Study guide

6.3 High-Yield Review & Test Strategy

Microsoft Security Operations Analyst SC-200

Track your progress as you read

Finished sections get a check mark.

1 Introduction & SC-200 Exam Overview

1.1 SC-200 Exam Facts 1.2 Role & Recommended Knowledge 1.3 Skills Measured & Study Strategy

2 Manage a Security Operations Environment: Defender XDR & Sentinel Setup

2.1 Configure Microsoft Defender XDR 2.2 Configure Microsoft Sentinel Workspace 2.3 Data Connectors & Ingestion

3 Manage a Security Operations Environment: Defender for Cloud, Analytics & Automation

3.1 Microsoft Defender for Cloud 3.2 Analytics Rules & Detection Content 3.3 Automation & SOAR

4 Respond to Security Incidents: Defender XDR Workloads

4.1 Respond with Defender for Endpoint 4.2 Defender for Identity, Office 365 & Cloud Apps 4.3 Threat & Vulnerability Management

5 Respond to Security Incidents: Microsoft Sentinel Investigation

5.1 Sentinel Incident Management 5.2 Investigation, UEBA & Threat Intelligence 5.3 Respond & Remediate

6 Threat Hunting & Final Review

6.1 KQL & Threat Hunting 6.2 Custom Detections, Bookmarks & Notebooks 6.3 High-Yield Review & Test Strategy

6.3 High-Yield Review & Test Strategy

Key Takeaways Explain This Topic AI

Map every scenario to the right product first: Defender XDR is unified detection/response across endpoint, email, identity, and cloud apps; Microsoft Sentinel is the cloud SIEM/SOAR for ingestion, analytics rules, and cross-source correlation; Defender for Cloud protects cloud workloads (servers, containers, storage, SQL, APIs).

Domain weighting drives study time: Manage a security operations environment is 40-45%, Respond to security incidents is 35-40%, and Perform threat hunting is 20-25% — spend the most prep on connectors, analytics rules, and automation.

SC-200 is ~40-60 questions in 100 minutes (about 1.5-2 minutes each); flag long case studies and active-screen items, bank the quick multiple-choice points, and never leave a question blank because there is no wrong-answer penalty.

Highest-frequency traps: hunting query vs analytics rule, scheduled vs NRT rule, Defender for Identity (on-prem AD) vs Entra ID sign-ins, automation rule vs playbook, and Analytics vs Data lake retention tier.

Confirm the April 16, 2026 skills outline (three domains) before exam day — guides still listing five domains or a standalone Copilot for Security domain are out of date.

Last updated: May 2026

Cross-Domain Recap: Which Product Owns the Scenario?

The single most valuable SC-200 skill is reading a scenario and instantly identifying the right product . Most wrong answers are plausible-sounding tools from a different product. Anchor on these responsibilities.

Product What it is You use it to

Microsoft Defender XDR Unified extended detection and response across endpoint, email, identity, and cloud apps (portal security.microsoft.com ) Triage the unified incident, run Advanced Hunting, automatic attack disruption, live response, agentic AI investigation with embedded Copilot for Security

Microsoft Sentinel Cloud-native SIEM and SOAR Ingest first/third-party logs via connectors, write analytics rules, correlate across sources, automate with playbooks, Sentinel Graph, Data lake retention

Microsoft Defender for Cloud Cloud workload protection platform (CWPP) and CSPM Protect and alert on servers, containers, storage, SQL, Key Vault, App Service, APIs, DevOps; surface posture recommendations

Microsoft Defender for Identity Identity threat detection for on-premises Active Directory Detect DCSync, Pass-the-Hash, recon, lateral movement on AD

Microsoft Entra ID Protection Cloud identity risk Detect risky users/sign-ins for cloud (Entra) identities

Microsoft Purview Data security and compliance Insider risk, eDiscovery/Content Search, audit, DLP investigations

The two distinctions that decide the most points: Defender for Identity = on-prem AD , while Entra ID Protection = cloud identities ; and Defender XDR is detection/response , while Sentinel is ingestion/correlation/SOAR . A scenario about correlating a firewall log with an Entra sign-in and an endpoint alert is Sentinel; a scenario about isolating a device from its alert is Defender XDR.

KQL Must-Knows

You will not write production KQL, but you must read it confidently. Lock in:

Pipeline order : where (with a time bound) first, then project / extend , then summarize , then sort / top .

String matching : == exact, has whole-term, contains substring, =~ / in~ case-insensitive.

Aggregation : count() total rows, dcount() distinct, make_set() / make_list() collect values, arg_max() / arg_min() latest/earliest row per group.

Time : ago(7d) , between(datetime(..) .. datetime(..)) , bin(Timestamp, 1h) for time bucketing.

join kinds : inner , leftouter , leftsemi , and especially leftanti for "what is missing."

Schema location : Device* / Email* / Identity* / CloudAppEvents in Defender XDR; SecurityAlert / SecurityIncident / SigninLogs / Syslog / CommonSecurityLog in Sentinel.

