Multi-factor authentication is one of the few security controls where the evidence for its effectiveness is so clear and so consistent that recommending it requires no caveats. MFA prevents the majority of credential-based account takeover attacks — the category of attack that underlies a significant proportion of all significant breaches. Cyber insurers ask for it. Cyber Essentials requires it for cloud services. The NCSC recommends it without reservation.
And yet a large proportion of small businesses either haven't implemented it at all, or have implemented it on some systems and not others — leaving gaps that a credential-theft attack will find and exploit. This guide explains what MFA is, where it must be implemented, how to do it without causing significant disruption, and what the common mistakes look like.
What Multi-Factor Authentication Actually Is
Authentication is the process of proving you are who you claim to be when accessing a system. The traditional method — a username and password — is single-factor: one thing you know. Multi-factor authentication requires at least two distinct types of evidence:
- Something you know: A password, PIN, or security answer.
- Something you have: A phone that receives a code, a hardware security key, an authenticator app.
- Something you are: A fingerprint, face scan, or other biometric.
The practical significance is this: if an attacker steals your password — via phishing, a credential dump, or by guessing — they still cannot access your account without the second factor. The password is useless without the phone. This is why MFA is effective against credential phishing even when the phishing itself succeeds.
Where MFA Must Be Enabled: The Priority Order
Not all systems carry equal risk, and a practical rollout prioritises the highest-impact systems first. Here is the order we recommend for most small businesses:
1. Email (Highest Priority)
Email is the master key to most other accounts — password reset links are sent by email, and control of an email account gives an attacker the ability to take over almost any service associated with it. Business email compromise, where an attacker uses a compromised email account to redirect payments or exfiltrate data, is one of the costliest attack types affecting small businesses. Enable MFA on Microsoft 365 or Google Workspace first, before anything else.
2. VPN and Remote Access
If your staff access company systems remotely — via VPN, Remote Desktop, or any other remote access mechanism — MFA on that entry point is essential. Open RDP without MFA is one of the most actively exploited attack vectors in existence. If you have remote access enabled without MFA, addressing this should be treated as urgent.
3. Cloud Services and SaaS
Every cloud service used for business purposes — CRM, HR systems, accounting software, file storage, project management — should have MFA enabled. The 2024 Cyber Essentials requirement for MFA on cloud services makes this a certification requirement as well as a security best practice. Work through your service inventory and enable MFA on each in turn.
4. Administrator and Privileged Accounts
Accounts with administrative access to your IT infrastructure — domain administrators, cloud console access, firewall management — should be treated as the highest-risk accounts in your environment. MFA here is non-negotiable: a compromised admin account gives an attacker the ability to disable your other security controls.
5. All Other Staff Accounts
Once the highest-priority systems are covered, extend MFA to all staff accounts on all remaining business systems. This is where the rollout takes the most organisational effort, but it's the step that closes the remaining gaps.
Choosing the Right MFA Method
Not all MFA methods are equally secure. From strongest to weakest:
- Hardware security keys (FIDO2/WebAuthn): Physical USB or NFC devices such as YubiKey. Resistant to phishing because the key cryptographically binds to the specific website. The strongest available option for high-risk accounts.
- Authenticator apps (TOTP): Microsoft Authenticator, Google Authenticator, Authy. Generate time-based one-time codes. Significantly better than SMS. Suitable for most business use cases.
- Push notifications: A notification sent to a registered device requiring approval. Convenient and reasonably secure, but vulnerable to "MFA fatigue" attacks — attackers repeatedly sending approval requests until a tired user clicks approve. Can be mitigated by enabling number matching in Microsoft Authenticator.
- SMS codes: A code sent by text message. Significantly better than no MFA, but vulnerable to SIM-swapping and interception. Use only if app-based MFA is genuinely not possible for a given user.
How to Roll It Out Without Breaking Everything
The most common reason MFA rollouts fail in small organisations isn't technical — it's organisational. Staff who don't understand why it's being introduced resist it. Staff who aren't given adequate notice lose access to systems at inconvenient moments. Rollouts that happen without IT support available create immediate problems that undermine confidence in the initiative.
A straightforward approach: announce the rollout at least two weeks in advance, explaining clearly why it's happening and what staff will need to do. Schedule a specific date for enabling MFA on each system, outside of peak working hours. Have IT or a knowledgeable colleague available to assist for the first few days. Create a simple written guide for staff who need to set up their authenticator app.
Plan for the "I've lost my phone" scenario before it happens: every system should have a documented account recovery process, and admin accounts should have a recovery method that doesn't depend on the phone being available.
What MFA Doesn't Protect Against
MFA is not a complete defence. Real-time phishing proxies — tools that sit between the victim and the legitimate website, harvesting credentials and MFA tokens simultaneously and replaying them immediately — can defeat TOTP-based MFA. Adversary-in-the-middle attacks of this type are increasingly common and are the reason that hardware security keys, which are immune to this technique, are recommended for the highest-risk accounts.
MFA also doesn't protect a compromised session after authentication — if a device is infected with malware that steals authenticated session cookies, the attacker can access your accounts without ever needing your credentials or MFA code. This is why MFA must be combined with endpoint detection, not treated as a standalone control.
Further Reading
MFA Is the Lock. Monitoring Is the Camera.
MFA stops most credential attacks. The attacks that get through — real-time phishing proxies, session hijacking, compromised endpoints — are detected by continuous monitoring. SOC in a Box provides both the endpoint detection and the network-level visibility to catch what MFA misses. Your named analyst sees authentication anomalies in real time.
Book your scoping call