Skip to content

Operations Guide

Applies to: All subscriptions

Purpose

Give the customer a written operating routine that catches problems before users do, and that survives the departure of the person currently doing it. Undocumented operational knowledge held by one person is a risk, not an asset.

Audience

The customer's day-to-day operations owner — often a finance systems administrator rather than an IT specialist. Also the IT team, for the infrastructure items.

Prerequisites

Steps

Routine Time needed Owner Page
Daily 10 to 15 minutes Operations Daily Checks
Weekly 30 minutes Operations and IT Weekly Checks
Monthly 1 to 2 hours Operations, IT and finance Monthly Checks
Year-end A planned exercise Finance, with IT Year-End Checks
Continuous — IT Backup and Restore · Housekeeping

The principle

Monitoring watches continuously and alerts on thresholds. These routines catch what monitoring cannot: things that are drifting rather than broken, things that need judgement, and things where the right response is a decision rather than a restart.

Monitoring catches The routine catches
A threshold crossed A trend approaching one
A component down A process nobody is following
An event that fired An alert nobody has acted on
A number out of range A number that is fine but should not be

Three things every routine checks

Whatever else changes, these three appear in every routine because their failure modes are silent and their consequences accumulate:

Check Why
Nothing is stuck Quarantined documents, queued jobs, write-back backlog. All accumulate silently
Nothing is ageing towards a deadline Unregistered documents, expiring e-Way Bills, expiring credentials and certificates
Someone acted on yesterday's alerts An alert that fires and is acknowledged without action is worse than no alert

Assign owners, and a deputy

Routine Needs
Daily An owner and a deputy. It must happen when the owner is on leave
Weekly An owner
Monthly An owner, with finance involvement
Year-end A plan, a date, and a rehearsal

A routine with one owner and no deputy stops during leave, and the backlog it was catching accumulates unseen for a fortnight.

Validation

The operational routine is healthy when:

Check Pass condition
Routines are documented Written down where the deputy can find them
Owners and deputies named For each routine
Routines are actually performed Evidenced — a log, a checklist, a ticket
Findings are acted on Not merely recorded
Nothing ages unnoticed Quarantine, backlog and deadline counts stay low
The restore has been tested Within the last quarter
Year-end is rehearsed Before the year it matters

Troubleshooting

Symptom Cause Action
Problems found by users first Routines not performed, or too shallow Compare what users find against what the routine covers, and close the gap
The routine stops during leave No deputy Name one and have them perform it at least once
Findings recorded, nothing changes No owner for the follow-up Every finding needs an owner and a date
Quarantine grows week on week Documents corrected but the cause never fixed Fix the source, not the instances
Month-end is always a crisis Deadline-sensitive items not tracked daily Move the ageing checks into the daily routine
Nobody knows how to do it Knowledge in one person's head Write it down. That is what this section is for