CMMS Implementation Checklist: A 6-Week Project Plan That Works

CMMS Implementation Checklist: A 6-Week Project Plan That Works

Most CMMS implementations don't fail because the software is bad. They fail because the rollout is treated as an IT install instead of a change to how the team works. The system goes live with 400 half-named assets, nobody trusts the data, the old spreadsheet quietly stays authoritative, and six months later the licence renewal becomes an awkward conversation.

The fix isn't complicated. It's a checklist, worked through in order, with one rule underneath it: go live small, on real work, then expand.

This is the plan I give teams moving off spreadsheets or replacing a system that never stuck. It's sized for operations with roughly 10–500 assets — the plan compresses for smaller teams and stretches for multi-site rollouts, but the phases don't change.

Want it as a working document? The Excel version has all 28 tasks across 6 phases, a week-by-week project plan, and a data prep template:

Phase 1: Scope (Week 1)

Skip this phase and every later phase gets harder. Four tasks:

Write down the three problems the CMMS must solve. Be specific. "Better maintenance management" is not a problem statement — "20% of PM tasks completed late" and "audit prep takes two weeks of manual compilation" are. These become your day-90 scorecard, and they'll settle a dozen configuration arguments later. (Starting with problems, not features is the single biggest predictor of implementation success.)

List the assets in scope for go-live. Not every asset you own — the ones where the three problems live. Critical production equipment, statutory-inspection items, the vehicles with MOT dates. You can add the office chairs later, or never.

Name an owner with time actually allocated. An implementation "owned" by someone with a full day job stalls at week three. Half a day a week for six weeks is a realistic minimum.

Agree what success looks like at day 90. Tie it to the three problems: 95% PM completion, zero missed statutory inspections, audit evidence pulled in minutes. Numbers, not vibes.

Phase 2: Data Prep (Weeks 1–2)

This is where implementations are won or lost, and it happens before you touch the new system.

Export everything you have. Spreadsheets, the old system, the paper logs in the workshop drawer. You're not migrating all of it — you're finding out what exists.

Standardise asset names and IDs. One convention, applied everywhere: PUMP-07, not "old pump", "pump 7 (rear)", and "P7" in three different tabs. If you're consolidating multiple spreadsheets, the spreadsheet migration guide covers the cleanup patterns in detail.

Collect compliance dates per asset. LOLER examinations, PAT testing, MOT dates, service contract renewals. These are the dates a CMMS earns its keep on — missing one after go-live undoes months of credibility.

Decide what history to migrate. Twelve months is usually enough. Older history has archive value, not operational value — keep the export, don't import it.

The Excel download includes a minimum-viable-register template: nine columns that are enough to go live. Resist cleaning more than go-live needs; perfectionism here is procrastination with a clean conscience.

Phase 3: System Setup (Week 3)

Now — and only now — the software:

1. Import the cleaned asset register. If Phase 2 was done properly, this takes an hour, not a week.
2. Set up your location structure — sites, buildings, zones — before assets, so everything imports into the right place.
3. Create users, roles, and permissions. Technicians should see their work, not the admin console.
4. Build PM schedules for critical assets first. Every schedule you build should trace back to one of the three problems. A schedule nobody asked for is a schedule nobody will follow.
5. Set up compliance schedules with lead time. A LOLER reminder on the due date is a fail; a reminder six weeks out books the examiner.
6. Configure work order categories and priorities — few enough that people pick the right one without thinking.

Phase 4: Pilot (Weeks 4–5)

The most skipped phase, and the one that decides whether the system sticks.

Run two weeks with one team, one site, critical assets only — and one non-negotiable rule: every real work order goes in the system. No parallel spreadsheet. Parallel running feels safe but it means the team never actually switches; the old spreadsheet stays authoritative and the CMMS becomes homework.

Hold a 30-minute review each week. Ask one question: what blocked you? You'll hear about naming that makes no sense on a phone screen, categories that don't match how the team thinks, and assignments landing with the wrong person. Fix all of it now — this is exactly what the pilot is for, and it's ten times cheaper to fix before fifty users learn the broken version.

Phase 5: Rollout (Week 6)

Train on real assets, not demo data. A technician who logs a work order against the actual compressor they fixed yesterday has learned the system; one who watched a slideshow hasn't.

Set a hard switch-off date for the old spreadsheet — and honour it. As long as the spreadsheet lives, it competes.

Add remaining assets in batches, using the same naming convention Phase 2 established.

Review the dashboards weekly for the first month. Not to admire them — to catch the schedule that's already slipping and the user who's silently reverted to WhatsApp.

Phase 6: Day 90 Review

Pull out the success measures from Phase 1 and score them honestly. Then do two unglamorous things: archive the PM schedules nobody follows (a dead schedule erodes trust in every live one), and collect team feedback while opinions are still fresh enough to act on.

If the numbers moved — PM completion up, audit prep down — you have the evidence for expanding scope. If they didn't, you have a specific conversation to have, which beats a vague sense that "the system isn't working."

The Two Failure Modes to Watch

Nearly every failed implementation I've seen died one of two deaths:

Big-bang go-live. Every asset, every user, every site, day one. The data quality problems all surface at once, trust collapses, and the team retreats to spreadsheets within a month. The pilot phase exists to make this impossible.

The eternal parallel run. The CMMS and the spreadsheet both stay live "just until we're confident." Confidence never arrives, because the spreadsheet keeps being the place where the real answers live. Kill it on a date, not on a feeling.

If you're still choosing a system, the CMMS selection guide covers evaluation; if you're wondering whether you've outgrown spreadsheets at all, these five signs will tell you. And if the data-prep phase is what's putting you off — that's the part we'll do for you: cleaning, mapping, and importing your existing records is included free with any annual plan.