Lesson 21 — The Successful Sign-In Wasn't Normal
The authentication result said Success.
The password worked. MFA was satisfied. Conditional Access did not block the session.
If the analyst looked only at the result code, there was nothing to investigate.
But the user normally signed in from Melbourne on a managed Windows laptop.
This session appeared from an unfamiliar location, used a different client and was followed by activity the user had never performed before.
Success was only one field
The verdict depends on the wider sign-in story: identity, location, application, device, authentication and what happened next.
Case briefing
Investigation objective
Investigate successful authentication by comparing the event with the user's normal sign-in behaviour and examining what happened before and after access was granted.
Investigator's rule
Success is an authentication outcome, not a security verdict. Always interpret it with identity, device, location, application and session context.
Stage 1 — establish what normal looks like
| Baseline field | Normal pattern | Current sign-in |
|---|---|---|
| Location | Melbourne / Victoria | Unfamiliar location |
| Device | Managed Windows laptop | No familiar managed-device context |
| Application | Microsoft 365 applications | Microsoft 365 |
| Authentication | Password + MFA | MFA satisfied |
| Typical hours | Business hours | 07:42 |
| Follow-on activity | Normal mailbox usage | Unusual mailbox activity |
One anomaly is rarely enough
A user can travel, replace a device or work unusual hours. The investigation becomes stronger when several independent contextual differences appear together.
Baseline the identity, not the organisation
What is normal for an administrator, salesperson and finance user may be completely different. Compare the sign-in with the behaviour of the specific identity whenever possible.
Stage 2 — reconstruct recent sign-ins with KQL
Start with the identity and compare the suspicious event against recent successful and failed authentication activity.
let TargetUser = "s.bennett@contoso.com";
SigninLogs
| where TimeGenerated > ago(7d)
| where UserPrincipalName =~ TargetUser
| project TimeGenerated,
UserPrincipalName,
IPAddress,
Location,
AppDisplayName,
ClientAppUsed,
ResultType,
ResultDescription,
ConditionalAccessStatus
| order by TimeGenerated desc
Look before the success
Failed attempts immediately before a successful sign-in can materially change the story. The successful event may be the end of an authentication sequence rather than the beginning.
Look beyond the IP address
An unfamiliar IP is a lead, not proof. Location, client, device context, authentication details and subsequent behaviour help determine whether it matters.
Stage 3 — build the authentication sequence
MFA success requires interpretation
MFA being satisfied is important evidence, but it does not automatically resolve who controlled the session. The analyst still needs to understand the authentication method and surrounding behaviour.
Do not jump straight to compromise
Suspicious context raises the priority of the investigation. It does not remove the need to test legitimate explanations such as travel, approved remote access or a changed device.
Stage 4 — compare source patterns
let TargetUser = "s.bennett@contoso.com";
SigninLogs
| where TimeGenerated > ago(30d)
| where UserPrincipalName =~ TargetUser
| summarize
SignInCount=count(),
FirstSeen=min(TimeGenerated),
LastSeen=max(TimeGenerated)
by IPAddress,
Location,
AppDisplayName,
ClientAppUsed
| order by LastSeen desc
First-seen context matters
A source combination that has never appeared in the user's recent history deserves closer inspection, especially when other anomalies occur at the same time.
Baselines are evidence, not rules
Thirty days of history can tell you what has been observed. It cannot prove that anything new is malicious. Use the baseline to focus the investigation.
Stage 5 — follow what happened after authentication
Post-authentication behaviour can be decisive
If a suspicious sign-in is followed by behaviour the user has never performed before, the combined evidence can be substantially stronger than the sign-in anomaly alone.
Keep the timeline chronological
Do not start with the conclusion and search only for supporting evidence. Record what happened before, during and after the authentication event in time order.
Stage 6 — test competing explanations
| Hypothesis | Evidence to test |
|---|---|
| User travelling | Travel context, known device, expected applications, user validation. |
| Approved VPN / proxy | Known corporate egress, expected client and historical use. |
| New legitimate device | Device registration, management state and user confirmation. |
| Credential compromise | Preceding failures, unfamiliar source, abnormal session activity and identity changes. |
| Session / token abuse | Authentication details, session behaviour and activity inconsistent with normal interactive use. |
Stage 7 — make the triage decision
Write the investigation finding
Lesson 21 key takeaways
- A successful sign-in is an authentication result, not proof of legitimate user activity.
- Compare suspicious authentication with the normal behaviour of the specific identity.
- Location, IP address, device, client application and authentication context should be evaluated together.
- Failed attempts before a successful sign-in can materially change the interpretation.
- MFA satisfaction is important evidence but does not automatically end the investigation.
- Unfamiliar IP addresses are leads, not standalone verdicts.
- Use KQL to reconstruct the user's recent sign-in timeline and source baseline.
- Follow the session after authentication to identify unusual mailbox, cloud, privilege or application activity.
- Test legitimate explanations instead of searching only for evidence of compromise.
- Make the final decision from the complete identity timeline.
Module 3 — Identity Incidents
Module 3 begins with one of the most important identity-investigation lessons: successful authentication can still be suspicious. Next, Agent Foskett investigates a user who received twelve MFA prompts.
Continue your SOC Analyst training
🔎 SOC Analyst Academy — Module 3: Identity Incidents
How to investigate a suspicious successful Microsoft Entra sign-in
Lesson 21 of the Agent Foskett SOC Analyst Academy teaches analysts how to investigate successful authentication by comparing location, device, client, application and authentication context with the user's normal sign-in behaviour.
KQL investigation of successful sign-ins and identity incidents
Learn how to use Microsoft Entra SigninLogs and KQL to reconstruct authentication timelines, identify unfamiliar sign-in patterns, review preceding failures and determine whether post-authentication activity supports escalation or containment.
