Guide Cybersecurity

SOC vs MDR: what they are, and which one your business needs

Vendors use "SOC" and "MDR" almost interchangeably in their marketing, but the terms describe different things. A SOC is a function. MDR is a managed service that can complement an existing SOC or deliver the capability to organisations that do not have one. This guide covers what each does, what drives the cost of building round-the-clock monitoring in-house or outsourcing it, and how to tell whether an MDR provider actually responds or only monitors and escalates.

20 September 2026
11 min read
Updated 9 November 2026
Key takeaways
  • A SOC is the function: the people, process and tooling that monitor for signs of threats and respond to them. MDR is a managed service that can complement an existing SOC or provide the capability for organisations that do not have one.
  • One seat staffed around the clock is 8,760 hours a year, about 4.2 full-time posts at 40 hours a week before leave. With holiday, sickness, training and shift rotation, most plans need five to six analysts per seat, plus roles beyond the analysts.
  • Salaries are only part of the bill. Add employer costs for each analyst, then SIEM, EDR and the rest of the tooling.
  • MDR prices depend on the assets covered, the data sources ingested, the response model and the SLA. Compare quotes built on the same scope, not headline per-user rates.
  • Ask providers for MTTD and MTTR with their exact definitions, percentiles and exclusions, and get in writing which actions they can take without your prior approval.

What a SOC does

A Security Operations Centre (SOC) is the function responsible for continuously monitoring for signs of security threats and responding when one appears. It is not a room, a product or a single tool, though large enterprises often build a physical one. It is people, documented process and technology working together around the clock.

A working SOC does four things on a continuous loop. It collects logs and telemetry from endpoints, network devices, cloud platforms and applications, usually through a Security Information and Event Management (SIEM) platform. It correlates that data to surface anomalies. It triages the alerts that correlation produces, separating genuine threats from noise. And it investigates and responds to the ones that turn out to be real, from an infected laptop through to an active intrusion.

Building this in-house has traditionally meant hiring analysts, licensing a SIEM and running shifts to cover nights and weekends. That is a substantial commitment, which is why most organisations outside the large-enterprise segment never had a SOC of their own until managed services made the function available.

What MDR is, and how it differs from a SOC

Managed Detection and Response (MDR) is a service in which a provider delivers detection, investigation and response using its own analysts and, usually, its own tooling. It can complement an existing SOC, for example by covering nights and weekends or a specific set of assets, or provide the capability for an organisation that has none. Most MDR services are built around an Endpoint Detection and Response (EDR) agent, often extended to identity, email and cloud signals as Extended Detection and Response (XDR).

The distinction that matters is response. Some services sold as MDR only alert: an analyst sees something suspicious and emails your IT team a ticket. Providers that do respond tend to use one of three models, and many contracts mix them:

  • Direct response. The provider takes pre-authorised containment actions itself, such as isolating a device, ending a malicious process or disabling an account.
  • Authorised response. The provider prepares the action and carries it out once a named contact on your side approves it, case by case.
  • Guided response. The provider investigates and tells your team exactly what to do, and your team acts.

Direct response is the fastest, and speed matters: CrowdStrike’s 2026 Global Threat Report puts the average eCrime breakout time, from initial access to moving on to other systems, at 29 minutes. Authorised and guided models can work, but only if someone on your side is reachable at any hour.

Check what "response" actually means

Before signing, ask the provider to describe which actions they can take on your network without contacting you first, which need your approval, and how quickly they can reach you at 3am. A service that always waits for approval before acting is mainly a monitoring and escalation service. That can be the right choice, but the contract and the price should say so.

SOC vs MDR vs MSSP vs SIEM

These four terms get used loosely and interchangeably, which makes vendor comparisons harder than they need to be. They describe different layers of the same problem, and the boundaries between them are not fixed.

1
SIEM
The tooling. A platform that collects and correlates logs from across your environment. Many SIEMs now include automation and response playbooks, but someone still has to tune the rules and decide what the alerts mean.
2
SOC
The function. The people and process that watch what the SIEM (or other tooling) surfaces and decide what to do about it, whether built in-house or delivered by a provider.
3
MSSP
Managed Security Service Provider. A broad term for outsourced security services, from firewall management to log monitoring. Offerings vary widely: some MSSPs mainly alert and escalate, others include active response or run an MDR service.
4
MDR
A managed service specialised in detection and response, usually built on EDR or XDR technology and combining continuous monitoring with human-led investigation and an agreed response model.

