Feature Incident Response
The Memo You Sign Before the 36-Hour Clock Starts
The 36-hour clock runs from your determination, not detection. Here is the notification incident determination memo, who may sign it, and how to record a holding answer.
Table of Contents
TL;DR
- The 36-hour clock under 12 CFR 53.3 runs from your determination, not from detection, containment, or the first Slack message. The determination is the event being regulated.
- The agencies replaced the proposed “good faith belief” standard with a determination standard because belief was “too subjective and imprecise.” A determination needs a signer, a threshold test, and a timestamp.
- The memo is short by design. It records what you knew, which prong you tested, who decided, and what would reopen the question. It is not an incident report and it is not an assessment.
At 02:14 the on-call engineer confirms the core platform is encrypted and the core vendor’s status page still says “investigating.” Somebody on the bridge asks the only question that actually matters: does this get called in?
Six people have an opinion. Two of them are loud. Nobody knows whether they are allowed to decide, and the clock that everyone is arguing about has not started yet, because the rule does not start it at detection.
What the 36-hour rule actually times
The notification duty sits at 12 CFR 53.3 for OCC-supervised institutions, with parallel provisions at 12 CFR 225.302 for the Board and 12 CFR 304.23 for the FDIC. In each, the deadline is “as soon as possible and no later than 36 hours after the banking organization determines that a notification incident has occurred.”
Two words carry the weight: determines and notification incident.
A notification incident is not any cyber event. Under 12 CFR 53.2(b)(7) it is a computer-security incident that has materially disrupted or degraded, or is reasonably likely to materially disrupt or degrade, one of three things: the bank’s ability to carry out operations or deliver products and services to a material portion of its customer base; a business line, where failure would cause material loss of revenue, profit, or franchise value; or operations whose failure would threaten U.S. financial stability.
The final rule preamble is blunt about what happens after the determination: 36 hours is reasonable “as it does not require an assessment or analysis.” The agencies also wrote that they do not expect a bank to determine that a notification incident occurred the moment it learns of a computer-security incident, and that a reasonable amount of time to determine is anticipated.
That is the entire design. Assessment time lives before the determination. Once the determination lands, you have 36 hours and no analysis obligation. Which means the determination itself is the control, and an undocumented determination is an unprovable one.
The agencies removed the original “good faith belief” trigger on purpose, agreeing with commenters who called belief “too subjective and imprecise.” Belief is what gets reconstructed from memory during an exit meeting. A determination is what gets signed.
The memo, field by field
The rule prescribes no document. Notification goes to the supervisory office or OCC-designated point of contact by email, telephone, or another method the agency allows. So the memo is not the filing. It is the internal record that survives the filing — the thing your counsel reads before the post-incident review and the thing an examiner reads before asking why a notification took nine hours.
Keep it to one page. The decision-log structure covers the broader logging discipline; this is the single artifact for the determination itself.
| Field | What goes in it | Failure mode it prevents |
|---|---|---|
| Incident ID and detection time | Existing ticket number, first signal, monitoring source | A memo that floats free of the incident record |
| Facts at decision time | Only verified facts, timestamped, each tied to a system or vendor observation | Facts from hour nine retro-fitted into an hour-two decision |
| Prong tested | Which of the three 53.2(b)(7) prongs, and the specific service, business line, or customer segment | A generic “materiality” conclusion nobody can map to the rule |
| Threshold reasoning | Two to five sentences applying the facts to the prong, including the “reasonably likely” limb if used | A conclusion with no bridge to the definition |
| Determination | Notify, or not a notification incident, or not determinable yet | Ambiguity about whether the clock is running |
| Signer, role, timestamp | Named individual, authority basis, minute-level time | A decision owned by a bridge call |
| Reassessment trigger and time | The specific fact that would change the answer, plus the next review time | A negative determination that is never revisited |
| Distribution | Who received it, including bank partner and outside counsel where applicable | A record nobody downstream can find |
The facts field has a floor. If you cannot state the affected service, the customer or business-line exposure, the duration or likely duration, and whether workarounds exist, you do not have enough to determine and you should say so rather than guess. Four facts, minimum, before anyone signs anything.
Who is allowed to sign
Most incident response plans name an incident commander and stop there. The commander runs the response; that is a different job from deciding the regulatory threshold, and conflating the two produces a decision made by whoever is loudest at 2 a.m.
