← Back to Overview

How We Score Machine Monitoring Platforms

How MFG Tech Review evaluates machine monitoring platforms, transparently and consistently, against published criteria.

How We Score

Every platform is evaluated independently across 12 criteria, each rated on a 0–10 scale where 10 is the best possible outcome and 0 means the capability is absent. Scores reflect publicly available information, direct product research, user reviews, and documented feature capabilities as of early 2026. No platform has paid to be included or to influence their score.

Scoring Criteria

CategoryScoring LogicWhy It Matters
Vibration Monitoring10=native vibration sensors; 7=third-party sensors supported; 0=unsupportedPrimary indicator of mechanical failure in rotating equipment
Power Monitoring10=dedicated power/current sensing; 5=partial; 0=noneElectrical signature detects motor health and machine state
PLC Integration10=native PLC; 5=via adapters; 0=bypasses/unsupportedEnables deeper machine data beyond run/stop signals
Price10=price published and low cost per machine; 7=published, mid cost; 4=published high cost, or buyer-reported only; 0=no public pricingTotal cost of ownership affects ROI timeline for SMBs
Contract Terms10=no contract month-to-month; 5=annual; 2=multi-year lock-inLong contracts increase switching costs and risk
Ownership Model10=software owned outright, or offered as both perpetual and subscription; 6=hardware purchased, software subscribed; 3=nothing owned, subscription or rental onlyChoice of how to acquire the system, not a preference for one model
Predictive Maintenance10=AI-driven PdM native; 5=rule-based alerts; 0=unsupportedPdM is the highest-value application of machine data
Open API10=documented REST/GraphQL; 5=limited; 0=noneEnables ERP, CMMS, BI, and custom workflow integration
Installation Speed10=minutes; 7=hours; 4=days; 2=weeks or longerLong deployments delay value and disrupt production
Installation Cost10=included; 7=charged, under $1,000 for a five-machine shop; 4=charged, $1,000-$5,000; 2=charged above that, or charged with no figure publishedAn unquantified install fee is a budget risk, not a rounding error
Self-Install Capability10=the buyer can install it; 0=a vendor or integrator mustVendor-performed installs add scheduling, cost and dependency
OEE Tracking10=native automated OEE; 5=manual/partial; 0=not availableOEE is the universal manufacturing KPI

Machine Intelligence Score (MIS)

The MIS is a 0–100 composite index measuring overall suitability for industrial machine intelligence, weighted across 6 dimensions:

DimensionWeightComponents
Sensor & Data Coverage20.0%Vibration Monitoring, Power Monitoring
Predictive Maintenance20.0%Predictive Maintenance
OEE Tracking15.0%OEE Tracking
Commercial Flexibility15.0%Price, Contract Terms, Ownership Model
Integration Depth15.0%Open API, PLC Integration
Usability & Deployment15.0%Installation Speed, Self-Install Capability, Installation Cost

How the weights are set, and who they are set for

The weights above are chosen, and they are chosen for a particular buyer: a job shop of roughly five to twenty machines, with no IT department, buying machine monitoring for the first time. A different buyer would justify different weights, and a reader who disagrees with ours can see exactly what to disagree with.

They come from shop owner interviews, not from our own judgement. Buyers were given the six dimensions and a hundred points to divide between them. Our own first draft had put commercial terms at thirty per cent, on the reasoning that price and lock-in are the gate a small shop fails at. That is not what buyers said. They put the most on what a system can sense and on whether it warns them before something breaks, less on commercial terms, and spread the rest almost evenly.

Read plainly, that allocation says nothing dominates. Sensing breadth and failure prediction lead by five points over everything else, and the remaining four dimensions are level. A buyer at this size does not think one question settles it, and the weighting now reflects that rather than our assumption.

These weights replace a formula that set them from how much each dimension's criteria overlapped. That formula answered a real question, but it was a statistical property of the rubric rather than a statement about who this site is for. Correlations are still measured from the published scores at build time and still shown here: they run from 0.01 across criteria in different dimensions up to 0.68 for installation speed and self-install capability, so overlap stays visible rather than being absorbed silently into a weight.

When the rubric changes

Corrections to this rubric change published scores. Each one is dated and explained on the changelog, including the revision that produced the weights above.