The First Hour: Ransomware Response for Small Teams
It is 6:50 on a Monday morning. The first person into the office calls you because the shared drive is "full of weird files". There is also a text document on the desktop that she did not put there. Your stomach drops. You have never done this before. There are two of you on the IT team. Everyone who will want answers is going to arrive in the next forty minutes. What you do in the next hour will decide whether this is a very bad week or a very bad quarter.
This guide is the checklist for that hour. It is written for small teams — one to five people, no security operations center, no incident commander, possibly no incident-response retainer. It is written in order, because order is the thing panic destroys first. Each step says what to do, why, who owns it, and how long it should take. You will also find the list of things not to do. In practice, that matters as much as the list of things to do. There is also a short note on what the recovery phase will need from the evidence and notes you gather now. Print it. Put it in the drawer. The day you need it is not the day to be reading it for the first time.
In this guide
- The ransom note is a late signal
- Before you start: one decider, one timeline
- Step 1: Isolate
- Step 2: Protect the backups
- Step 3: Preserve evidence
- Step 4: Communicate
- Step 5: Engage help
- Step 6: Scope
- Step 7: The don't list
- The first-hour checklist table
- What recovery will need from you
- Where to go next
The ransom note is a late signal
Start with an uncomfortable truth, because it shapes everything that follows. If you learned about the incident from a ransom note, the attack has been running for a while. The note appears at the end of the encryption stage, which is itself the end of a lifecycle that usually spans days. That lifecycle includes a foothold, quiet check-ins, discovery, lateral movement, and staging. Each of those stages left traces on your network, and the rest of this pillar is about catching them earlier. Right now, though, you are at the end of the story, and that means two things.
First, the attacker has likely reached more than the one machine that showed you the note. Assume the scope is bigger than it looks until you have proven otherwise. Second, encryption may still be in progress on machines that have not displayed a note yet. Every minute counts in a literal way. A file server encrypting at full speed loses shares by the minute. The difference between isolating it now and isolating it after a meeting can be the difference between restoring one share and restoring all of them. That is why the first step is isolation, and why it comes before understanding.
Before you start: one decider, one timeline
Two things take thirty seconds and pay for themselves many times over. First, decide who is deciding. In a small team that is usually whoever is most senior and present. It does not matter much who it is as long as it is one person. A response with two people making independent isolation decisions is worse than a response with one making imperfect ones. Second, open a timeline note — a text file, a notebook, a message thread, anything with timestamps — and write down what you know right now. You will add to it all day, and it becomes the backbone of everything later. That includes the report to leadership, the briefing to the incident-response provider, the insurance claim, and the retrospective. A starting entry looks like this:
06:52 Call from reception: shared drive "full of weird files", note on desktop.
06:55 Confirmed: \\FS01\Shared shows renamed files, new extension, note file.
06:56 Decider: me. Timeline: this file. Second person: paged.
06:58 FS01 switch port shut (core-sw1 port 14). Reception PC port shut (sw3 port 7).
07:01 Backup server: disconnected from network. Offsite copy: confirmed reachable
only from vault console, last good run yesterday 23:10.
Write the time before you do the thing, not after. Precision to the minute is enough. You will not remember the order of events by lunchtime, and the order is exactly what everyone will ask about.
Step 1: Isolate
Do: disconnect the machine or machines that are actively encrypting from the network, right now, before anything else. The fastest reversible ways are to shut down the switch port the machine is plugged into, or to move that port to a quarantine VLAN with no route to anything. If you can do it from your switch management console, do that. If you cannot, physically unplug the cable. Then look at what is on the same network segment. Encryption may be visible on more than one machine in a segment. Or you may have lateral-movement evidence pointing at other hosts. In either case, isolate the segment at its uplink rather than chasing machines one at a time.
Why network isolation and not power: powering off feels decisive, and it is the wrong reflex. A running machine holds evidence in memory — what was executing, what connections were open, sometimes material an investigator can use to understand the attack. That evidence vanishes at power-off. Pulling the network cable stops the spread just as completely, because ransomware cannot reach other machines without a network. Disconnecting the network keeps the evidence intact. Power off only if you have no way to disconnect the network and the machine is actively destroying data you cannot afford to lose more of. Treat it as the last resort it is.
Which machines: isolate the ones showing notes, and the ones the file-server audit log names as the source of the rename storm. Also isolate the ones your flow data shows moving heavy traffic to shares, and any that your lateral-movement alerts flagged in the past few days. When in doubt, isolate; a user without a network connection for an hour is a trivial cost against one more encrypted server. Write each port and time in the timeline.
