Staffing a round-the-clock monitoring desk looks simple on a whiteboard and turns difficult by the second month. This walkthrough sets out ten steps for building a SOC Ops rota in a hospital IT department, from counting the work to protecting the people who carry it. It stays in the administrative and workforce domain throughout; no clinical material is involved, and the example numbers are round, illustrative figures rather than data from any real facility.
Step 1: Count SOC Ops arrivals before you count the staff
The Ops Atlas treats the desk as a waiting line. The delay an alert experiences is driven by three inputs: the arrival rate, the minutes each case consumes, and the headcount present. Begin by exporting a month of alerts and recording arrivals per hour of the day. Most departments discover a lumpy pattern, with a daytime hump and a quiet but unwatched night.
Keep the unit of analysis simple. A spreadsheet with hour, count and average handling minutes is enough to support the next steps.
Step 2: Convert workload into seat-hours
Multiply arrivals by handling time to obtain the working hours the desk must supply each week. Then add headroom, because a queue running at full utilisation grows without limit whenever a busy hour arrives. The Atlas walks through a worked month in which several hundred weekly seat-hours translate into just under a dozen full-time triage staff once leave, training and breaks are included.
Another useful rule of thumb comes from the MITRE material the Atlas summarises: a single seat covered continuously, every hour of the year, needs roughly five full-time people. If management wants “one person watching at all times”, the budget conversation must begin with that arithmetic.
Step 3: Respect the two-person floor
The same sources suggest that no shift ever runs with a lone analyst, so two is the minimum. One person cannot take a break, cannot sanity-check an unusual case, and cannot absorb a sudden spike. For a small hospital team, this floor is often the deciding factor between building a desk in-house and contracting a managed provider.
Before moving on, add a filter step. Analysts should never receive an unfiltered feed. Enrichment, such as attaching the owner of a device or the usual hours of a user account, should happen before a human opens the case. Every minute spent looking up basic context is a minute of scarce rota capacity, and the arithmetic in Step 2 assumes handling times that already include such help.
Step 4: Define roles as a career ladder
The Atlas describes three tiers: first-line triage, deeper investigation, and specialists. It also insists the entry rung be presented as a place to grow, never as a lower caste. This matters for retention. When triage is treated as a punishment, experienced people leave and the desk loses the context that makes decisions quick.
- Publish what each tier is expected to decide and what must be escalated.
- Rotate people through investigation work so that triage does not become a dead end.
- Tie progression to demonstrated judgement, not to ticket counts.
Tiering is optional, too. The sources caution against rigid per-alert time quotas, since a stopwatch rule pushes analysts to close cases early.
Step 5: Write the handover ritual
Rotas fail at the seams. Adopt a pass-down log and make the rule explicit: the outgoing analyst lists open items, the incoming lead reads and signs, and anything not accepted remains with the outgoing owner. Ten minutes of overlap is cheap compared with a lost case. If you have wondered how other operational teams keep equipment and people aligned over time, our guide to routine monitoring and upkeep for robotic equipment shows the same logic of scheduled checks and documented ownership in a different setting.
Step 6: Give every alert a recorded outcome
Agree on a short list of closing states and require a reason with each. The Atlas lists four: real threat, benign but genuine, mistaken detection and undetermined. A reason field converts closures into raw material for improvement. After a month, benign cases often reveal where the organisation should share context with analysts, such as a list of expected maintenance windows, rather than where detection needs fixing.
Step 7: Measure the desk without measuring the individual
Publish the typical wait plus the ninetieth-percentile wait until someone first acts, split by severity and by shift. The Atlas warns that individual scorecards invite gaming, and that a target based purely on closing time increases reopened cases. Team-level measures, reviewed weekly, give honest signals while preserving trust.
Step 8: Look after the people
A rota is also a promise to the humans on it, and promises are tested in the third week of winter nights.
Burnout is listed in the Atlas among the central concerns of the discipline, not a footnote. Practical protections include capping consecutive night shifts, keeping the alert feed filtered so people never face raw noise, and reserving time for tuning work that makes next month easier. A desk that spends all its capacity on repetitive closures never improves. Ask each analyst once a month which three recurring cases annoy them most, and give someone the time to remove the cause.
For readers who want a broader map of where this work sits among related operating disciplines, the longer overview in our overview of how modern operations disciplines and learning paths fit together is a useful companion. The definition used throughout this walkthrough comes from the Atlas’s description of the SOC Ops function, which also lists reference frameworks for self-assessment.
Step 9: Decide what to build and what to buy
Many hospital departments cannot staff a full rota themselves. The Atlas covers both managed providers and in-house models, and its message is consistent: contracting out the labour does not contract out accountability. If you do buy, insist on the same artefacts you would build, namely recorded outcomes, median and slow-tail times, and a signed handover. Write the escalation route into the agreement, including who may declare an incident and the criteria on paper.
Step 10: Review yearly and adjust
Finally, schedule a maturity review. Put the date in the calendar now, before the rota becomes everyone’s excuse to postpone it. The Atlas points to an openly published self-assessment model spanning five domains from business through to services, intended to be repeated annually. Treat the first result as a baseline, not a grade.
The steps above are general workforce planning guidance for IT operations, not a recommendation for any particular hospital, vendor or staffing level. Adapt every figure to local data.
