Small business security monitoring for computers and email
Small business security monitoring means software on each computer and in your email system records what accounts and programs do, and a person reads the alerts it raises and acts on them, nights and weekends included. For a 3 to 30 person office it covers three things: sign-ins, programs starting that shouldn't, and signs that an email account was taken over, such as a new forwarding rule. Antivirus alone doesn't do this, and neither does a tool that sends its alerts to nobody. The software is cheap or already part of Microsoft 365 Business Premium; the hard part is deciding who answers the alert.
This post is about computers and accounts, not burglar alarms or cameras. Monitoring is item 10 on the small business cybersecurity checklist, after the safeguards that stop most attacks from starting. It is there for the attack that gets past them.
What does small business security monitoring watch?
Three kinds of events matter most in a small office.
Sign-ins. Failed and successful sign-ins to Microsoft 365 or Google Workspace and to each computer. A run of failures means someone is guessing passwords. A successful sign-in from a country where nobody works, the day after a phishing email, means someone has the password. Goal 3.E of CISA's Cross-Sector Cybersecurity Performance Goals (the CPGs) asks organizations to log every unsuccessful login and notify security staff "after a specific number of consecutive unsuccessful login attempts in a short period."
New programs. A program that has never run in your office, started from an email attachment or the downloads folder, or a remote-control tool nobody asked for. An ordinary admin tool can do an intruder's work, and antivirus has no reason to block it. What tells you something is wrong is who ran the tool, when, and what it did next.
Email account takeover. Microsoft 365 has built-in alert policies for signs of a taken-over mailbox, among them "Creation of forwarding/redirect rule", "Suspicious email forwarding activity" and "User restricted from sending email." Microsoft's descriptions of the last two say they can mean the account is compromised, and Microsoft turns its built-in policies on by default (alert policies in Microsoft 365). The forwarding rule alert has a gap: it only tracks rules created in Outlook on the web or through Exchange Online PowerShell. If you suspect one account right now, the post on what to do after clicking a phishing link covers the checks to do by hand.
Is antivirus the same as security monitoring?
No. Antivirus compares files and programs against known malware and blocks what it recognizes. Endpoint detection and response (EDR) is software that records what programs and accounts do on each computer, so it can flag behavior that matches no known virus, and so someone can look back afterward and see what happened.
CISA's goal 4.A asks for both: signature-based detection, the kind antivirus uses, and "non-signature-based mechanisms (focusing behavior, heuristics, or anomalies)" to catch malicious code on each computer.
Microsoft Defender Antivirus is built into Windows 10 and Windows 11. Keep it on, or keep your paid antivirus current. EDR is the part you add, and every alert it raises needs a person.
What Microsoft Defender includes in Microsoft 365
What you get depends on the plan. From Microsoft's documentation, read October 18, 2026:
| Business Basic or Standard | Business Premium | |
|---|---|---|
| Antivirus on Windows | Microsoft Defender Antivirus, part of Windows | Same |
| EDR | Not included. Defender for Business can be added, listed at $3.00 a user a month, paid yearly | Defender for Business included |
| Email threat protection | The built-in protection every cloud mailbox gets | Adds Defender for Office 365 Plan 1 (Safe Links, Safe Attachments) |
| Sign-in log kept | 7 days | 30 days |
| Audit log of user and admin actions kept | 180 days | 180 days |
Defender for Business is Microsoft's EDR for organizations with up to 300 users. Microsoft lists endpoint detection and response, automated investigation and remediation, automatic attack disruption, attack surface reduction and vulnerability management among its features. It works on Windows, Mac, iOS and Android.
The catch is the alert email. Defender for Business shows alerts and incidents in the Microsoft Defender portal, and it emails only the people an admin names in a notification rule, under Settings, Endpoints, Email notifications (Microsoft Learn). If nobody set up that rule, an alert can sit in a portal nobody opens.
How long does Microsoft 365 keep the records?
Looking back at an incident depends on the records still existing. We read Microsoft's retention pages on October 18, 2026 and matched them to the business plans.
- The sign-in log, the record of each sign-in to Microsoft 365, is kept 7 days on Microsoft Entra ID Free and 30 days on Entra ID P1 (Entra data retention). Entra ID Free comes with every Microsoft 365 subscription, and P1 comes with Business Premium (Entra licensing). Upgrading later doesn't bring back sign-ins that already expired.
- The audit log of user and admin actions in Microsoft 365, Microsoft Purview Audit (Standard), comes with Business Basic, Standard and Premium and keeps records 180 days, but on those plans it isn't turned on by default: an admin starts it once from the Audit page in the Microsoft Purview portal (auditing solutions, Purview service description).
