Skip to content
EncryptTechBD
Home Industries Our Work Blog Pricing About Contact Discuss Your Project
বাংলা (BN)
Privacy & Compliance 6 min read

Business email security policy: a risk register approach

Rank email risks by likelihood and impact, choose controls that address the real failure modes, and write policy clauses people can follow under pressure.

E
EncryptTechBD · 0 comments
Passport, phone and earphones on a desk
Photo: Rawpixel

Most email policies are written as a list of prohibitions: do not open suspicious attachments, do not share passwords, do not use personal email. They describe the desired behaviour but not the risks, the controls or the person accountable when a control fails. A risk register does the opposite. It forces a business to state what could happen, how likely it is, what is being done and what remains.

This guide is for an SME that runs its own mail domain, uses Microsoft 365 or Google Workspace, and handles payments or customer data by email. It treats email as a business process with several failure points, not as an IT setting.

Scope the policy around the mail flow

Before writing clauses, inventory what email actually does in the business. Include the main domain and any local or campaign domains, shared inboxes such as accounts@ and support@, aliases that forward outside the company, distribution groups, marketing senders, mobile access and any application that sends mail on the company's behalf.

Then mark the mailbox owners. A shared inbox with no owner is where payment requests go unanswered and where leavers retain access. A domain nobody remembers registering is where a spoofing incident starts.

The risk register

The ratings below are illustrative for a typical SME and should be re-scored against your controls. Use a five-level or four-level scale and keep the definitions consistent.

RiskLikelihoodImpactPrimary controls
Credential phishing leads to mailbox takeoverHighHighMFA, phishing-resistant authentication for admins, link and attachment filtering, reporting route
Supplier payment request is altered (BEC)MediumHighCallback verification to a known number, dual approval for bank changes, external sender banner
Domain is spoofed or a lookalike domain is usedMediumHighSPF, DKIM and DMARC enforcement, registrar lock, lookalike monitoring
Malicious attachment or link is openedHighMediumSandboxing, macro blocking, browser isolation where available, endpoint protection
Legacy mail protocol bypasses access controlsMediumHighDisable legacy authentication, conditional access, device compliance
Hidden forwarding rule exports mailMediumHighRestrict external forwarding, alert on new rules, periodic mailbox audit
Sensitive file sent to the wrong recipientMediumMediumData classification, recipient warnings, DLP rules, documented recall procedure
Mailbox data is deleted or encryptedLowHighRetention policy, immutable backup, tested restore, separate admin accounts

Likelihood and impact without fake precision

Do not assign decimal probabilities you cannot defend. Define four levels and describe the evidence that would move a risk up or down.

  • Likelihood: rare, possible, likely, almost certain. Base the rating on whether the control exists, whether it is tested, and whether similar businesses in your sector have reported the event.
  • Impact: minor, moderate, major, severe. Use money, downtime, customer harm and regulatory exposure, and state which one dominates.

A high-likelihood, high-impact risk needs a control with an owner and a test date. A low-likelihood, high-impact risk may be accepted with a recovery plan. The point is to make the acceptance explicit, not to produce a perfect score.

Controls by layer

Identity. Enforce multi-factor authentication on every mailbox, especially administrators. Prefer phishing-resistant methods for privileged accounts. Block legacy protocols, apply conditional access by device and location, and remove dormant accounts on a schedule.

Domain. Publish SPF, DKIM and DMARC. Move DMARC from monitoring to enforcement when the reports show legitimate senders are authenticated. Lock the registrar account, enable transfer protection and monitor for lookalike domains that target your brand.

Content. Filter links and attachments, block macros from the internet, and add an external sender warning. These controls reduce volume but do not replace user judgement; assume some malicious mail will arrive.

Mailbox. Restrict automatic forwarding to external addresses, alert on new inbox rules, review delegate access and separate administrative roles from daily-use accounts. Retain audit logs long enough to investigate an incident.

People. Give staff a one-click report button and tell them what happens next. Train on the specific patterns used against finance and management, not on generic "cyber awareness". Praise reporting; punish concealment.

Recovery. Back up mailboxes or rely on a retention and export capability you have tested. Know how to restore a single mailbox, revoke sessions and preserve evidence without destroying it.

Policy clauses that change behaviour

  • Payment changes require a callback. Any change to bank details is confirmed by phone using a number already on file, never a number in the email.
  • Report first, investigate second. Staff who report a suspected phishing message within the first hour are protected from blame for clicking.
  • Business email only. Customer data and attachments go through company accounts and approved file-sharing, not personal mailboxes.
  • Shared inboxes have a named owner. The owner reviews access quarterly and on every departure.
  • Forwarding is exceptional. External auto-forwarding requires approval and expires.
  • Administrators use separate accounts. Day-to-day email never carries domain administrator rights.

Residual risk and sign-off

For each risk, record the controls in place, the evidence that they work, the residual rating and the person who accepts it. The finance director should accept payment-related residual risk; the IT manager should not accept it on the finance director's behalf. Review the register quarterly and after any incident or major change, such as a migration or a new payment process.

Questions for the board

  • Which mailbox compromise would stop the business fastest, and who would notice first?
  • Is DMARC enforcing, and who reviews the reports?
  • Can a finance payment be changed without a verbal confirmation to a known contact?
  • How quickly can we revoke a session, restore a mailbox and preserve evidence?
  • Which third parties send mail as us, and how is that access controlled?

If an account is compromised

Contain first: reset the password, revoke sessions and tokens, and disable suspicious forwarding rules. Check sent items, inbox rules, delegate access and sign-in logs for the previous weeks. Tell finance and the bank if payment details may have been changed. Preserve logs before cleaning up, notify affected customers according to your obligations, and write a short incident note with the timeline and decisions. Then fix the control that failed, not just the account.

Email controls are delivered through Managed Technology & Cybersecurity. The pricing page explains the difference between a policy review and ongoing monitoring. For the training side, read the phishing prevention cost guide.

Sources

Services that deliver this

Want this implemented properly?

Pricing →

Every engagement starts with a written scope. If a requirement is outside it, we quote it separately before starting.

Responses (0)

No comments yet. Add the first one below.

Leave a response

Your email address will not be published. Required fields are marked *.

Keep reading

More field notes