Skip to content

170 practical guides on SAP, cybersecurity and AI, $9.99 each. Shop all books ›

0

Blog›Cybersecurity

What Is a SOC (Security Operations Center)? How It Works

A Security Operations Center is the team, processes and tools that monitor for cyber threats and lead the response. Learn what a SOC does, its tools and roles, build vs buy options, metrics and what to monitor first.

At 3 AM on a Sunday, someone logs into a finance manager’s account from another country and starts downloading files. Who notices? In most mid-size and large organizations, the answer should be the Security Operations Center. It’s the team whose job is to watch for attacks around the clock and act quickly when something looks wrong. You’ll hear “the SOC” in security job ads, audit reports and incident updates, so it’s worth knowing what it actually does.

A Security Operations Center (SOC) is the team, processes and tools an organization uses to continuously monitor its systems, detect cyber threats, investigate suspicious activity and respond to security incidents. A SOC can be built in house, outsourced to a provider or run as a mix of both. It’s often staffed around the clock.

In one sentence: A SOC is an organization’s security watch team: it collects signals from across the business, spots the ones that mean trouble and leads the response.

Key takeaways

  • A SOC combines people, processes and technology to monitor, detect, investigate and respond to threats.
  • Core tools usually include a SIEM, endpoint detection and response (EDR) and automation (SOAR).
  • Many SOCs organize analysts in tiers, from alert triage up to deep investigation and threat hunting.
  • SOCs can be in house, outsourced or hybrid, depending on size, budget and risk.
  • Good SOCs measure detection and response speed, tune out noise and practice incident response.

What does a SOC do?

The day-to-day work follows a cycle:

Function What it involves
Collect Gather logs and signals from endpoints, servers, cloud services, identity systems, networks, email and business applications
Detect Use rules, analytics and threat intelligence to raise alerts on suspicious activity
Triage Decide quickly which alerts are real, which are noise and which are urgent
Investigate Work out what happened, which systems and accounts are involved and how far it has spread
Respond Contain the threat, for example by isolating a device or disabling an account, and coordinate recovery
Improve Tune detections, close gaps found during incidents and report on risk and performance

The exact mix depends on the organization, but the cycle stays the same. Many SOCs also run threat hunting, where analysts look for signs of attackers that haven’t triggered alerts, and vulnerability monitoring, so they know which weaknesses attackers are most likely to use.

What tools does a SOC use?

Tools vary, but most SOCs rely on a common set:

Tool What it does
SIEM (security information and event management) Collects logs from across the organization, correlates them and raises alerts. It sits at the center of most SOCs.
EDR / XDR (endpoint or extended detection and response) Monitors laptops and servers for malicious behavior and can isolate a device remotely
NDR (network detection and response) Watches network traffic for signs of attack
SOAR (security orchestration, automation and response) Runs routine response steps automatically, such as enriching an alert or disabling an account
Threat intelligence platform Brings in information about current attacker tools, infrastructure and techniques
Case management Tracks each incident from first alert to closure

Tools only help if they get the right data. A SIEM with no logs from your cloud apps, identity provider or ERP system has blind spots exactly where attackers like to work. Deciding which data to collect, and paying for its storage, is one of the most important SOC decisions.

Who works in a SOC?

Many SOCs use a tiered analyst model, with supporting roles around it:

Role Focus
Tier 1 analyst Front-line alert triage: confirm whether an alert is real and escalate if needed
Tier 2 analyst / incident responder Deeper investigation and containment of confirmed incidents
Tier 3 analyst / threat hunter Advanced investigations, threat hunting and malware analysis
Detection engineer Writes and tunes detection rules so alerts are accurate
SOC or security engineer Builds and maintains the tools, data pipelines and automation
SOC manager Runs operations, staffing, priorities, metrics and reporting

Smaller SOCs often blend these roles, with a few people covering triage, investigation and engineering together. Larger ones add specialists in areas like cloud security, forensics and threat intelligence.

Tier 1 is a common entry point into cybersecurity careers. It teaches people how real attacks look and how systems behave, though it can also be repetitive and tiring, especially when alert volumes are high.

Should you build or buy a SOC?

Running a 24/7 SOC takes several shifts of trained people plus expensive tools. Many organizations, especially smaller ones, can’t justify that. The main options:

Model How it works Fits
In-house SOC Your own team, tools and processes Large organizations with complex environments or strict requirements
Outsourced SOC / MDR A managed security provider monitors and often responds for you Small and mid-size organizations, or those without security staff at night
Hybrid Internal team during business hours plus a provider for nights, weekends or specialist skills Many mid-size and large organizations

When comparing providers, ask what they actually do at 3 AM: only alert you, or contain the threat themselves? Which of your systems can they see? How fast do they respond, and how do they prove it?

Outsourcing doesn’t outsource responsibility. Someone inside the organization still needs to make decisions during an incident, such as shutting down a system or contacting customers, and the provider needs clear authority and contacts to act fast.

What makes a SOC good?

MITRE’s book 11 Strategies of a World-Class Cybersecurity Operations Center, published in 2022 and free to download, is one of the best-known guides. Its strategies include knowing what you’re protecting and why, giving the SOC the authority to do its job, building a structure that fits the organization, hiring and growing quality staff, prioritizing incident response, selecting and collecting the right data, communicating and collaborating well, and measuring performance to improve it.