Step 2: Protect the backups
Do: immediately verify that a copy of your backups exists that the attacker cannot reach. That means an offline copy, an immutable copy that cannot be altered or deleted for a set period, or a copy at a location that the compromised network has no credentials for. If your backup server is on the same network and reachable with domain credentials, disconnect it from the network now, before it becomes the next target. A backup server that is still online during an active incident is a backup server the attacker may be working on. If you have cloud or offsite backups, confirm from an independent device that they are intact and that the retention has not been shortened.
Why this is step two and not step five: your backups are the only thing that makes recovery possible without paying, and the attacker knows it. Backup tampering — deleted snapshots, shortened retention, cancelled jobs — is a routine part of the encryption stage. If it has not happened yet, it may be happening while you are reading this. Securing the backups is the single action that most changes the outcome of the incident. It comes right after stopping the spread.
Do not restore yet. That is worth saying twice. Restoring onto a network that still contains the attacker means restoring into an environment where the restored data can be encrypted again. Restoration belongs to the recovery phase, after scoping and after the environment is clean. Right now, the job is to make sure the backups survive.
Step 3: Preserve evidence
Do: stop the automatic loss of logs. Firewall logs, DNS logs, domain controller logon logs, file-server audit logs, flow records, and your monitoring system's history all roll over on a schedule. On a busy day some of them keep only hours. Extend retention or export what you have to a location off the affected network. Take a screenshot or photograph of the ransom note and of any affected screens. Note the exact text of the note, including any identifiers it contains, but do not act on anything it says. You may have the ability to take a disk image or a memory snapshot of an isolated machine. If you have that ability and someone who knows how, do it. Otherwise, leave the machine isolated and running for the people who do.
Why: everyone who helps you from here will ask two questions. That includes the incident-response provider, the insurer's investigators, law enforcement, and your own future self writing the retrospective. How did they get in, and what did they reach? The answers live in logs that are quietly being overwritten. Twenty minutes spent freezing them now is the difference between knowing and guessing, and guessing wrong means rebuilding an environment with the same hole in it.
Step 4: Communicate
Do: tell leadership first, in plain words, within the first thirty minutes. Tell them what has happened, what you have done, what you do not yet know, and when you will update them next. Then tell users, briefly and honestly, that there is a security incident and some systems are unavailable. Tell them not to try to work around it by copying files elsewhere. Tell them not to turn machines off or on unless told to. Tell them where updates will appear. Set a rhythm for updates — every hour, say — and keep to it even when the update is "no change." The general shape of that first announcement is the same as for any outage. So is why a promised next-update time matters more than the details. Both are covered in The First Fifteen Minutes of an Incident.
Why leadership first: because the decisions from here on are not yours to make alone. Leadership, with legal advice, decides whether to engage an outside provider, whether to notify the insurer, and whether customers or regulators must be told. Every decision related to the ransom demand belongs to them too. Your job is to give them accurate facts fast enough to make those decisions well. Tell them what you know, flag what you are unsure of, and resist the temptation to reassure. "We have contained it" is a claim you will regret if a second machine lights up an hour later. "We have isolated the machines we know about and are checking for others" is both true and useful.
Step 5: Engage help
Do: make the calls. If your organization has an incident-response retainer or a managed service provider with a security capability, call them now, not after you have "figured out what happened." If you have cyber insurance, the policy almost certainly requires notification within a short window and may require using the insurer's approved responders. In that case, find the policy, find the number, and let leadership make the call. Legal counsel should be looped in early, because notification obligations and the handling of evidence both have legal dimensions. And law enforcement, as a concept, is a party that should be informed of a ransomware incident. Leadership and counsel decide when and how.
Why now: because outside responders are most useful at the beginning, when the evidence is fresh and the decisions are still open. It is also because every one of these parties is going to ask what time you contacted them. The instinct to handle it in-house first, to avoid looking incompetent, costs hours you do not have. You are not expected to be a forensics team. You are expected to stop the spread, protect the backups, keep the evidence, and get the right people on the phone. That is the whole job of the first hour, and doing it well is the opposite of incompetent.
Step 6: Scope
Do: with the spread stopped and the calls made, start answering the three scoping questions and writing the answers in the timeline. Which hosts? List every machine showing a note and every machine named as a source in file-server audit logs or flow data. Also list every machine your lateral-movement or fan-out alerts flagged in the past week. Which shares? Walk the file servers and note which shares contain encrypted files and which do not. The boundary tells you how far encryption got before isolation, and the untouched shares are the ones you protect hardest. Which accounts? From the audit logs and domain controller logon events, identify the account or accounts that performed the encryption. Also identify any that showed the one-account-many-machines pattern. All those accounts get disabled and their passwords reset. Every administrative credential they could have exposed is treated as compromised.
Why after the calls: scoping is the first step that is analysis rather than action, and it is the step the outside responders will lead if you have them. Doing the first pass yourself gives them a running start and gives leadership the "how big is it?" answer they are waiting for. Do not let scoping expand into investigation. "How did they get in?" is a fair question and the wrong one for this hour. It will be answered from the evidence you preserved, later, by people with time.
