Skip to content

Explainer, read 10 Nov 2026

A SOC is the team and tools that watch your systems; MDR is a way to get that team from an outside provider

What is the difference between MDR and a SOC?

A SOC is not a product and not a vendor. It is a function inside an organization, or outsourced for one: the people and process that watch for attacks and respond to them. MDR is a service that supplies some or all of that function. So the comparison is between three ways of getting a SOC: build it, buy it as MDR, or split it.

The UK National Cyber Security Centre says SOCs vary widely in scope but most are responsible for detecting and responding to cyber attacks, and that their role is to limit the damage from attacks that get past preventive controls. It adds that a SOC can include activities such as vulnerability assessment, compliance and system configuration (NCSC, read 10 Nov 2026). That wide scope is why a SOC and an MDR service rarely match one for one.

What the NCSC lists as SOC work

The NCSC guidance is organized around an operating model, onboarding systems and log sources, threat modeling, detection approaches, detection practices, threat intelligence and incident management. It describes onboarding as adding systems to the scope of the SOC so their logs are collected and monitored, and it notes that no IT estate stays the same, so onboarding never finishes.

Read that list as a checklist of jobs. For each one, decide whether your team keeps it, the provider takes it or you share it. Most small teams keep the business context and hand over the night shift.

What a provider takes over

Published service descriptions show typical hand-offs. Expel lists SOC threat analysts, threat hunters, detection engineers and a global response team as parts of one 24x7 operation. Service tiers differ by provider: SentinelOne names three service levels, Threat Hunting, MDR Essentials and MDR Elite. Sophos lists service level targets of 2 minutes to create a case and 30 minutes from case creation to first outreach, evidence collection or response action, and says these are targets and not an SLA.

Each of these is a provider's statement, read on 10 Nov 2026. None of them says the provider replaces your side of the work. Someone in your company still decides what is critical, approves response actions, talks to counsel and runs recovery.

  • Usually handed over: alert triage, 24x7 monitoring, first investigation, containment within agreed limits.
  • Often shared: detection tuning, threat hunting, incident communications.
  • Usually kept: asset knowledge, business priority, approvals, recovery, legal and regulatory decisions.

Three models for the same function

Build: you hire analysts, license tools and run shifts. A week has 168 hours, and one analyst works about 40 of them before leave and sickness, so continuous coverage needs several people before you count a manager. The cost lands in salaries before any tool is bought.

Buy: you contract MDR for the night shift, or for the whole function. The provider owns the shift pattern and you own the escalation path.

Share: your analysts work days, the provider works nights and weekends, and both read the same cases. Deepwatch, for example, says it connects to the customer's existing SIEM, EDR, cloud, identity and SaaS tools, and that its response actions run automatically, with analyst approval or in monitor-only mode, as the customer sets. That is the vendor's description of a shared arrangement. The seam between day and night is where cases get dropped, so agree how a case opened at 6 p.m. reaches your analyst at 8 a.m., and who owns it in between. Put the hand-off in the runbook and test it with a drill alert in the first month.

What to settle before you decide

Count the hours your team can actually cover and the incidents it has handled alone. Write both down for the past year, with dates, so the provider sees the same picture you do. Then write the escalation path: who gets the call, in what order, with what authority. Agree what the provider may do without a call. The NIST incident response publication, SP 800-61 Rev. 3 (April 2025), places incident response inside CSF 2.0 risk management, which fits this view: response is a governance matter as well as a technical one.

Finally, plan the exit. If the provider is your SOC, ask how you would run detections, cases and history on the day the contract ends.

  • Hours covered by your own staff, by week.
  • Actions the provider may take without calling anyone.
  • Data and case history you can export.
  • Who owns detection rules written for you.
  • Escalation contacts and their back-ups.

Sources

  1. NCSC, Building a Security Operations Centre, read 10 Nov 2026
  2. NCSC, Onboarding systems and log sources, read 10 Nov 2026
  3. NIST SP 800-61 Rev. 3 publication page, read 10 Nov 2026
  4. Expel, meet the SOC experts, read 10 Nov 2026
  5. SentinelOne, MDR page, read 10 Nov 2026
  6. Sophos, MDR service description, read 10 Nov 2026
  7. Deepwatch, home page, read 10 Nov 2026
  8. Deepwatch, Active Response, read 10 Nov 2026

Last reviewed 10 Nov 2026.