
Build a SOC escalation policy with severity levels, escalation matrix, response times and Singapore reporting rules, so every alert reaches the right person fast.

SOC Escalation Policy: How to Write One That Works

A SOC escalation policy is the written rule set that says which alerts go to whom, how fast, and through which channel when a security operations centre (SOC) cannot close an event on its own. A good one defines severity levels, names a primary and a backup person for each level, sets response clocks, and is rehearsed before it is needed. In Singapore it also has to line up with statutory reporting duties. The best way to keep the policy short and enforceable is to cut the noise upstream, so that operators escalate exceptions rather than raw alarms. That is the design principle behind our own monitoring platforms.
This guide is written for security managers, facility teams, integrators and agencies who run or build a SOC, covering physical security operations and the cyber incidents that spill into them.
TL;DR
- An escalation policy answers four questions for every event: how severe is it, who must know, how fast, and who takes over if that person does not respond.
- Use four severity levels with measurable triggers, a named primary and backup contact per role, and a response clock per level. Keep the matrix to one page.
- In Singapore, owners of Critical Information Infrastructure (CII) must call the Commissioner of Cybersecurity within 2 hours of a notifiable incident, so your internal clock has to be faster than that.
- Test the matrix with a one-hour tabletop drill. Every hesitation or missing phone number is a defect to fix.
- Fewer, better-qualified alerts make any escalation policy work. Exception-based monitoring protects the people at the top of the chain from noise.
Table of Contents
- What an escalation policy does in a SOC
- Who sits where: the tiers behind the matrix
- Define severity with triggers, not feelings
- Build the escalation matrix
- Set response clocks that fit Singapore's reporting rules
- What every handoff must contain
- Reduce what reaches the top of the chain
- Promo: one platform behind the policy
- Test, document and improve the policy
- A starter checklist for your first draft
- FAQ
- Sources
What an escalation policy does in a SOC
The policy turns a judgment call made at 3 a.m. into a decision someone already made in daylight. Without it, operators decide for themselves whether to wake a supervisor, and they decide differently depending on the shift, their confidence and who is on call.
A workable policy has five parts:
- Severity definitions with objective criteria, not adjectives.
- Escalation triggers, the specific conditions that move an event up a level.
- Escalation paths, meaning who is contacted, in what order, and by what channel.
- Time targets for acknowledgment, containment and notification.
- Handoff content, the minimum facts the receiving person needs.
Singapore's regulator treats this as a basic obligation for the most critical operators. The Cybersecurity Code of Practice for CII, issued under the Cybersecurity Act 2018, requires each CII owner to keep an incident response plan with defined roles, responsibilities and contact details, communication structures that ensure timely escalation to the response team and senior management, and thresholds and procedures to activate the response. Even if your site is not CII, that list is a sound template for any SOC.
Who sits where: the tiers behind the matrix
Most SOCs use a tiered model. Tier 1 operators watch alerts, filter false positives and escalate what they cannot resolve. Tier 2 responders investigate, judge the scope of the problem and decide on containment. Tier 3 specialists take on major incidents and hunt for threats that no rule has caught yet.
The policy should say what each tier is allowed to decide alone. A common failure is an unclear Tier 1 mandate: operators either escalate everything to protect themselves or sit on events because nobody told them they could act. Write down the actions Tier 1 may take without approval, such as dispatching a guard, locking a door group or isolating a camera network segment, and the events that always go up.
Above the SOC, the structure splits into decision and execution. The SGX RegCo guide for listed companies, summarized by Rajah & Tann, recommends a Crisis Management Team of senior management that makes the key decisions, and a Cyber Incident Response Team drawn from relevant functions that runs the plans and carries out those decisions. Your matrix should show where the SOC hands over to each.