A workable authority design, and one that survives an examiner asking “show me the delegation”:
| Role | Signs? | Backup | Quorum note |
|---|---|---|---|
| CISO or security lead | Facts only | Deputy security lead | Always present for the fact record |
| Second-line risk or compliance officer | Yes, threshold determination | Named compliance officer on the on-call rota | Cannot be the incident commander |
| General counsel or delegate | Yes, legal exposure and privilege | Outside counsel during business hours | Required when customer data or litigation is likely |
| CFO or Treasurer | No | — | Consulted when the prong is business line or franchise value |
| CEO | Escalation only | — | Not a default signer; too slow for a 36-hour window |
One of the three decision roles is the designated signer, named by title in the plan, with a published escalation order. The plan should also set a decision deadline inside the response itself — a hard internal checkpoint, commonly two hours from severity escalation — so the determination does not wait on perfect information.
Two authority gaps show up repeatedly. The first is a plan that requires “CRO approval” without naming a backup, which stalls every incident that starts on a weekend. The second is a plan with no rule for disagreement, which means the most senior person in the room decides and the dissent disappears. Record the dissent. A memo saying “Compliance recommended notification; GC assessed no notification incident because X” is stronger evidence of a reasoned process than a unanimous one-line conclusion.
What a completed memo looks like
Labeled hypothetical, built on the rule’s own illustrative examples:
Determination: NOTIFY. 02:41. Prong 1 tested. Core deposit platform reachable only in read-only mode since 01:58; card authorizations failing for two of three BIN ranges; mobile and online banking returning errors to approximately 60% of sessions. Core vendor confirms ransomware on a shared application tier, recovery time undetermined. Facts verified by CISO from the vendor bridge (02:22) and our own synthetic transaction monitor (02:31). Threshold reasoning: inability to post transactions for a material portion of the retail customer base meets prong 1; “reasonably likely to materially disrupt” is independently satisfied given the undetermined recovery time. Signed: J. Okafor, VP Compliance (delegated authority, IRP §4.2). Reassess at 04:00 or on any change in recovery estimate. Distribution: CEO, CFO, GC, bank partner ops contact, outside counsel.
Note what the memo does not do. It does not diagnose root cause, estimate loss, or assign remediation. Those arrive later. The memo answers one question in the present tense and creates the timestamp that starts the clock.
When the answer is “not yet”
The most common real determination at hour two is neither yes nor no. It is not determinable.
That is a legitimate state, and the preamble supports taking time to reach the answer. But “not determinable” has to be managed as a holding entry with an owner, the specific missing fact, and a review time — usually 60 to 120 minutes out, tightening as the incident matures. A holding entry with no review time is how a Friday night incident becomes a Monday notification.
Two preamble passages should shape how you treat the in-between. The agencies wrote that if a bank is in doubt about whether it is experiencing a notification incident, they encourage it to contact the regulator. They also wrote that they recognize a bank may notify on a mistaken determination, and that they “generally do not expect to take supervisory action in such situations.”
So the cost of a false positive is a phone call and a memo that says so. The cost of a false negative is a reconstructed timeline and a question about why the threshold test took as long as it did.
One incident, several clocks, one memo
The OCC memo will not satisfy NYDFS. Under 23 NYCRR 500.17(a)(1), a covered entity must notify the superintendent “as promptly as possible but in no event later than 72 hours from a determination that a cybersecurity incident has occurred” — a different defined term with a different scope, including incidents with no OCC prong implication. The SEC’s Item 1.05 analysis runs on materiality for a different audience again.
Use the OCC memo as the shared fact record, then run each regime’s own test in its own short addendum. The multi-regulator sequencing problem is where institutions most often lose time, because a single determination gets treated as covering every obligation at once.
File it, retain it, reopen it
Name the memo to the incident ID, not to the topic: IR-2026-041_determination_0241.pdf. Store it in the same system of record as the incident, not in a shared drive named after the month. Freeze it — corrections go in a dated addendum, so a later reader can see what changed and when.
Retention should match your incident records schedule; seven years is a common starting point for examination-sensitive artifacts, and it is a starting point, not a rule requirement. Confirm against your own schedule and counsel’s view.