Domain → Must-Know Table

Domain (weight) Highest-yield must-knows

Manage a security operations environment (40-45%) Connectors: Windows Security Events/Syslog/CEF via AMA , WEF, Azure activities, threat indicators. Analytics rule types: scheduled, NRT, Microsoft security (built-in), threat intelligence, ML/anomaly. Automation rule (orchestrates incidents) vs playbook (Logic App actions). Sentinel RBAC (Reader/Responder/Contributor). Retention tiers: Analytics (fast, query/alert), Auxiliary/Basic , Data lake (cheap long-term, KQL jobs + Summary rules). ASR rules, AIR automation level, automatic attack disruption.

Respond to security incidents (35-40%) Unified incident queue and correlation in Defender XDR. Response per workload: Defender for Office 365 (purge/ZAP), Defender for Endpoint (isolate device, live response, evidence, automatic attack disruption), Defender for Cloud Apps, Defender for Identity, Entra ID, Purview (Content Search, Audit, Graph activity logs). Multi-stage/lateral-movement investigation. Agentic AI investigation with embedded Copilot for Security and case management.

Perform threat hunting (20-25%) Pick the right Advanced Hunting table; read KQL pipelines; hunting query (no alerts) vs analytics rule (alerts). Livestream for live watch. Bookmarks → incidents. Promote hunt → XDR custom detection or Sentinel analytics rule (required columns). Sentinel Graph and blast-radius/entity relationships. KQL jobs + Summary rules in the Data lake tier . Notebooks + Sentinel MCP Server . Threat Analytics for context.

100-Minute Pacing Plan

SC-200 typically delivers 40-60 questions in 100 minutes — roughly 1.5 to 2 minutes per question . Case studies and active-screen (interactive) items take longer, so manage the clock deliberately.

Phase Time budget Action

First pass — standard items ~60 min Answer every multiple-choice/multi-select; flag anything that takes more than ~2 min and move on

  • Case studies / active screen ~25 min Work flagged interactive scenarios and case studies with full attention
  • Final review ~10-15 min Revisit flagged items; verify multi-select count; ensure nothing is blank

Rules of thumb: there is no penalty for wrong answers , so never leave a blank. If a case study lets you review questions before locking the section, read all its questions first so you only parse the scenario once. For multi-select, count exactly how many selections the stem requires.

Common SC-200 Traps

Hunting query vs analytics rule : a hunting query never alerts; if the scenario needs an automatic alert/incident, the answer is an analytics or custom detection rule.

Scheduled vs NRT rule : NRT for lowest latency on a single simple condition (about every minute); scheduled for tunable intervals and joins/aggregation.

Defender for Identity vs Entra ID : "on-premises Active Directory" → Defender for Identity ( IdentityLogonEvents ); "cloud sign-in/risky user" → Entra ID Protection ( SigninLogs ).

Automation rule vs playbook : automation rule orchestrates (assign, tag, close, run playbook); playbook is the Logic App that performs the actions.

Analytics vs Data lake tier : investigate/alert on recent data → Analytics tier; cheap long-term storage with KQL jobs/Summary rules → Data lake tier.

AMA vs legacy agent : current Sentinel connectors use the Azure Monitor Agent (AMA) — Syslog and CEF flow via AMA; the legacy MMA/Log Analytics agent answers are distractors.

Copilot for Security scope : in the 2026 update it is embedded in incident response , not a standalone domain; standalone depth is the SC-5006 Applied Skills credential.

Outdated outline : five-domain or standalone-Copilot-domain descriptions are pre-April-2026 and wrong.

Study-Plan Timeline

Microsoft suggests roughly 80-120 hours over 6-10 weeks . A workable sequence:

Week(s) Focus Outcome

1-2 Defender XDR + Sentinel foundations: portal, workspaces, RBAC, retention tiers, connector planning (AMA) Can place any scenario on the right product

3-4 Detections + automation: scheduled/NRT/TI/ML analytics rules, automation rules vs playbooks, ASR, attack disruption Can choose the correct rule type for a scenario

5-6 Incident response across the stack: Office 365, Endpoint, Cloud Apps, Identity, Entra, Purview; agentic AI + case management Can drive an end-to-end investigation

7-8 Threat hunting + KQL: Advanced Hunting schema, hunting queries, bookmarks, custom detections, Sentinel Graph, Data lake KQL jobs, notebooks/MCP Can read/modify KQL and promote a hunt

9-10 Timed practice: two full timed runs of the 100-question bank + the free Microsoft Practice Assessment; review every miss Consistent ≥80% on practice before scheduling

Go into the exam having confirmed the April 16, 2026 skills outline, with the product-responsibility table memorized and the KQL pipeline second nature. That foundation plus disciplined pacing is what passes SC-200.

Loading diagram... SC-200 Scenario Routing: Pick the Product First