What is Multifactor Authentication (MFA)?

Multifactor authentication (MFA) requires users to present two or more different types of evidence, known as authentication factors, before they’re granted access to an application, service, or network. Passwords alone are no longer adequate. They’re routinely phished, guessed, reused, and stolen in data breaches, and MFA ensures that a stolen password by itself isn’t enough to compromise an account. However, not all MFA is created equal, and some common methods are now easily defeated by attackers.

Last updated: November 9, 2026

Authentication Factors

Authentication factors fall into three categories. True MFA combines at least two different categories.

Something You Know

Knowledge factors are secrets the user memorizes, such as a password or PIN. They’re the most common factor and the easiest to steal, because users can be tricked into revealing them.

Something You Have

Possession factors prove the user has a specific physical or digital object, such as a mobile device with an authenticator app, a hardware token like an RSA SecurID or FIDO2 security key, a smart card, or a private key stored in a device’s Trusted Platform Module (TPM).

Something You Are

Inherence factors are biometric characteristics, such as a fingerprint or facial features. On modern devices, biometrics are typically used locally to unlock a credential stored on the device, rather than being sent to a server.

Two factors from the same category, such as a password and a security question, don’t constitute MFA.

Risk Signals Are Not Factors

Location, device health, network, and user behavior are sometimes described as additional factors. They’re more accurately described as risk signals. Policy engines such as Microsoft Entra Conditional Access use these signals to decide when to require MFA, require a compliant device, or block access entirely. Risk signals are valuable, but they complement authentication factors rather than replace them, and NIST SP 800-63B-4 notes that they don’t change the authentication assurance level.

Types of MFA

MFA methods vary widely in the protection they provide. From weakest to strongest, the common types are as follows.

  • SMS and voice codes. Vulnerable to SIM swapping, interception, and phishing. Better than a password alone, but the weakest form of MFA.
  • Email codes and links. Only as secure as the email account itself, which is often protected by the same password.
  • One-time passcodes. Time-based codes from an authenticator app or hardware token. Stronger than SMS, but users can still be tricked into entering them on a phishing site.
  • Push notifications. Convenient, but vulnerable to MFA fatigue attacks. Number matching, which Microsoft Authenticator now requires, significantly reduces this risk, but push notifications can still be phished.
  • Phishing-resistant MFA. FIDO2 security keys and passkeys, Windows Hello for Business, and certificate-based authentication with smart cards or TPM-protected keys. These methods can’t be phished in the traditional sense.

Phishing-Resistant MFA

Phishing-resistant MFA uses public key cryptography. The private key never leaves the user’s device or security key, and the authentication is cryptographically bound to the legitimate service, so a fake website or an attacker-in-the-middle proxy can’t capture and replay it. Phishing-resistant methods include the following.

  • FIDO2 security keys and passkeys. Device-bound passkeys on a hardware security key or platform authenticator, and syncable passkeys that synchronize across a user’s devices.
  • Windows Hello for Business. A TPM-protected credential unlocked with a PIN or biometric.
  • Certificate-based authentication. Smart cards, virtual smart cards, and certificates with TPM-protected private keys.

Standards and platforms increasingly require phishing resistance. NIST SP 800-63B-4, finalized in 2025, requires a phishing-resistant authenticator with a non-exportable private key at the highest assurance level (AAL3) and allows syncable passkeys at AAL2. In Microsoft Entra ID, administrators can require the built-in phishing-resistant MFA authentication strength in Conditional Access policies.

Drawbacks and Common Attacks

MFA is one of the most effective security controls available, but it isn’t perfect. Organizations should understand the following drawbacks and attacks.

MFA Fatigue

In an MFA fatigue (or push bombing) attack, an attacker who has a user’s password repeatedly triggers push notifications until the user approves one out of frustration or confusion. The LAPSUS$ group used this technique in its widely publicized 2022 breach of Uber. Number matching, which requires the user to enter a number shown on the sign-in screen, is the most effective mitigation for push-based MFA. Users should also be trained never to approve a request they didn’t initiate.

Adversary-in-the-Middle Phishing

Modern phishing kits act as a proxy between the user and the real sign-in page. The user enters their password and one-time code or approves a push notification, and the attacker captures the resulting session token. This attack defeats SMS, one-time passcodes, and push notifications alike. Only phishing-resistant MFA reliably stops it.

SIM Swapping

Attackers can convince a mobile carrier to transfer a victim’s phone number to a new SIM card, giving them access to SMS and voice codes. Organizations should phase out SMS and voice as MFA methods wherever possible.

User Experience

Frequent MFA prompts add friction, reduce productivity, and train users to approve prompts without thinking. Risk-based Conditional Access policies, which only require additional verification when the risk warrants it, and passwordless methods such as Windows Hello for Business and passkeys improve both security and the user experience.

Complexity and Cost

Integrating MFA with legacy applications and protocols, such as RADIUS and LDAP, can be complex, and some legacy systems can’t support modern MFA at all. Centralizing authentication in an identity provider using federation protocols such as SAML and OpenID Connect reduces this complexity. Hardware tokens and security keys also carry purchase and distribution costs.

Account Recovery

Account recovery is often the weakest link. Attackers target help desks with social engineering to reset MFA methods. Organizations should verify identity carefully before resetting credentials and use secure onboarding and recovery options, such as a time-limited Temporary Access Pass in Microsoft Entra ID.

A False Sense of Security

MFA doesn’t eliminate all risk. Weaker methods can be bypassed, session tokens can be stolen from compromised devices, and users can still be socially engineered. MFA should be one layer in a broader Zero Trust strategy that includes device compliance, least-privilege access, and monitoring.

MFA and Zero Trust

Strong authentication is a foundational principle of Zero Trust and Zero Trust Network Access (ZTNA). Every access request should be authenticated with strong, preferably phishing-resistant MFA, and evaluated against risk signals. For remote access specifically, see Always On VPN and Multifactor Authentication.

How I Can Help

I help organizations move to strong, phishing-resistant authentication, including the following.

Additional Information

Digital Identity Guidelines: Authentication and Authenticator Management (NIST SP 800-63B-4)

Conditional Access Authentication Strengths (Microsoft Learn)

Number Matching in Multifactor Authentication Push Notifications (Microsoft Learn)

Digital Certificates for Strong, Phishing-Resistant Authentication

Always On VPN and Multifactor Authentication

Start a Conversation

Planning a move to phishing-resistant MFA or certificate-based authentication? Tell me about your environment, and I’ll follow up, typically within one business day.

← Back

Thank you for your response. ✨