Then close the loop. At post-incident review, the question is not “did we notify correctly.” It is “how long did the determination take, and which missing fact caused the delay.” OCC Bulletin 2025-24 moved community bank supervision toward documented, defensible risk decisions rather than checklist conformity, which raises the value of a memo that shows a reasoned two-hour decision and lowers the value of a retrospective narrative.
So What?
Three things, this week, in this order. Pull up the incident response plan and find the sentence that says who determines the threshold — if the answer is a role with no named alternate, write the alternate in. Build the one-page template with the eight fields above and drop it in the incident folder so nobody is drafting it during an incident. Then run one tabletop where the injected scenario is deliberately ambiguous at hour two, and grade the team on the memo, not on the notification.
The clock does not start when everyone agrees. It starts when someone signs.
If you want the clock, notice matrix, and severity scoring already built, the Incident Response & Breach Notification Kit has the workbook and plan templates.
FAQ
Does the memo need to be privileged? Mark it “prepared at the direction of counsel” only if counsel is actually directing the process, and route it through counsel. Over-marking is common and mostly harmless; under-marking a determination memo that discusses legal exposure is not.
What if our primary federal regulator is the FDIC or the Fed? Same structure, different addressee. The FDIC’s parallel provision is 12 CFR 304.23 and the Board’s is 12 CFR 225.302, both with the same determination trigger and 36-hour outer limit. Update the contact block, keep the memo.
Do bank service providers need their own memo? The provider’s duty under §53.4 is to notify at least one designated point of contact at each affected customer bank as soon as possible once it determines a covered-service disruption of four hours or more. A provider memo recording that determination and each notice sent is the provider-side equivalent. Ask your critical vendors for it.
Is a memo needed for every incident, including small ones? No. Use a severity gate — anything escalated to the incident commander gets one, which in practice means a dozen or so a year at a mid-size institution. Everything below the gate stays in the ticket.
◆ Need the working template?
Start with the source guide.
These answer-first guides summarize the required fields, evidence, and implementation steps behind the templates practitioners search for.
◆ Related template
Incident Response & Breach Notification Kit
Run an incident and every notice clock it starts: bank partner, the bank regulators' 36-hour rule, NYDFS, FTC, SEC and breach laws in 54 states and territories. Workbook, guide, four playbooks, tabletop kit and Word plan templates.
◆ Immaterial Findings · Weekly
Sharp risk & compliance insights. No fluff.
◆ FAQ
Frequently asked questions.
What is a notification incident determination memo?
Who should be authorized to sign a notification incident determination?
Can we avoid a memo if the incident is obviously not a notification incident?
How fast does the determination have to be made?
Author
Rebecca Leung
Rebecca Leung has 8+ years of risk and compliance experience across first and second line roles at commercial banks, asset managers, and fintechs. Former management consultant advising financial institutions on risk strategy. Founder of RiskTemplates.
◆ Related framework
Incident Response & Breach Notification Kit
Run an incident and every notice clock it starts: bank partner, the bank regulators' 36-hour rule, NYDFS, FTC, SEC and breach laws in 54 states and territories. Workbook, guide, four playbooks, tabletop kit and Word plan templates.
◆ Keep reading
Related posts.
Incident Response
NCUA's 2026 Cybersecurity Report: Most Credit Union Incidents Involve Vendors the Regulator Can't Examine
The NCUA's 2026 Cybersecurity and Credit Union System Resilience Report documents 588 reported incidents, including third-party provider incidents at vendors NCUA has no authority to directly examine. NCUA's earlier data put nearly 7 in 10 incidents at a vendor, and one 2023 ransomware attack on a service provider knocked out 60 credit unions. Here's what this structural gap means for your incident response program.
Oct 7, 2026
Incident Response
Your Incident Response Plan Has the Wrong Trigger. Here's What Starts the SEC's 4-Business-Day Clock.
Two years of SEC Item 1.05 cybersecurity disclosures reveal one consistent mistake: IR plans built around incident discovery, not materiality determination. Here's what the trigger actually is — and how to fix your escalation process before it matters.
Oct 6, 2026
Incident Response
Most Incident Response Plans Skip This Step. It's the One Regulators Will Ask About Six Months Later.
Financial institution incident response plans focus on notification timelines. Forensic evidence preservation — the step that determines what you can tell regulators and courts — is usually an afterthought. Here is what needs to happen in the first four hours of a cyber incident, and what it costs when it doesn't.
Oct 3, 2026