Define severity with triggers, not feelings
Four levels are enough for most operations. Fewer than four blurs the middle, and more than four slows classification.
| Level | Typical trigger | Who is told | Target clock |
|---|---|---|---|
| SEV1 / P1 | Active data theft, ransomware, intrusion in progress, life-safety threat, loss of a critical site | SOC lead, incident commander, executive, legal | Immediate, 24/7 |
| SEV2 / P2 | Confirmed compromise or breach of a single area, major system degradation | Incident commander, technical lead | Within about 30 to 60 minutes |
| SEV3 / P3 | Suspicious but unconfirmed activity, contained minor issue | Technical lead, shift supervisor | Within a few business hours |
| SEV4 / P4 | Informational alert, policy deviation | Assigned analyst | Next business day |
The tiers above follow the pattern in published templates. One enterprise playbook template sets SEV1 as critical systems down, active exfiltration or ransomware with immediate 24/7 response, SEV2 within one hour, SEV3 within four hours and SEV4 the next business day. A sample incident response plan from a Singapore software company uses three levels (Low, Medium, High) with a role table that lists a primary, backup and emergency contact method for each role. Treat both as starting points and set the numbers to your own risk, contracts and staffing.
Write triggers a night operator can apply
For each level, list two or three conditions that can be checked in under a minute. "Perimeter breach on a critical zone confirmed by video" is a trigger. "Serious security event" is not. Add an upgrade rule as well: any event that remains unresolved past its clock moves up one level automatically, and any event touching life safety or regulated data starts at SEV1.
Physical and cyber events share one matrix
In converged sites, a tailgating event at a server room, a camera network compromised through a weak protocol or a contractor badge used out of hours can start as a physical alarm and become a cyber incident. Use one severity scale for both, so the operator does not have to guess which procedure applies. Our guides on alarm triage workflow and security monitoring workflow for facility teams show how to get an alarm to a classified, verified state before it enters the matrix at all.
Build the escalation matrix
The matrix is the one-page working document. For each severity level it lists roles, not names alone, because people change jobs and take leave.
| Role | Primary | Backup | Emergency channel | Activated at |
|---|---|---|---|---|
| SOC shift supervisor | Name, mobile | Name, mobile | Phone call | SEV3 and above |
| Technical lead | Name, mobile | Name, mobile | Phone call, then chat | SEV2 and above |
| Incident commander | Name, mobile | Name, mobile | Phone call | SEV1, SEV2 |
| Site or facility head | Name, mobile | Name, mobile | Phone call | Physical impact |
| Legal and compliance | Name, mobile | Name, mobile | Phone call | Data or regulatory exposure |
| Executive sponsor | Name, mobile | Name, mobile | Phone call | SEV1 |
Four rules keep it honest:
- One call, then a record. Notify by phone for SEV1 and SEV2, and log it in the ticket. A chat message is not an acknowledgment.
- Backups are mandatory. If the primary does not acknowledge within a set number of minutes, the system or the operator moves to the backup without asking permission.
- Parallel, not serial, for the top level. For a suspected ransomware case, one published example escalates from analyst to technical lead to incident commander, who then notifies legal and executive leadership at the same time rather than one after the other.
- Include external parties. Insurers, vendors, system integrators and regulators belong in the matrix with their own time targets.
The same enterprise template gives a stakeholder timetable you can adapt: executives within 15 minutes for the two highest levels, legal counsel within an hour where a data breach or regulatory exposure is possible, communications staff within two hours when customers are affected, third-party vendors within four hours and cyber insurers within 24 hours.
