Pillar guide·Maintenance

Choosing a CMMS: 12 questions to ask vendors

Twelve questions that separate real CMMS contenders from glossy demos - covering data import, mobile UX, integrations, hosting, and the hidden costs that bite later.

UtilityRadar Team May 9, 2026 8 min read

Twelve questions separate real CMMS contenders from glossy demos. Bring this list to every vendor call. Anyone who deflects more than two of them is not the right partner for a wastewater utility.

Group A: data and migration

The single biggest source of failed CMMS rollouts is bad data migration. The vendor who has actually done this before will answer specifics; the one who has not will hand-wave.

1. Can you import our existing asset list and PM schedule on day 1? The right answer is "yes, send us a sample export from your current system in any reasonable format and we will turn it around in two weeks". The wrong answer is "we have an import wizard". An import wizard means your team will spend three months cleaning the data, not the vendor.

2. What is the data export path if we leave? A CMMS that cannot produce a clean export of every asset, work order, PM, and attachment is a hostage situation. The vendor should commit, in writing, to a documented export format and a reasonable timeframe (30 days is fair). If they cannot describe the export schema on the call, walk away.

Group B: field operation

The mobile experience is where rollouts succeed or fail. Test it on the actual hardware your crews will use.

3. Show me the mobile app working offline for 4 hours. Do not accept "it works offline" as an answer. Ask the vendor to demonstrate a full work-order open, edit, photo attach, and close cycle on a tablet in airplane mode for at least 30 minutes, with the queued work syncing once connectivity returns. If they have to "set up an environment" to do that, you have your answer.

4. How are work orders auto-created from condition data? Show me a vibration reading crossing a threshold, then show me the work order it created. Show me a SCADA alarm forwarding into an emergency work order. If condition-based work order creation is "a roadmap item", the system is reactive only and you will run it on calendar PMs forever.

5. Can a non-IT operator build their own report? Open the report builder live. Have the operations manager build a "PMs overdue by station" report in front of you. If it requires SQL, a ticket to the vendor's support team, or a paid services engagement, your operators will never get the data they need.

💡 Demo on your own data Insist that the demo run against a sample of your real asset list, not the vendor's polished demo plant. Five hundred assets you recognise tells you more in 20 minutes than two hours of generic walkthrough.

Group C: integrations and platform

A wastewater CMMS that does not integrate cleanly with SCADA, GIS, and the historian is an island. Islands sink.

6. What is the integration story with our SCADA, historian, and GIS? Ask for named customer references where each integration is in production today. "OPC UA" is a standard, not a working integration. "We connect to Wonderware Historian via the OPC HDA driver and pull the following tags into the asset record" is a working integration. Same question for ESRI ArcGIS or QGIS for asset location overlays.

7. Where does the data live and who owns it? Multi-tenant cloud, single-tenant cloud, on-premise, or hybrid? Which jurisdiction hosts the data, and does that satisfy your regulator and your security policy? In Europe, ask specifically about GDPR data residency. In the US, ask about FedRAMP or state-equivalent requirements if you are a public utility.

8. Show me the audit trail for one work order from creation to close. Open a work order in the demo system. Show me every state change, every field edit, every attachment, every signature, with timestamps and user IDs. If the audit trail is fragmented across two screens or stored in a separate "history" tab, your inspectors will not accept it.

Group D: total cost and ops

The contract is signed in year 1. The bill is paid in years 2 through 10. Year 1 pricing is a marketing exercise; year 3 pricing is reality.

9. What does it cost in year 3 (not year 1)? Get the price escalation in writing. Get the cost per additional user, per additional asset, per additional integration, per additional GB of attachments. Compute the all-in 5-year cost of ownership. Many vendors are 30% cheaper in year 1 and 80% more expensive in year 5.

10. How do you handle a plant-wide outage of your service? The CMMS goes down on a Friday afternoon. What happens to the 24/7 operation? The right answer involves an offline mobile cache that lasts at least 48 hours, a documented incident response process, an SLA with credits if the vendor breaches it, and a public status page. The wrong answer is "we have 99.9% uptime".

11. Can you talk to your closest customer running a similar utility? Not a logo on a slide. A name, a phone number, and the customer's permission to have a 30-minute call. Ask the customer what went wrong during the rollout, not what went right. The vendor's answer to this question, before you make the call, tells you most of what you need to know.

12. Who is our implementation lead and what is their background? A name, a CV, and the percentage of their time committed to your project. The implementation lead is the single biggest predictor of rollout success. A senior consultant with three wastewater-utility deployments under their belt is worth more than a junior project manager and a 200-page methodology document.

Why these 12

The list is not arbitrary. It is the residue of patterns across CMMS rollouts that succeeded and ones that failed. Failed rollouts almost always trace back to a "yes" the vendor gave on questions 1, 3, 6, or 12 that turned out, six months in, to mean "yes, with caveats we did not mention". The 12 questions are designed to expose those caveats during the sales cycle, when you still have leverage.

Three questions test the platform (offline mobile, condition-based triggers, audit trail). Three test the data and integration story (import, export, SCADA/historian/GIS connectivity). Three test the commercial reality (year 3 pricing, outage handling, contractual residency). Three test the human factors that determine whether the project lands (implementation lead, reachable reference customer, non-IT report builder). Together they cover roughly 90% of the post-mortem material from rollouts that go sideways.

Red flags during a demo

Even before the 12 questions, certain demo behaviours predict trouble.

⚠ Vendor types in advance If the demo screen is full of pre-typed text and pre-attached files, the vendor is hiding how slow the real workflow is. Ask them to clear the form and start over while you watch.

No offline demo. The vendor says "we can show you offline mode in a follow-up call". They cannot, because offline mode is broken or not yet built. Decline to schedule the follow-up; move on.

No exit clause in the contract draft. Standard CMMS contracts include a defined termination notice period and a defined data export deliverable. A vendor who pushes back on either of those clauses is signalling that they intend to make leaving expensive.

Cannot name the implementation lead. "We will assign someone after contract signature" means you are buying a project manager from a pool, not a partner. Insist on a named individual in the contract.

Reference customer is a logo, not a name. Every vendor has a slide of customer logos. Few will hand over a phone number for one. The phone number is what you need.

Demo runs on a perfect dataset. Real wastewater data is messy: orphan assets, half-populated PMs, inherited part numbers from three predecessor systems. Ask the vendor to demo on a sample of your real data, with all the warts. Their willingness to do so predicts the quality of the rollout.

UtilityRadar
More
Press Esc to close · Advanced search