In practice, strong SOCs share a few habits:

  • They know the business. An alert on the payroll system at month-end means something different from an alert on a test server.
  • They fight alert fatigue. Too many false alarms train analysts to ignore alerts. Good SOCs tune detections hard and automate routine steps.
  • They have playbooks. Written, practiced steps for common incidents like phishing, compromised accounts and ransomware.
  • They have authority. Analysts can disable an account or isolate a laptop without waiting hours for approval.
  • They practice. Tabletop exercises and simulated attacks test both the tools and the people.

How does a SOC handle a confirmed incident?

When an alert turns out to be a real attack, the SOC moves from monitoring into incident response. Many teams follow the lifecycle NIST describes in its long-standing incident handling guidance:

  1. Preparation: plans, playbooks, contacts, tools and practice, done before anything happens.
  2. Detection and analysis: confirming the incident, understanding its scope and deciding how serious it is.
  3. Containment, eradication and recovery: stopping the spread, removing the attacker’s access and tools, and restoring systems safely.
  4. Post-incident activity: a review of what happened, what worked and what to change, without blame.

Bigger incidents pull in people beyond the SOC: IT operations to rebuild systems, legal counsel on notification duties, communications teams for customers and staff, and senior leaders for decisions with business impact. A good SOC knows who those people are before it needs them, keeps their contact details somewhere that still works if email is down, and runs exercises with them at least once a year.

The post-incident review is where SOCs improve the most. Each incident should leave behind a better detection, a fixed gap or a clearer playbook.

How do you measure a SOC?

Common metrics include:

  • Mean time to detect (MTTD): how long it takes to notice an attack.
  • Mean time to respond or contain (MTTR): how long it takes to stop it.
  • Alert volume and false positive rate: how much noise analysts face.
  • Coverage: which systems and attack techniques the SOC can actually see. Many teams map detections to the MITRE ATT&CK framework to find gaps.
  • Escalation quality: how often escalated alerts turn out to be real.

Be careful with any single number. A SOC can lower its response time by closing alerts too quickly. What counts is catching real attacks early, and a fast-looking dashboard can hide the opposite.

What should a SOC monitor first?

With limited budget, focus on where attacks usually start and what matters most:

  1. Identity: sign-ins, MFA changes, new admin rights and impossible travel. Stolen credentials are a leading way in, as covered in our guide to identity and access management.
  2. Email: phishing reports, suspicious forwarding rules and unusual sending.
  3. Endpoints: EDR alerts on laptops and servers.
  4. Internet-facing systems: VPNs, firewalls and web apps, which attackers scan constantly.
  5. Cloud and SaaS: admin activity, configuration changes and data downloads.
  6. Critical business systems: ERP systems like SAP, payment systems and HR data, which hold the most valuable data and are often missing from SOC monitoring.

That last item is a common gap. Many SOCs see the network and laptops well but have little visibility into what happens inside SAP. Feeding the SAP security audit log and key change logs into the SIEM, with detections for things like emergency access use or bank detail changes, closes it.

How is AI changing the SOC?

AI is being used to summarize alerts, suggest investigation steps, write queries and automate routine triage. Done well, that frees analysts for judgment calls and reduces burnout. It also brings new work: SOCs now need to watch AI agents and tools themselves, including what they access and what they do, and to defend against attackers who use AI to write better phishing and move faster. Human review stays essential for decisions with big consequences, such as shutting down production systems.

A worked example: the 3 AM login

Back to the finance manager’s account. At 3:04 AM, the identity provider flags a sign-in from an unusual country, followed by MFA approval. The SIEM correlates it with a mailbox rule created two minutes later that forwards all invoices to an outside address. The combination triggers a high-priority alert.

The Tier 1 analyst on night shift checks the alert against the playbook. The manager is not traveling, according to HR data, and the device is unknown. She escalates to the on-call responder, and the SOAR playbook automatically revokes the account’s sessions and blocks sign-ins while they investigate. The responder removes the forwarding rule, checks what was accessed and finds the user approved an MFA prompt they didn’t start, a classic push fatigue attack.

By 3:40 AM the account is locked down, by 9 AM the manager has a new password and phishing-resistant sign-in, and finance is warned to verify any payment changes made that week. Without the SOC, the first sign might have been a supplier asking why they hadn’t been paid.

Frequently asked questions

What is the difference between a SOC and a NOC?

A network operations center (NOC) focuses on uptime and performance, keeping systems running. A SOC focuses on security, detecting and responding to threats.

What is a SIEM in a SOC?

A SIEM collects and correlates logs from across the organization and raises alerts. It’s the central tool most SOC analysts work from.

Do small businesses need a SOC?

They need SOC capabilities, such as monitoring and response, but many get them from a managed detection and response provider instead of building their own team.

How do I start a career in a SOC?

Tier 1 analyst roles are a common entry point. Learn networking, operating systems, logs and common attacks, and practice with free labs and capture-the-flag exercises.

The bottom line

A SOC is the team, processes and tools that watch for cyber threats and lead the response when one appears. It runs on good data, tuned detections, clear playbooks and the authority to act fast. Whether you build one, buy one or combine both, make sure it can see identity, email, endpoints, cloud and your most critical business systems, and measure it on catching real attacks early.

Sources: MITRE: 11 Strategies of a World-Class Cybersecurity Operations Center; CyberDefenders: SOC explained; UpGuard: what is a SOC.

Go deeper with these books