Set response clocks that fit Singapore's reporting rules
Internal clocks must be tighter than external statutory ones, or the policy fails when it matters most. According to a 2026 practice guide on Singapore cybersecurity, under the CII incident notification regime as expanded from 31 October 2025, a CII owner must call the Commissioner of Cybersecurity within 2 hours of the incident or trigger circumstance, supply written details within 72 hours and file a final report within 30 days after that. The triggers now also cover effects visible to the public, zero-day exploitation and suspected advanced persistent threat (APT) activity.
If the law gives you two hours to notify, a SOC that needs 90 minutes just to reach an incident commander has no margin. Work backward:
- Minutes 0 to 15: the operator classifies the event and calls the incident commander.
- Minutes 15 to 30: the commander confirms the level, activates the response team and tells the executive sponsor.
- Minutes 30 to 90: legal and compliance decide whether the event is notifiable and prepare the call to the authority.
- Before the 2-hour mark: the notification is made, if required.
Not every site is CII. The Cybersecurity Act also gives the Commissioner powers over serious cybersecurity threats or incidents, including directing remedial measures and requiring affected systems to be preserved. SingHealth in 2018 is the well-known case: the incident was escalated on 10 July 2018 to the Cyber Security Agency, senior management and the health ministries, and a national response team was deployed on site. Your policy should say when you call national responders and who makes that call.
Where an owner has suppliers that interconnect with the CII, the picture widens. A 2024 amendment, as debated in Parliament, extended reporting to other computers under the owner's control and to those under an immediate supplier's control that connect to the CII. Add your integrators and managed service providers to the matrix, with a contract clause that obliges them to alert you within a set time.
For wider compliance context across the region, see our overview of security compliance frameworks for Southeast Asia.
What every handoff must contain
A good handoff lets the receiver act without calling back to ask basic questions. Use a fixed template, and keep it to the facts:
- What happened, in one sentence, and the severity assigned.
- Where: site, zone, system, asset or user.
- When it started and when it was detected.
- Evidence seen so far: video clip, log line, access record, sensor reading.
- Actions taken and by whom, so nobody repeats or reverses them.
- What is needed from the receiver and by when.
Trainers sometimes call this the five Ws. The point is that the ticket, not the operator's memory, carries the story. Where logs and video sit in separate tools, the handoff takes longer and mistakes creep in. Access data matters here too: role-based access control and contractor access control records often decide whether an incident is SEV2 or SEV3 within the first few minutes.
Reduce what reaches the top of the chain
Escalation policies fail from overload more often than from design errors. If operators see hundreds of alarms per shift, most of them false, they learn to dismiss, and the one real event is judged like the rest. The fix sits upstream: verify and filter before you escalate.
In physical security, that means giving the operator a verified picture of the event rather than a bare alarm code. Video analytics, virtual patrol records and visitor data can be correlated so that the operator receives an exception, a candidate event with context attached. This is what we mean by exception-based monitoring: automation handles the routine, and people handle the cases that need judgment. It keeps SEV3 and SEV4 volumes down, shortens classification time, and leaves the senior contacts in your matrix to be woken only when it counts.