Step 7: The don't list
The things not to do are as important as the things to do, because each of them is a natural, well-intentioned instinct that makes the incident worse.
- Don't reboot everything. Rebooting destroys memory evidence, can trigger encryption routines that were waiting for a restart, and tells you nothing. Isolate instead.
- Don't wipe or reimage the affected machines yet. They are evidence. Reimaging before an investigator has looked at them erases the answer to "how did they get in."
- Don't restore backups into the compromised environment. Restoration comes after the environment is confirmed clean. Restoring early risks the restored data.
- Don't communicate with the attacker, negotiate, or make any payment decision yourself. Those decisions involve leadership, legal counsel, the insurer, and possibly law enforcement. They involve considerations — legal, financial, reputational — that are not an IT team's to weigh. Your contribution is accurate facts about what was encrypted and what the backups can restore.
- Don't reconnect an isolated machine "just to check." Every reconnection is a chance for the spread to resume.
- Don't reuse the compromised accounts, including for your own response work. Use a fresh administrative account created from a known-clean machine.
- Don't promise timelines. "Back by Wednesday" is a guess that will be quoted back to you. "Next update at 9:00" is a promise you can keep.
Tip: print the checklist table below and keep a paper copy somewhere that is not on the file server. A response plan stored only on the share that just got encrypted is a plan you do not have.
The first-hour checklist table
Owners assume a two- or three-person team; adjust to yours, but keep one decider. Time targets are cumulative from the moment you learn of the incident.
| # | Step | Owner | Target | Done when |
|---|---|---|---|---|
| 0 | Name the decider; open the timeline note | Senior person present | 1 min | First timestamped entry exists |
| 1 | Isolate encrypting hosts; then the segment if spreading | Network admin | 5 min | Ports shut or quarantined and logged; no new encryption observed |
| 2 | Protect backups: verify an unreachable copy; disconnect backup server if exposed | Sysadmin | 10 min | Offline or immutable copy confirmed from an independent device |
| 3 | Preserve evidence: extend log retention, export logs off-network, photograph the note | Sysadmin | 20 min | Logs and screenshots stored off the affected network |
| 4 | Communicate: leadership, then users; set the update rhythm | Decider | 30 min | Leadership briefed; user notice sent; next update time promised |
| 5 | Engage help: IR retainer or provider, insurer, legal, law enforcement as leadership directs | Decider with leadership | 40 min | Calls made and logged with times |
| 6 | Scope: hosts, shares, accounts; disable and reset compromised accounts | Both admins | 60 min | First-pass lists written in the timeline |
| 7 | Review the don't list; confirm nothing on it has happened | Everyone | Continuous | No reboots, wipes, restores, reconnections, or attacker contact |
If you get through the table in the hour, you have done the job. If step one takes you twenty minutes because you had to find the right switch, that is fine. That delay is also the first item for the readiness work in the last guide of this pillar, because next time it should take two.
What recovery will need from you
Recovery is a different phase with a different tempo, usually led by whoever you engaged in step five, and it can take days to weeks. It is not the subject of this guide, but the first hour sets it up. And it helps to know what it will ask of you. Recovery needs the timeline, because the order of events determines which backups are trusted and which are suspect. It needs the scope lists — hosts, shares, accounts — because recovery is done host by host and share by share. And a machine missing from the list is a machine that gets rebuilt last or not at all. It needs the preserved evidence, because the environment cannot be declared clean until the entry point is known and closed. Rebuilding without that knowledge is rebuilding with the same hole. It needs a known-good backup point, which is why protecting the backups came second. It also needs a clean, isolated network to restore into, which is why you did not restore in the first hour. And it needs you rested enough to do careful work, which is the least technical item on the list and the one small teams most often get wrong. Hand off, sleep, come back.
Where to go next
The first hour is isolate, protect the backups, preserve, communicate, engage, scope, and refrain. Everything else in this pillar is about not needing this guide. The Ransomware Lifecycle from the Network's Point of View explains why the note is a late signal and where the earlier ones are. Lateral-Movement Traffic is the detection that would have caught it days earlier. Mass Encryption Signatures shows the instruments that would have fired in the first minute. Readiness is the program that puts it all in place, including the drill that makes this checklist take an hour instead of a day. The whole set is at Early Warning Signs of Ransomware on the Wire.
Put this into practice with SysPeep Monitor
Step three of this checklist assumes your monitoring history survives the incident, and step six leans on it to scope the damage. SysPeep Monitor keeps device, interface, and server metrics in a central database — SQLite, MySQL/MariaDB, or SQL Server. Its dashboards and reports show what a file server or a switch port was doing at any hour. Optional clustering provides high availability so the monitor itself stays up on the worst day. Free trial on any Windows machine.
Download Free TrialLearn About SysPeep