Preventive and predictive maintenance get presented as competing philosophies, and on an actual site they are nothing of the kind. One is a base layer covering everything you own; the other is a thin layer over the handful of assets where it pays. Almost every site that runs predictive well runs it on top of a working preventive programme, and almost every predictive programme that produced nothing was installed instead of one. The layering is the useful way to think about it, and it changes what you buy first.
The base layer covers everything, because it needs nothing
Intervals work on any asset with no instrumentation, no baseline and no specialist. That universality is why preventive is the base: it is the only strategy you can apply to the whole list on a Monday. It is also the layer that produces the failure history, which is the evidence any later decision about monitoring will need, and which cannot be reconstructed after the fact.
The thin layer goes where the warning exists and the failure is expensive
Monitoring pays on assets whose failure is costly and whose failure mode announces itself: bearings, motors, large rotating plant, sometimes transformers. That is a small share of any asset list. Instrumenting broadly is how the cost of a predictive programme runs away while the benefit stays concentrated in the few assets that would have justified it alone.
What changes on a monitored asset, and what does not
On a monitored asset, condition decides when work happens instead of the clock, so the interval becomes a backstop rather than the trigger. What does not change is everything else: the asset is still on the same list, findings still route into scheduled work, and the record still lives against the asset. The two layers share one system, and a site running them in two separate systems has doubled its admin for nothing.
Questions people ask about preventive and predictive maintenance
Should we run both?
Nearly every site does, whether or not it says so. The decision worth making explicitly is which assets are in the thin layer, and that is decided by consequence and by whether the failure gives warning.
Does predictive make preventive unnecessary on that asset?
It changes the trigger and rarely removes the work. Most sites keep a long-interval preventive backstop on monitored assets so that a monitoring failure does not mean no maintenance at all.
Where does this product sit?
Upkeepvo is the base layer: assets, intervals, checklists and the record of completed work. Condition-monitoring hardware and its interpretation are deliberately outside what it does.