Threat intelligence analysis turns collected observations into a judgment a decision owner can test and use. Start with one decision, trace every material claim to its evidence, challenge the leading explanation, and state what remains unknown. The result should be a bounded assessment, not a pile of alerts or a claim of certainty.
This guide starts after collection and initial source checks. The threat intelligence lifecycle guide covers the wider process, while the threat intelligence collection plan addresses what to gather. Here, a fictional sign-in pattern shows how to move from several observations to a defensible monitoring decision.
Which Decision and Time Horizon Should Threat Intelligence Analysis Answer?
Ask what the recipient can decide before choosing an explanation. In the fictional example, a SOC lead must decide whether to increase monitoring for a business application during the next review period. The question is narrower than “Is the organization under attack?” It names an action, an owner, and a time horizon.
Write the decision question at the top of the working record: “Does this sign-in pattern justify targeted monitoring while account ownership is checked?” Then identify what would change the answer. A confirmed scheduled test could close the concern, while an unauthorized account change could trigger incident triage.
The priority intelligence requirements template helps define the earlier requirement. At analysis time, narrow that requirement to one decision and the evidence available by its deadline. The Cyber Threat Intelligence platform is the product route for teams connecting threat context with defensive decisions.
NIST SP 800-150 frames cyber threat information by how it helps organizations assess, monitor, and respond to threats. Its information-sharing guidance supports purposeful use of sources, though it does not prescribe this worksheet.
Decide what the assessment can safely say before evidence review begins. A sign-in anomaly can justify a temporary check without proving account misuse. Conversely, a short decision window may justify a reversible monitoring action while a key fact is still being verified.
How Do You Separate Observations, Source Claims and Analyst Judgments?
Use separate labels for what a record shows, what a person reports, and what the analyst concludes. In the fictional case, an authentication log records repeated failed attempts against a service account. A help desk note says no related user complaint was reported. Neither statement by itself establishes intent or compromise.
A configuration record shows that an integration was recently changed. That record is an observation about a change, not proof that the change caused the failures. The analyst may judge that the pattern warrants monitoring while the account owner explains the change. Label that conclusion as a judgment, and write down the causal assumption it depends on.
ODNI’s ICD 203 distinguishes underlying information from assumptions and judgments. It asks analysts to identify assumptions that bridge important gaps and explain how a judgment changes if they fail. The directive governs United States Intelligence Community products; its distinction is a useful quality reference here, not a mandatory CTI standard.
Keep timestamps, source provenance, access limits, and handling restrictions beside each observation. A copied alert may trace back to the same log as another report. Treat those as one underlying observation until independent collection is established. NIST SP 800-150 warns that unknown original sources can make independent confirmation hard to establish.
Which Evidence Supports or Challenges the Leading Explanation?
Make a small evidence ledger before writing the conclusion. The leading explanation here is that the failed attempts deserve targeted monitoring because the account activity is not yet explained. A competing explanation is that an approved integration test generated the same pattern. Both explanations must face the same records.
| Item | What It Supports | What It Cannot Establish |
|---|---|---|
| Authentication log | Failed attempts occurred within the review window. | Who initiated them or why. |
| Configuration record | An integration changed near the same period. | Whether that change caused the failures. |
| Help desk note | No related complaint was recorded in that channel. | That no user or service was affected. |
| Account owner reply pending | A route to test the authorized-test explanation. | Any conclusion before the reply arrives. |
Read supporting and contrary evidence together. The configuration record makes the authorized-test explanation plausible, but timing alone is weak causation. The help desk note is a narrow absence in one channel, not a clean bill of health. The account owner’s change record could resolve the most consequential gap.
Brandefense Blackhole recorded 1,353 dark-chat posts matching a broad “threat intelligence” query from August 24 through November 22, 2026. The unit is matching posts, not verified incidents, distinct actors, or affected organizations. That aggregate cannot corroborate this fictional sign-in case; it illustrates why a relevant-looking count needs a defined question and unit.
For a real case, document whether two reports share an original author, sensor, or copied claim. Do not raise confidence merely because the same statement appears in several feeds. NIST SP 800-150 describes independent confirmation and reduction of ambiguity as ways to enrich shared threat information.
How Do You Test an Alternative Explanation Before Escalating?
Choose an alternative that could change the action, then name a discriminating check. Here, the integration-test explanation predicts an approved change window and an accountable owner who can explain the attempts. Ask for the change record and compare its scope and timing with the log. Do not treat an informal reassurance as the record itself.
