CMMS features lists are long because feature lists sell software, and they are a poor way to choose because almost every product in the category has almost every feature. What differs is how easy each product makes it to do the four things that matter, and how much of the rest you will actually turn on. This page separates the features that decide the outcome from the ones that decide the demo, including the two that demo best and are used least: reporting dashboards and integrations.
The four that decide: raise, one-asset, hours, findings
Raising a job at the machine in under a minute. One job against one asset, enforced. The schedule visible as hours rather than dates. A found field that is hard to skip. Every other feature is downstream of these, because they determine whether there is any data worth having.
Reporting demos best and is used monthly at most
Dashboards are what a manager sees in a demo and what a manager looks at once a month. They matter, and they matter far less than the daily raise-and-close loop that produces the numbers on them. Weight them accordingly when comparing.
Integrations are usually wanted later and specified too early
The common ones are finance for asset numbers and stores for parts. Both are more useful once the register is stable, and both are frequently specified during selection, paid for, and never configured. Ask whether the integration is included or a project.
Features worth checking that nobody demos
Bulk edit, because you will rename things. Attachment size limits, because photographs are the found field people actually use. Export, because it decides your position at renewal. History retention, because a schedule is judged on two years of findings and some plans keep one.
Questions people ask about cmms features
Do we need mobile offline?
If plant rooms, basements or yards have no signal, yes, and test it rather than trust it. A job lost offline is worse than no mobile at all, because the technician believes it saved.
Is barcode or QR scanning worth it?
Usually yes. It removes the commonest source of bad data, which is a job logged against the wrong one of two similarly named machines.
What about AI features?
Judge them on whether they change the raise, the schedule or the findings. Anything that only summarises data you already have is a convenience rather than a reason to choose a product.