Every small-business IT guide hands over a checklist. Almost none explains why most of them stop working by March. The document is not the problem — the ownership model is. A checklist saved as a file has no owner, no deadline and no memory, so it quietly becomes a wish list everyone agrees with and nobody executes.
Here is the uncomfortable version. The checklist is not the asset. The asset is the checklist plus a calendar, a named owner for every task, and a log that proves the work actually happened. Without those three, the twelve best-intentioned tasks decay into a PDF nobody opens twice.
The full checklist — twelve tasks across weekly, monthly and quarterly intervals — sits in the complete guide on tanit365.com. What follows is the operational half the guides skip: how to run maintenance as procedure, so the routine survives busy months, staff changes, and your own forgetfulness.
Give every task one owner
A task owned by "the team" is owned by nobody. The arrangement that survives contact with a real month: each person handles the visible part of their own machine — free space, updates, the Downloads folder. One named person owns the calendar, the backup log and the decision to escalate. A technician touches only what requires opening hardware or real diagnosis.
Ownership also decides duration. If the weekly block takes more than a few minutes per machine, the split is wrong — someone is doing someone else's task. The monthly block fits one morning; the quarterly block belongs to a quiet afternoon, done one machine at a time. Opening every machine at once is how screws get lost and cables get swapped.
Weekly is observation, monthly is numbers, quarterly is hands
The split is not arbitrary — it matches the kind of attention each task needs.
Weekly work is looking: free space on the system drive (act when it drops under roughly 15%), whether security updates actually installed, whether the last backup actually ran, and whether anything unfamiliar appeared in the startup list. Four glances, minutes per machine. Nobody installs anything during business hours; you only record which machines fell behind.
Monthly work is reading numbers: disk health trend, user accounts, the backup log, one test-restore of a random file, office software and browser updates. The trend beats the absolute value — a drive that slid from 98% to 91% across two months is telling you something; a drive parked at 88% for half a year is fine.
Quarterly work is hands-on: dust out of the hardware, a software audit, administrator-access review, a restore of a larger data set. Anything that needs the case opened waits for the quarter, not the week.
The backup test that decides everything else
A backup only counts when a restore has proven it. The monthly ritual is one random file, restored to a temporary folder and opened. Businesses discover their backup was incomplete during an actual recovery — the worst possible moment — because nobody ever tested a restore on an ordinary Tuesday.
Order matters too: back up first, maintain second. A cleaning session can loosen a cable; an uninstall can remove a shared driver. With a fresh backup in place, every task on the list becomes reversible. The layered arrangement that works: a system image per machine taken right after setup, a monthly data backup, and an offsite copy for what cannot be rebuilt — contracts, client files, photos. Whether that offsite layer is a synced drive or a cloud folder matters less than the rule behind it: no single login and no single laptop is load-bearing. The storage comparison in the full guide covers which plans behave that way.
Keep a log that makes troubleshooting cheap
Four columns: date, machine, task, anything unusual. Add a version number for software updates and the outcome of each restore test. The log's value arrives months later, when a machine misbehaves and the first question is "what changed?" A PC that slowed gradually over three months is a different problem from one that slowed right after a new install — but only if notes exist to tell them apart.
The log also catches the signal people miss: the same fault appearing on several machines in one week. Identical faults across machines usually mean a shared cause — the network, an infected shared file, an update that just landed — not failing hardware. Fixing machines one by one burns hours without touching the root.
Know where the checklist ends
Stop and call a technician in four situations: the drive shows mechanical noise or rising errors, the machine reboots without explanation, data is already lost, or the machine is under warranty — opening that case yourself can void coverage. Everything else on the list is deliberately low-risk: no tools, no disassembly, no expertise beyond patience.
The thirty-minute setup
Once: list the machines, assign each of the twelve tasks an owner, put the weekly glances on a recurring calendar slot, and create the log. Most offices have the checklist but score zero on owners, calendar and log. That score is not a criticism; it is the roadmap. Thirty minutes moves maintenance from a document to a procedure — and the next hardware failure becomes a scheduled replacement instead of a lost morning.
FAQ
How do you divide IT maintenance tasks in a small team?
By visibility and risk: each person watches their own machine weekly (space, updates, backup status), one named person owns the calendar, backup log and escalation decisions, and a technician handles anything that opens the case. If the weekly block takes longer than minutes per machine, tasks are assigned wrong.
How often should a small business test its backups?
Monthly, by restoring one random file and opening it. A backup that has never been restored is a hope, not a backup. The monthly test takes minutes and is the only reliable way to catch an incomplete job or an excluded folder before a real recovery depends on it.
What belongs in an IT maintenance log?
Date, machine, task performed, and anything unusual observed — plus version numbers for updates and outcomes of restore tests. It needs to be findable months later, when the question "what changed?" decides whether troubleshooting starts from data or from zero.
When should a small business stop fixing computers themselves?
When a drive shows mechanical symptoms, when machines reboot without explanation, when data is already lost, or whenever a warranty term forbids opening the case. One more pattern people miss: the same fault on several machines in the same week points to a shared cause, and that belongs to a professional.
Do we need monitoring software to run this routine?
No. Every weekly item is a glance, every monthly item is reading a number the system already shows. Tools help at scale, but the discipline — owners, calendar, log, restore tests — is what makes the routine work; software on top of no discipline just automates the skipping.
Top comments (0)