The unauthorized-activity explanation predicts that the owner cannot account for the pattern, or that related account changes appear outside the approved work. Check only the records needed for the decision. An analyst should not invent a full incident narrative when the current question is whether monitoring should increase.
ODNI’s ICD 203 calls for plausible alternative hypotheses, including indicators that would affect their likelihood. In this guide, a signed change record would weaken the need for escalation. An unexplained account modification would strengthen it. Those are proposed tests for a fictional case, not claims about a real event.
If the account owner cannot respond before the decision deadline, say so. The SOC lead can authorize a narrow, reversible monitoring change under local procedures and set a review time. Escalation should follow the organization’s incident criteria when new evidence meets them. The analyst’s role is to show what the evidence supports now.
How Should Confidence and Unknowns Be Stated Without False Precision?
Confidence describes the strength of the basis for a judgment. Likelihood describes the chance of an event or explanation. Keep them separate, and explain the sources behind either term. ODNI’s ICD 203 explicitly distinguishes the two and asks for the causes of uncertainty.
For the fictional case, state: “We assess with moderate confidence that targeted monitoring is warranted while ownership is verified.” Explain the basis separately. The log shows a repeatable pattern, but the integration change has not been explained. Then state the limit: this assessment does not establish malicious activity or account compromise.
Avoid converting missing evidence into evidence of absence. The lack of a help desk complaint says only that this channel recorded none during the review window. It does not show whether the service account was used appropriately. The unreturned owner request is an open gap, not a negative finding.
Do not assign a numerical score merely to make the sentence look precise. The underlying records do not support a calibrated probability. Record what would raise or lower confidence instead: verified account ownership, a matching approved test, or a new independent observation. Use the same terms consistently across revisions.
What Does a Reviewable Assessment Record Look Like?
A reviewable record lets another analyst reproduce the reasoning without guessing which source mattered. Put the decision question, time window, source ledger, leading explanation, alternative, judgment, confidence basis, unknowns, action owner, and reopening trigger together. Preserve source links internally according to handling rules; the published assessment can summarize them without exposing sensitive records.
Here is the fictional assessment in compact form. The judgment is “Targeted monitoring is warranted while the service-account change is verified.” The basis is repeated failures in the authentication log, alongside a nearby integration change. The alternative is an approved test, and the key unknown is the owner’s explanation.
- Decision questionShould the SOC lead increase targeted monitoring now?
- Source ledgerLog, change record and help desk note retain their limits
- Alternative testAsk whether an approved integration test explains the pattern
- Bounded judgmentMonitor while account ownership is verified
- Review triggerReassess after the owner reply or a new account change
The flow shows a decision path, not a real case or an automated verdict. A reviewer should be able to challenge the causal assumption, find the contrary evidence, and see why monitoring was selected. Record who approved the action and when it expires. A monitoring decision without an expiry can quietly outlive the evidence that justified it.
MITRE’s ATT&CK CTI training connects analysis of mapped behavior to defensive recommendations. Map a behavior only when the underlying evidence supports it. This fictional log pattern does not justify an ATT&CK technique label or actor attribution.
When Should the Assessment Be Reopened or Withdrawn?
Set a trigger before sharing the assessment. Reopen this case when the account owner supplies the approved change record, when new authentication evidence appears, or when the monitoring period ends. A reviewer then compares the new facts with the original assumption and records what changed.
If a verified approved test explains the pattern, withdraw the implication that the activity needs continued threat-focused monitoring. Preserve the earlier version and the correction so readers understand the decision history. If new independent evidence indicates unauthorized changes, revise the judgment and follow the incident route. Do not silently replace the original assessment.
NIST SP 800-150 recommends mechanisms for updating or retracting shared indicators and associated context when they are wrong. ODNI’s ICD 203 asks analytic products to explain meaningful changes from earlier judgments. Those ideas support a visible revision trail, although this workflow is an editorial application for CTI teams.
End the record with the next check, its owner, and its deadline. A SOC lead can then act on the current assessment while knowing exactly which finding would reverse it. The threat intelligence executive briefing guide covers the later task of presenting such a decision to leadership.
Frequently Asked Questions
What is the output of threat intelligence analysis?
A bounded assessment that connects evidence, alternatives and uncertainty to a decision. It should identify the owner and the condition that would change the judgment.
Can several matching reports raise analytic confidence?
Only if their original observations are independent and relevant. Reports that repeat one source do not provide separate corroboration.
When should an assessment be withdrawn?
Withdraw or revise it when verified new evidence defeats a key assumption. Keep a visible revision trail and tell the decision owner what changed.
