The NoOps Checklist for Small Firms That Hate Downtime

A two-partner practice we will call Rao & Mehta ran its website, its document system, and its appointment calendar the way most small firms do: one partner’s nephew “who is good with computers,” a hosting bill nobody read closely, and a quiet prayer that nothing breaks during filing season. Last monsoon, something broke. The site went down for three days, clients could not download forms, and the nephew had gone off to college. In the recovery conversation that followed, someone used a word the partners had only seen in technology columns: NoOps — running your systems with no dedicated technology staff, on the strength of automation and managed services. Could a firm like theirs actually do that?

The honest answer is “partly, and only with an audit first.” That is true for any small practice, legal or otherwise. What follows is the audit: eleven questions, each answerable with documents or demonstrations rather than assurances. Firms that score honestly tend to discover something more useful than a yes-or-no verdict — they discover exactly where their systems are sturdy, where they are luck, and which duties quietly belong to a person who left in 2023. The whole exercise costs an afternoon and a pot of tea, which puts its pricing somewhere below a single hour of most professionals’ billable time.

The Question Behind the NoOps Question

First, define the term properly, because vendors will not. NoOps is an aspiration: automate and outsource the running of systems so thoroughly that a dedicated operations team becomes unnecessary. It is a direction, not a product, and the careful public explanation at the xOps School reference on NoOps is blunt about the limits — operations gets displaced to providers, software, and remaining staff, rather than abolished. With that framing in place, the eleven questions make sense.

The Checklist: Eleven Questions Before You Let Go

1. Can you list every recurring task your systems need? Updates, renewals, backups, access reviews, certificate refreshes. If producing this list takes more than an afternoon, the estate is not yet understood — and work you cannot enumerate is work you cannot delegate safely. Most firms that try this discover at least one task whose only explanation is habit.
2. Does every task on that list carry a name? A person, a vendor, or a documented automation — not “the nephew.” The share of tasks with no name is your risk number, and the only acceptable target is zero.
3. Do you know the four things that never transfer to a provider? The files you hold, the logins you allow, the settings you chose, and the permissions you granted — all four remain your responsibility on every service tier, on every plan, at every price. Name who reviews each, and how often. Weak logins and forgotten permissions cause more small-firm breaches than any cloud failure ever has.
4. When did you last restore from a backup — actually restore? Not a green tick on a dashboard; a real recovery of a real file, timed and recorded. “Managed” hosting is not a backup strategy, and the major providers’ own guidance leaves backups with you. A restore that has never been performed is a claim, and claims about recovery deserve the same scepticism you would apply to an unverified alibi.
5. Who answers when something breaks at 11 p.m. on a Saturday? Whatever the arrangement, check it is fair, known in advance, and not always the same exhausted person. Someone always holds the phone; dignity lies in choosing who, deliberately.
6. Is your uptime promise built from your whole chain? Your site depends on hosting, domain name service, payment tools, and integrations. Its real availability is the combination of all of them — never the prettiest number on any single contract. Clients who cannot reach your forms do not ask which link snapped.
7. What would leaving your main vendor cost tomorrow? Time, money, who does the work. If nobody has written this down, you do not have a supplier relationship; you have a dependency with good manners.
8. Is every piece of automation understandable by at least two people? One-person scripts are secrets, and secrets resign, marry, and move to college. Documentation is cheaper than emergencies.
9. Do you know what systems attention costs in people-hours each month? Invoices show the software; they never show the Tuesday afternoons. Owners who count the hours make better purchase decisions than owners who count only rupees.
10. If an AI tool proposes changes to live systems, who approves — and who answers for mistakes? The 2026 wave of agents that investigate and fix problems is genuinely useful, and it changes nothing about accountability: a person approves, and a person owns the tool’s permissions and errors.
11. What would one day of downtime cost your practice — in fees, not feelings? A number written on paper: missed filings, missed consultations, clients who quietly conclude the firm is careless with small things and might be careless with large ones. Every other question on this list is easier to answer sensibly once this one has a figure attached.

Scoring Yourself on the NoOps Checklist

Count the questions you answered with evidence rather than hope. Nine or more: your practice is ready to slim its operations deliberately — automate further, reassign duties with names attached, and keep the audit in the drawer for next year. Five to eight: you are on the low-ops road already, whether you use the word or not; close the specific gaps before trimming any support. Fewer than five: keep whatever help you have, commission the inventory from question one, and consider the fee the cheapest insurance you will buy this year. Re-run the exercise annually; estates drift, staff move, and last year’s evidence goes stale quietly.

Notice what the honest destination is called in every serious treatment of the idea: low-ops by design. Not an empty office with a heroic robot — a smaller set of duties, each counted, each owned, each rehearsed against failure. The firms that get closest to the NoOps ideal are, without exception, the ones that understood operations best before they let any of it go.

The Rest of Rao & Mehta’s Story

The partners spent one long afternoon on the eleven questions and scored a humbling four. But the checklist turned anxiety into a work list: a hosting migration with a written exit plan, a restore test that found a gap nobody wanted discovered during filing season, and a standing arrangement with a part-time consultant who now answers the Saturday calls so the partners never do. A year later their systems run with no technology staff at all — and with less risk than at any point in the firm’s history. The nephew, asked his opinion, reportedly said he had been doing half of this all along and was relieved someone wrote the rest down.

That is the quiet lesson hiding in the buzzword. NoOps, done properly, is not the absence of operations. It is operations with no mysteries left in it: every duty counted, every key assigned, every failure rehearsed before it is suffered. Small firms that hate downtime do not need to learn the vocabulary of the cloud. They need one honest afternoon, eleven questions, and the discipline to write the answers down — habits, as it happens, that the legal profession invented long before the software industry did.