Promo: one platform behind the policy
Our BeyondWatch, BeyondPatrol, BeyondVisitor and BeyondSecure product lines bring AI-powered surveillance software, virtual guard tours, automated visitor management and unified security operation dashboards into one multi-vendor platform. For integrators and security teams in Singapore, Malaysia and the Philippines, that gives the SOC one place to verify an event before it enters the escalation matrix.

If you are designing a SOC or rewriting its procedures, our Solution Integration service is the place to start a conversation about fitting escalation into your deployment.
Test, document and improve the policy
A matrix that has never been used is a hypothesis. Three practices turn it into a working tool.
Distribute and version it. Put the escalation tree in the incident response plan, share it with every stakeholder, and review it whenever staff, vendors or systems change. A contact list older than a quarter is already wrong.
Run a one-hour tabletop. Use a realistic SEV1 scenario, such as ransomware reaching a building management network or a breach of a data centre cage combined with a camera outage. Time each step and write down every point where someone hesitates or cannot find a number. Those notes are the revision list.
Keep a rhythm. A typical schedule pairs an annual executive tabletop with quarterly technical tabletops and contact verification, plus annual ransomware and data-breach simulations and a red team exercise. The Singapore CCoP also requires a post-incident review that identifies and implements corrective measures, so build that review into the policy as its closing step.
For templates, the Cyber Security Agency's SingCERT publishes an incident response checklist built around the NIST framework (Identify, Protect, Detect, Respond, Recover), which stresses that a rehearsed plan limits damage, and a set of incident response playbooks that includes a malware infection playbook and a ransomware response checklist. Use them to check that your matrix covers the scenarios most likely to hit you.
A starter checklist for your first draft
- List your top ten incident scenarios, physical and cyber.
- Assign each to a severity level with a measurable trigger.
- Name a primary and backup for each role, with an emergency channel.
- Set acknowledgment and notification clocks per level, tighter than any statutory limit.
- Add the external contacts: regulator, insurer, integrator, key vendors.
- Write the handoff template and put it into the ticketing tool.
- Drill it within the first month, then revise.
- Review after every SEV1 and SEV2 event, and at least twice a year regardless.
Our view: escalate exceptions, not alarms
We build and deploy sensor-based security systems across Singapore, Malaysia and the Philippines, and the pattern we design for is simple: an escalation policy is only as strong as the quality of the signal feeding it.
That is why we focus on automation at the front of the chain. Our platform combines AI-driven tools, multi-vendor integration and localized technical validation so that regional teams can meet their compliance expectations. In our approach, AI never blinks, so the routine verification and patrol work can run continuously, and your people are brought in when a situation is a real exception.
Write the matrix first, then choose technology that makes each severity level easier to classify and each handoff faster. A unified dashboard that shows the event, its video, its access trail and its history in one view is what lets a Tier 1 operator make the call the policy asks of them.
Eumir
FAQ
What is the difference between an escalation policy and an escalation matrix?
The policy is the full rule set: severity definitions, triggers, time targets and responsibilities. The matrix is its one-page table that lists who to contact at each level and how. Most teams keep both in the incident response plan, with the matrix printed and available at every console.
How many severity levels should a SOC use?
Four is the most common choice, labelled SEV1 to SEV4 or P1 to P4. Each level needs an objective trigger, a named recipient and a time target. Three levels can work for small teams, as long as the top level is reserved for events that need an immediate 24/7 response.
Who should be able to declare a SEV1?
Any Tier 1 operator should be able to declare it when a listed trigger is met, without seeking approval first. The incident commander then confirms or downgrades the level. This keeps the first minutes fast and puts accountability for the final classification with a senior person.
How often should we test the escalation policy?
Run a one-hour tabletop drill before you rely on the policy, then repeat technical tabletops and contact checks quarterly and an executive exercise yearly. Revise the policy after every real SEV1 or SEV2 event and whenever roles, vendors or systems change.
Sources
- Cybersecurity Code of Practice for Critical Information Infrastructure, 2nd Edition Rev 1
- SGX RegCo Cyber Incident Response Guide, summary
- Incident Response Playbook Template
- Incident Response Plan (IRP), Singapore sample
- Cybersecurity 2026, Singapore trends and developments
- Panoramic Cybersecurity, Singapore
- Singapore Cyber Landscape 2018
- Cybersecurity (Amendment) Bill debate
- SingCERT Incident Response Checklist
- SingCERT Incident Response Playbooks
Recommended
Read More Articles

CCTV Tender Specifications in Singapore: What to Write and Check
How to write or answer a CCTV tender specification in Singapore: camera, retention, network, analytics and acceptance requirements, with SPF, MOM, IMDA and PUB figures.

Alarm Deduplication for Security Teams: A Practical Guide
How alarm deduplication, correlation rules, severity mapping and escalation matrices work together to cut duplicate alarms without hiding the ones that matter.

Vehicle Classification Analytics: A Guide for Singapore Sites
How vehicle classification analytics and vehicle type detection work, which classes to use in Singapore, and how to plan camera placement, integration and data.

Evacuation Muster Reporting: Singapore Site Guide
How to run evacuation muster reporting and roll call on Singapore sites: who counts whom, what the report records, SCDF duties, and how to close the gaps.
Let's Build YourSecurity Ecosystem.
Whether you're a System Integrator, Solution Provider, or an End-User looking for trusted advisory, our team is ready to help you navigate the BeyondSensor landscape.
Direct Advisory
Connect with our regional experts for tailored solutioning.