Key Points
- Asset prioritization is a four-factor decision. They are the consequence of failure, monitoring viability, failure-mode predictability, and execution capacity. All have to be determined before an asset is a smart first move.
- The sensor's measurement modalities must match the failure modes the prioritized asset actually exhibits, or detection arrives after the early window has closed.
- Lead time between potential failure and functional failure determines whether the alert is actionable. Detection without lead time is a record, not a tool.
The most common first move is usually the wrong one
The implementation team has a budget for condition monitoring. The case has been made and accepted for asset monitoring. Now, the question that will shape the next 12 months sits before the reliability lead. That question is “Where to start?”
The whiteboard list of critical assets is the (seemingly) obvious first move for this team. And yet, the prioritization factor used to create that list is the same misstep that causes many asset monitoring programs to drain the expected return on their deployments.
Deploying with an incomplete determination of the factors to consider leaves sensors on the ‘highest-stakes’ assets without a clear answer about, for example, whether those sensors can actually detect the failures they are meant to act against. Or, does it produce alerts that arrive without enough lead time for the team to plan against? And, does it produce alerts that close as missed catches because the parts are not on the shelf or the procedure is not documented?
Deloitte research finds that poor maintenance strategies can reduce a plant's overall productive capacity between 5 and 20 percent, and the prioritization decision sits upstream of every one of those losses.
The framework below walks through the prioritization logic across four factors, then describes what the asset monitoring system must do once prioritization is settled. Read both halves together. The first tells the team where to start, and the second determines whether starting there actually produces the outcomes the prioritization promised.
The four factors at a glance
Four Factors That Should Govern Asset Monitoring Prioritization
Four factors determine whether monitoring an asset yields decision-grade information or becomes dead weight on the program. Each one quietly determines a different part of the outcome, and most of them get overlooked when criticality is treated as the whole conversation. Working through them in sequence is what turns prioritization from a ranking exercise into a decision the team can defend later.
Consequence of failure
This is the familiar starting point. What an asset's failure actually costs in production, safety, and downstream impact. A primary motor on a packaging line, the loss of which idles every downstream conveyor for a shift, sits well above a secondary cooling fan whose failure the production schedule can absorb without a noticeable hit.
Critical assets belong in the decision. The trap is treating them as the whole decision. Stopping here leaves sensors on the highest-stakes assets without a clear answer about whether those sensors can actually detect the failures that matter on those assets. Criticality alone is a starting line, not a finish.
Monitoring viability
The second factor asks whether the failure modes on the prioritized asset are actually detectable by the available sensing approach. Failure modes have physical signatures. Sensors have measurement ranges. The two have to overlap.
Take a centrifugal pump in which the dominant failure path is bearing lubrication breakdown. The earliest signs of that breakdown express as high-frequency acoustic energy from friction at the asperity contact, which lives in the ultrasonic range near 30 kHz. A vibration accelerometer with a useful response up to a few kilohertz will eventually detect the failure, once the bearing has progressed far enough to produce mechanical signatures in the audible band.
By then, the early window will have closed. The asset was the right call to prioritize. The sensing modality was the wrong tool for the failure mode it was meant to detect.
The pattern repeats across rotating equipment. Electrical faults on motor windings express electromagnetically before they show up mechanically. Cavitation in pumps lives in the ultrasonic range before it registers as vibration. A sensor covering only one of those domains will miss the rest, no matter how high the asset ranks on criticality.
Failure-mode predictability
A monitoring program creates value when the alert arrives in time to do something with it. The third factor is whether failures on the prioritized asset develop with enough lead time for the team to plan against.
The framework that captures this is the P-F curve, which maps the distance between two points on a failure timeline. The moment a developing fault first becomes detectable is the potential failure point. The moment the asset stops performing its function is the functional failure point. The interval between them is the lead time that monitoring is buying.
A lubrication-driven bearing failure on a gearbox-coupled motor might run on a P-F interval of six to eight weeks, which gives the team time to order parts, schedule a shutdown window, and plan labor. A sudden electrical insulation breakdown on the same motor might run on a P-F interval measured in seconds. The detection works. The window is gone before the technician sees the notification.
Predictability shapes whether the alert is a tool or a record.
Execution capacity
The fourth factor closes the loop between detection and action. An alert with no spare part in stock, no documented procedure, and no available labor window during the lead time it provides closes as a missed catch.
Consider a critical air compressor whose seal kit has a 12-week lead time from the supplier, and the maintenance team has no written sequence for the replacement because the technician who knew it has retired. The sensor on that compressor can detect seal degradation with perfect clarity. The alert can arrive weeks ahead of functional failure. But if the part is not on the shelf and the procedure exists only as one person's memory, the program has gathered an early warning that the team cannot convert.
Execution capacity rests on three things that have to be in place when the alert lands.
- Spare parts available on time.
- Standard operating procedures (SOPs) documented and accessible at the point of work.
- Labor with enough open hours to absorb the planned intervention.
Programs that monitor without these are running surveillance, not reliability.
Example scorecard comparing two assets
What the Asset Monitoring System Must Deliver Once Prioritization Is Set
Prioritization only pays off if the system the team selects can serve what prioritization identified.