So an office on Business Standard that notices a stolen password nine days later has lost the sign-in history from the first two days. CISA's goal 3.Q leaves the length of time to "risk or pertinent regulatory guidelines." Seven days is short for an office that finds out from a client asking about a strange invoice.
These Microsoft figures apply to most of the offices in our own research. In our public DNS check of 316 small tax and accounting firms in Alabama, Tennessee and Georgia on October 14, 2026, 156 of the 262 whose domain receives email use Microsoft 365 (about 6 in 10).
Who reads the alerts and how fast
Answer this before you buy EDR. The usual choices in a small office:
- The owner or office manager. Free, but they see alerts only when they check, during the day, and nobody trained them to tell a false alarm from an intruder.
- An outside IT person or managed service provider. Ask what hours they watch, whether alerts reach a person or a ticket queue, and what they charge for an incident.
- A service with analysts on shift around the clock. Costs more than the software alone. Check that it covers the computers and the email accounts.
CISA's goal 4.B asks organizations to define what counts as an adverse event, escalate suspected ones under the incident response plan, automate the analysis where possible and train the people who do it. No regulation sets a response time in minutes for a small office. What decides the outcome is whether a person sees the alert raised at 2 AM on a Saturday before staff log in on Monday.
Ask whoever you choose four questions. Who receives alerts, by name? Do they cover nights, weekends and holidays? Can they cut a computer off the network, and does that wait for your approval? What do you get in writing afterward?
Crashcart answers them this way. Every Windows computer and Mac runs endpoint detection and response, and a security team watches every computer, 24/7. An analyst investigates suspicious activity as it happens, including 3 AM on a Sunday. Most alerts are false alarms; for the rest you get the findings and the recommended response. When analysts confirm an attacker, that computer can be isolated from your network after your approval, or right away if you've allowed it in advance. Windows computers also send sign-ins, program launches and account changes to a security monitoring system where they're kept and reviewed. If you allow it, Crashcart reads the Microsoft 365 activity log every 5 minutes for signs of a taken-over account, such as a new forwarding rule. You decide whether isolation waits for you.
What CISA's performance goals say about logging and detection
The CPGs, version 2.0, are 34 baseline safeguards for organizations of any size, grouped into six functions. Two goals sit under Detect, and two more under Protect cover logging. CISA rates each one:
| Goal | What it asks | Cost | Impact | Ease |
|---|---|---|---|---|
| 3.E Monitor unsuccessful (automated) login attempts | Log failed logins; alert security staff after repeated failures in a short time | Moderate | High | Moderate |
| 3.Q Maintain log collection and storage | Collect admin and security logs in one central place only authorized people can change; alert when a critical log stops | Moderate | High | Moderate |
| 4.A Establish malicious code detection | Signature and behavior detection on every computer; antivirus updated and scanning email and USB drives | Moderate | High | Moderate |
| 4.B Identify adverse events | Define what counts as an incident, escalate under the response plan, automate analysis, train analysts | Moderate | High | Complex |
All four are rated high impact and none low cost. 4.B is the only one rated complex, and it is the people part: someone has to know what an alert means and what to do with it.
What security monitoring doesn't replace
Monitoring tells you an attack is under way. It doesn't close the hole the attacker used. Most of that work is routine: updates installed on a schedule staff can live with, which the post on how to schedule Windows updates and restarts covers, and MFA on every email account.
Old computers are the other gap. A PC that no longer gets security updates stays exposed however closely someone watches it. The guide to Windows 10 end of support covers how to find those PCs and the choices for each one.
Security monitoring checklist
- Confirm Microsoft Defender Antivirus, or your paid antivirus, is on and current on every Windows computer.
- Decide on EDR. Business Premium includes Defender for Business; on Business Basic or Standard it's an add-on, or part of a monitoring service.
- Name the person who reads alerts, and a backup for nights, weekends and vacations.
- In Defender for Business, create a notification rule for that person under Settings, Endpoints, Email notifications.
- In the Microsoft Defender portal, under Email & collaboration, Policies & rules, Alert policy, check who gets the built-in email alerts, including the forwarding rule alerts.
- Check your plan's sign-in log retention: 7 days or 30.
- Write down who can cut a computer off the network, and who calls your insurer, when an alert turns out to be real.
- Keep updates on schedule and replace computers that no longer get them.
More in this topic
Crashcart runs IT and security for small offices for $35 a computer a month. Start a 14-day free trial.