Imagine the reception desk of a busy organisation on a Monday morning. The check-in screen freezes, the phone starts ringing, someone suggests restarting the computer, someone else says they already did, and a queue forms. Within ten minutes there are five people giving five instructions and nobody has told the waiting visitors what is going on. Nothing about this scene is medical. It is an operations problem, and the way technology teams deal with it has a name that is worth knowing: IncidentOps. This article explains the idea in plain language for administrators, office managers and anyone who keeps an organisation’s everyday systems running.
To keep the explanation honest, we lean on the Ops Atlas, a catalogue written to teach operations subjects. According to the Atlas page on IncidentOps, the term is an informal, seldom-used way of describing incident management, so there is no exam, no licence and no single rulebook. What you get is a toolbox of habits. We will go through it piece by piece.
What does IncidentOps actually cover?
In one sentence: how a team responds when a service that people rely on breaks unexpectedly. It starts the moment someone decides “this is an incident”, and it ends after the team has met to learn from it. The in-between pieces are deciding how serious it is, picking who is in charge, passing messages to those affected, fixing or working around the problem, and writing down what to change. Planned maintenance does not count, and neither does every small alert that fixes itself.
It also helps to know what lies outside. The Atlas separates out attacks by hostile people, which need specialised security handling; the work of removing deep causes over weeks, which service-management people call problem management; and the restoration of systems after a disaster. IncidentOps is the first-response layer in between.
Piece 1: Deciding early
The single most skipped step is declaring. Staff hesitate because it feels dramatic. A simple rule fixes that: write down, in advance, when an incident is called. For example, when two departments are involved, when visitors can plainly see something is wrong, or when an hour has passed and nobody has a solution. If it turns out minor, you lose a few minutes. If you wait and it turns out major, you lose a morning.
Piece 2: How serious is it?
A severity scale gives everyone the same shorthand. Level one might mean the whole check-in process is stopped; level three might mean one workstation is slow. The scale decides who gets called, how often messages go out and how soon the review happens. Keep it to three or four levels, and give each a single sentence.
Piece 3: Who does what
The Atlas describes roles as hats, not job titles. A coordinator keeps the picture and makes decisions but does not do the repair. A small hands-on group does the fixing, and nobody outside it should be changing the system. A communicator writes updates for everyone else. A note-taker records what happened and when. Extra roles such as a deputy or liaisons are added only for bigger events.
In the reception example, the office manager coordinates, the IT contact fixes, a front-desk lead tells the queue what is happening, and an assistant jots down times. Notice that the manager never touches the computer. That separation prevents the classic mess where two people undo each other’s work.
Piece 4: Saying something, on a schedule
The Atlas describes a rhythm borrowed from technology operations: an opening message around the five-minute mark, a second one soon after to explain the scope, and then notes at gaps of twenty minutes or fewer for the first stretch. A closing message should state whether anything was lost. Status tools generally label the stage with four words, Investigating, Identified, Monitoring and Resolved, which gives readers an at-a-glance sense of progress.
At a desk, this means a sign or a spoken announcement every twenty minutes, even if it is “we are still working on it”. People tolerate waiting far better than they tolerate silence.
Piece 5: Make it better first, explain later
When the check-in system freezes, the quickest relief may be to switch to paper forms. You do not need to know why the system froze. Technology teams call this mitigation, and their common moves include going back to the previous version, turning off a recent addition, switching to a spare, or limiting how much work the system is asked to do. The Atlas stresses that these moves depend on knowing where the fault is, which is a smaller question than knowing why it happened.
For operations that rely on connected machines, our overview of tools for running a fleet of robots day to day shows the same philosophy at work: keep things moving first, diagnose second.
Piece 6: Learning without blame
After the event, the team holds a review. The Atlas offers some practical rules. Decide who owns the review before the incident is over. Have someone not directly involved run it. Look for conditions that made the slip easy rather than ending at “someone made an error”; the Atlas quotes a safety-science view that hindsight makes past choices look more careless than they were. End with a short set of actions, every one tagged with a single owner and a deadline, and check later that they were completed. Long lists, vague promises and quietly downgraded items are the usual ways reviews fail.
Why non-technical people should care
It is easy to think this belongs to IT alone. In practice, the quality of an incident response depends on people far from the server room. The person at the desk who spots the first sign, the manager who is willing to say “call it an incident”, the colleague who writes a calm message, and the leadership team that resists the urge to interrupt all shape the outcome. The Atlas notes that when an executive asks for updates mid-incident, a polite and effective reply is a question: do you wish to take command? It reminds everyone that being in charge is a role with duties, and not a right to hover.
A worked example, in ten lines
- 8:50 am: The check-in screen freezes. The front-desk lead declares an incident.
- 8:52: The office manager takes the coordinator hat and assigns the others.
- 8:55: First message to the queue: “Our check-in system is down. We are using paper forms. Next update at 9:15.”
- 9:00: The IT contact finds a recent update caused the freeze and rolls it back.
- 9:05: The system works. The communicator posts “Monitoring”.
- 9:15: Everything is stable. The communicator posts “Resolved” and notes that no records were lost.
- 9:20: The manager names who will run Thursday’s review.
- Thursday: The review finds that updates were applied during opening hours and picks three actions.
- Action one: schedule updates after closing time. Owner and date attached.
- Action two: keep printed forms at the desk. Owner and date attached.
Three questions beginners often ask
Do we need special software? Not at first. A shared document for role cards, a group chat and a calendar for the rota are enough. When you add a paging tool or public status page, pick ones hosted well away from the systems they are meant to report on, so they survive the very outage they announce.
Should everyone join the response? No. A crowd slows things down and tends to push the real workers into private conversations. Keep a small working group and a separate place for updates.
Is a root cause needed before we act? No, and waiting for one often prolongs the disruption. Relief first; investigation later, once things are calm.
How to start small
You do not have to adopt everything at once. The Atlas suggests a ladder, starting with simple tickets, adding one on-call schedule and a message template, and then, when incidents regularly cross departments, introducing named roles and a status page that lives outside your main systems. A few people can learn the coordinator hat, which also avoids relying on a single hero. If you enjoy reading about the community side of operations, a practical guide to technology for nonprofits and community groups covers the people side of running systems on a tight budget.
Whatever size you are, the plain-language summary is the same. Decide early, name the hats, speak on a rhythm, relieve first and explain later, and learn without blame. None of it needs special software. It needs a one-page agreement and a rehearsal on a quiet morning, long before the screen freezes for real.