Why most businesses can't build their own 24/7 SOC

The arithmetic of round-the-clock coverage defeats most in-house plans before the tooling budget comes up. One seat staffed 24/7 is 8,760 hours a year, roughly 4.2 full-time posts at 40 hours a week before anyone takes leave. Add annual leave, sickness, training and the limits of sustained night work, and a resilient 24/7/365 seat usually needs five to six analysts.

Salaries vary by region, sector and security clearance requirements, so price the team from current salary data for your own market rather than a national average. Then add employer costs to each salary: National Insurance contributions, pension contributions and the cover needed for holidays and training. Across five or six analysts, payroll alone becomes a significant annual commitment before any tool is licensed.

Analysts are not the whole team. A functioning SOC also needs someone to lead it, engineers to build and tune detection rules and maintain integrations, and people able to run a serious incident. In a small team one person may cover several of these roles, but the work still has to be done.

Tooling adds further cost on top of headcount. SIEM licensing is typically priced by data volume ingested, and costs scale with the size of your environment. EDR licensing is priced per endpoint. Threat intelligence feeds, case management software and on-call escalation tooling add more. None of this is optional if the SOC is meant to function rather than exist on paper.

5 to 6
analysts typically needed to staff one 24/7/365 seat with leave cover
8,760
hours a year in one 24/7 seat, about 4.2 full-time posts before leave
247 days
global mean time to identify and contain a breach, IBM Cost of a Data Breach Report 2026

MDR is usually priced per endpoint, per user or per asset, and quotes vary widely. The main cost drivers are the number and type of assets covered, which data sources are ingested beyond the endpoint (identity, email, cloud, network), whether EDR licences are included or billed separately, the response model and SLA, log retention, and whether incident response hours are included. Ask each shortlisted provider to quote on the same scope so the numbers can be compared.

For most organisations below several hundred staff, a managed service is the realistic way to get 24/7 coverage at all. The decision is about which scope to buy.

Check what the service covers

Many MDR services start with endpoints and add other sources as options. Map your environment against what the provider will monitor, and write down what stays outside the service.

AreaWhat to confirm
EndpointsWhich operating systems are supported, how agents are deployed and what happens when a device stops reporting
ServersOn-premises and cloud servers, including Linux, and whether critical servers can be isolated without breaking services
Identity and emailYour identity provider and email platform, and whether the provider can disable accounts and revoke active sessions
CloudWhich cloud platforms and SaaS logs are ingested, and whether configuration changes are monitored
NetworkFirewall, VPN and DNS logs or network sensors, and whether they are included or charged as an extra
Critical applicationsBusiness systems such as ERP or patient records: whether their logs can be ingested and who writes the detection rules for them

What a good MDR service includes

Pricing pages make MDR services look interchangeable. In practice, the difference between a strong provider and a weak one shows up in six areas.

Genuine 24/7/365 coverage. Not "monitored during business hours with an on-call escalation", but analysts actively watching alerts around the clock, including bank holidays.

A managed EDR or XDR agent. The provider deploys, tunes and maintains the detection technology itself, rather than expecting your IT team to keep an agent installed, up to date and working correctly on every device.

Threat hunting, not just alert response. The strongest providers proactively look for signs of compromise that automated rules miss.

An agreed response model. A documented, signed-off list of the actions the provider can take without contacting you first, such as isolating a host, disabling an account or blocking a malicious domain, and of the actions that need your approval.

Reporting with context. A monthly report that explains what was investigated and why it mattered, not a page of alert counts with no narrative.

A defined handoff to remediation. MDR detects and contains. Rebuilding a compromised machine, resetting credentials, recovering affected systems and fixing the gap that let the incident happen is typically your IT team or MSP’s job. Know where that line sits before you need it.