The four factors point at specific failure modes on specific assets with specific lead times and specific execution requirements, and the system has to do something specific with all of that. The capability of the platform is one side of this. The infrastructure that supports it is the other.
Diagnostic capacity matched to the failure modes that matter
The first thing the system has to do (beyond detection) is diagnose. A threshold notification tells the team something has crossed a line. A diagnosis tells them what is wrong, how severe it is, and what action to take next. Prioritized assets need the second.
Named failure-mode diagnosis is the core capability.
When the system identifies "outer-race bearing wear at stage 2 on motor 4" rather than "vibration RMS exceeded threshold," the technician knows what they are looking at without manual spectrum interpretation or specialist analyst review. Severity ranking sits on top of that. A stage-1 lubrication issue and a stage-3 misalignment require different responses, and the system has to communicate the difference.
And the diagnosis ideally should flow into maintenance execution without a manual handoff. The procedure to follow and the parts the work requires should be reachable from the same alert that produced the diagnosis.
A program that diagnoses and routes the response is the one that defensibly moves the uptime number.
Operational infrastructure on the plant floor
A capable system that the infrastructure cannot support is one that delivers on paper but not on the plant floor. The platform has to be evaluated against whether it can actually reach and serve the assets prioritization just identified, which usually comes down to three things.
- Sensor-to-platform connectivity must work without relying on the plant's Wi-Fi.
Prioritized assets are often in basements, on remote pump skids, or inside metal-walled compressor rooms where the wireless network does not reliably reach. A monitoring system that requires reliable Wi-Fi at every sensor location will lose coverage exactly where it matters most.
- Older equipment must be brought into the same monitoring layer as newer assets.
Most facilities are a mix of recent installations and machinery running for a decade or more, and a system that only works on the newer side leaves the older critical assets monitored via inspection routes, which is presumably what the team is trying to move past.
- The mobile interface has to work where the technicians are, including offline.
An alert that requires the technician to walk back to a desktop loses the time the prioritization just earned. A team that gets prioritization right and then picks a system whose infrastructure cannot carry the deployment will watch a good decision underperform.
How Tractian Supports the Full Framework
Tractian's condition monitoring solution operates across every part of the framework, from prioritization decisions through execution of the work the alerts identify.
The Smart Trac multimodal sensor expands monitoring capabilities by integrating four sensing technologies into a single device.
- Triaxial vibration up to 64,000 Hz captures mechanical degradation.
- A piezoelectric transducer up to 200 kHz captures the ultrasonic signatures of friction, lubrication breakdown, and cavitation.
- A magnetometer captures electromagnetic signatures of electrical faults and tracks RPM on variable-speed assets.
- Temperature rounds out the coverage. One sensor adapts across rotating equipment categories without dedicated hardware per machine type, which means monitoring viability holds up across the diversity of a prioritized fleet.
Auto Diagnosis handles both failure-mode predictability and diagnostic capacity. The AI recognizes all major failure modes with named diagnoses, evidence, and severity ranking, so an alert arrives as "stage-2 bearing wear with the fault frequency attached" rather than as a threshold notification. Criticality-based alerting means more critical assets trigger warnings at earlier P-F points, aligning the consequence-of-failure factor with the timing of the response.
Execution capacity sits inside the same platform. Every alert ships with prescriptive guidance from the attached procedures library, and the diagnosis routes natively into a Tractian-enriched CMMS as a work order, with inventory linked and mobile execution working offline. The path from detection to completed repair runs through one system rather than across three.
Learn more about Tractian's asset monitoring system to see how high-quality, decision-grade IoT data transforms your program into AI-powered closed-loop workflows.
FAQs about Asset Monitoring Systems
How do I decide which assets to monitor first?
Use the four factors together. Consequence of failure, monitoring viability, failure-mode predictability, and execution capacity. Criticality is a starting point, not the whole decision. It does not tell you whether the sensor can see the failure, whether the team has time to act, or whether parts and procedures are ready.
Is asset criticality enough to decide what to monitor?
No. Criticality identifies what failure would cost, not whether the failure modes are detectable, whether the lead time supports action, or whether the team can execute. All four factors have to be clear before the asset is a smart first move.
What is monitoring viability?
Whether the failure modes the asset actually exhibits can be detected by the sensor's measurement range. A vibration-only sensor on a pump whose failures express ultrasonically will miss the early window regardless of how critical the asset is.
What is the difference between an asset monitoring system and a sensor?
A sensor produces data. A system diagnoses what the data means, ranks the response by severity, and routes the work to execution. Sensors alone produce dashboards. The surrounding system produces reliability outcomes.
How long does an asset monitoring system take to pay off?
Published benchmarks for advanced platforms cite a three-month payback when both prioritization and execution are in place. The variable is rarely the technology. It is whether the prioritized assets clear all four factors and whether the team can act on what arrives.
What is the most common mistake when starting an asset monitoring program?
Treating prioritization as a criticality ranking and stopping there. The result is sensors on the right assets but the wrong failure modes, alerts without lead time, or alerts the team cannot act on because parts and procedures are not ready.
