt-in fallback. Recovery moves to verified email plus risk checks rather than knowledge-based questions.
Phase 4: Retire passwords for enrolled users. Once a customer has two working passwordless methods, offer to remove their password entirely. For B2B and device scenarios, move partners onto certificate-based access backed by automated certificate lifecycle tooling. If you are weighing whether to run that certificate service yourselves or buy it as a service, our article on managed PKI versus an in-house CA walks through the decision.
Common Mistakes to Avoid
- Launching passkeys without a tested recovery path, which leads to support spikes and frustrated customers.
- Calling email magic links "phishing-resistant". They remove the password but not the interceptable secret.
- Issuing client certificates from an ad-hoc CA with no revocation process or monitoring.
- Migrating every user at once instead of measuring adoption by segment.
- Ignoring accessibility: biometric prompts must always have a device-PIN alternative.
Final Thoughts
Going passwordless is a strategic identity decision rather than a single feature toggle. Passkeys should lead for most consumer journeys, certificate-based authentication backed by well-run PKI suits partners and devices, and weaker methods should be kept only as fallbacks with strong recovery controls. Planning the journey, the recovery model and the certificate infrastructure together is what turns a pilot into a durable programme.
If your team is planning a CIAM or PKI project and would like an experienced partner to review the approach, contact Schutz IT to discuss your requirements.
Comments
Post a Comment