The metrics that separate good MDR from a dashboard

Two numbers come up in every MDR sales conversation: mean time to detect (MTTD) and mean time to respond (MTTR). They only help if you know how the provider measures them. Providers define MTTR differently: some stop the clock when an analyst acknowledges the alert, others when the threat is contained or the case is closed. Ask for:

  • the exact start and end point of each metric
  • medians and 90th or 95th percentiles, not only averages
  • what is excluded, such as time spent waiting for your approval
  • time to triage, time to notify you and time to contain, measured separately
  • the percentage of critical incidents that met the contractual target over the last quarter
  • how many of your assets are sending telemetry, since a device that does not report cannot be monitored

Speed matters because breaches that run longer cost more. IBM’s Cost of a Data Breach Report 2026 puts the global mean time to identify and contain a breach at 247 days: 183 to identify and 64 to contain. That figure covers organisations of every kind, with and without managed detection, so it is not comparable with a provider’s MTTD or MTTR. It does show how long attackers can go unnoticed when nobody is reviewing the alerts.

The question is not whether you need monitoring. Every business already generates the logs. The question is whether anyone is watching them at 3am, and what they are allowed to do if something is wrong.

How to evaluate an MDR provider

Once shortlisted providers all claim 24/7 coverage and rapid response, these questions separate the ones that deliver from the ones that describe.

Which actions can you take without prior approval? Get the response model in writing before signing. A vague answer here is the biggest red flag in an MDR evaluation.

How is your automation validated? Automated triage is normal. Ask which detections are closed automatically, how often those rules are reviewed, and how analysts check that automation is not discarding real threats.

Where is the SOC, and how many analysts are on shift overnight? Ask directly rather than accepting a marketing claim. A skeleton overnight team changes what "24/7" means in practice.

Can it ingest what you already run? If you already have an EDR agent, cloud logging or an identity provider in place, confirm the provider can work with it rather than forcing a full replacement.

How does onboarding work, and how are coverage gaps reported? Ask how long deployment takes, what the provider needs from your team, and how you will be told when assets stop reporting or a data source is missing.

Which SLA commitments are in the contract? Check the targets for notification and containment, how they are measured and what happens when they are missed.

Where does MDR end and incident response begin? Containment is part of MDR. Full investigation, forensics and recovery after a serious incident may need a separate incident response retainer. Confirm what is included, how many hours, and who does the remediation and rebuild work.

How are your logs accessed, stored and retained? Ask where analysts work, where log data is stored, how long it is kept and how this fits your UK GDPR obligations, including any transfers outside the UK.

Frequently asked questions

Is MDR the same as antivirus or EDR?

No. EDR (Endpoint Detection and Response) is the technology installed on your devices that generates detection data. MDR is the managed service built around that technology: the analysts who watch what it surfaces and act on it. You can use EDR without MDR, but then someone on your own team has to review its alerts, including overnight.

How much does MDR cost for a small business?

It depends on scope. Providers price per endpoint, per user or per asset, and the quote moves with the data sources covered, the response model, the SLA and whether EDR licences and incident response hours are included. Get quotes based on your actual user and device count, with the same scope from each provider.

Can our existing IT provider or MSP add MDR on top of what they already do?

Some can, often by partnering with or reselling a specialist MDR platform rather than running the SOC themselves. That can work well, but ask specifically who is watching alerts at 3am and what authority they have to act, rather than assuming your MSP’s helpdesk doubles as a security operations team. The two functions need different skills and staffing models.

Do we still need our own IT team if we have MDR?

You still need an IT function, whether in-house or outsourced to an MSP. MDR covers detection and response to security threats. It does not cover patching, helpdesk tickets, asset management or rebuilding a machine after an incident is contained. Agree a documented handoff for who does what once a threat has been dealt with.

Ryland Deakin
About the author
Lead Consultant, Cyvra · CISM · CompTIA Security+ · MCP

Ryland has delivered cybersecurity, compliance, and IT management programmes for regulated organisations across the UK and the Netherlands for over 20 years. View full profile

Talk to Cyvra

Work out whether you need a SOC, MDR, or both

We help UK and Netherlands businesses decide between in-house monitoring, MDR, and hybrid models, then design the logging architecture and incident response playbook to back whichever one you choose.

Disclaimer: This article is for general informational purposes only and does not constitute legal, regulatory, or professional advice. Cyvra makes no warranty as to the accuracy or completeness of this content, which may not reflect the most current regulatory developments. Readers should seek independent legal and regulatory advice appropriate to their specific circumstances. Cyvra accepts no liability for any loss arising from reliance on this content.