OEE Software and Performance Analytics, One Module of the ProAlert Platform

The short version. OEE is Availability times Performance times Quality. Getting the number is the easy part and a spreadsheet can do it. The hard part is that everyone measures planned time differently, so the figures are rarely comparable, and that a low score tells you nothing about what to change. Both of those are solved by where the number lives, not by how it is calculated.

What OEE actually measures

Overall Equipment Effectiveness is one percentage describing how much of the production a machine could theoretically have delivered during its planned running time it actually did deliver. It resolves into three components, and the reason it is useful is that the three fail in completely different ways and need completely different responses.

Availability

Run time divided by planned production time. It falls when the machine is stopped during time it was supposed to be running: breakdowns, waiting for materials, waiting for a changeover, waiting for someone to come and look at it. This is the component response coordination and maintenance move.

Performance

Actual output against what the machine should have produced at its designed cycle rate over the time it was running. It falls through small stops too brief to log and through running slower than nameplate. It is the component most often invisible, because nobody writes down a nine-second stoppage.

Quality

Good units divided by total units produced. It falls through scrap and through anything needing rework. Counting a reworked part as good is the single most common way an OEE figure gets quietly inflated.

The arithmetic, and the trap in it

Multiply the three and express the result as a percentage. ProAlert computes it as Availability times Performance times Quality, divided by 10,000 because each component is held as a whole-number percentage.

The trap is that the formula is not where OEE programmes go wrong. The definitions are. Two plants can report 70 percent and 85 percent while performing identically, because one counts changeover inside planned production time and the other does not, or because one treats a scheduled break as downtime. This is why the commonly quoted benchmarks... 85 percent as world class, around 60 percent as typical... are worth much less than they appear. An OEE figure is genuinely useful compared against the same asset's own history under a fixed definition. Compared against another company's number, it is close to meaningless.

The practical consequence is that the value of OEE software is mostly in enforcing one definition consistently across every asset and every shift, so the comparison you make internally is real. We have written up exactly how the downtime clock starts and stops, and which time window the math runs against, in How ProAlert Measures OEE Availability.

What ProAlert does with it

Calculated per asset, continuously

The three components are computed at the machine level from captured state and output rather than from a shift-end return, and broadcast to connected screens as production runs, so a supervisor can intervene while the shift is still recoverable rather than reading about it next month.
Availability, Performance and Quality licensed independently

Each tier has its own subscription window, so a plant can start with Availability alone, which is the component most plants have the worst data on, and add Performance and Quality when the instrumentation and the process discipline are ready for them.
Shift-aware planned production time

Shift schedules, breaks, lunches, changeovers and planned maintenance are configured as planned downtime, so Availability is calculated against the time that should have been running rather than raw clock time. This is the definition problem above, solved by configuration rather than by convention.
Slow production and lull detection

When the rolling average cycle time exceeds its configured threshold, or activity stops for longer than the threshold allows, the asset is flagged... catching the Performance losses that are individually too small for anyone to record by hand.
Targets resolved from the tooling, not just the machine

In stamping and injection molding the die determines the expected output, not the press. ProAlert resolves cycle targets and per-cycle-output thresholds from the active die's cavity count, so a press running the wrong die is not scored as though it were running the right one. See die-driven manufacturing.
Rollup to an executive view

An Organization Health dashboard scoped by business unit and period, with configurable thresholds and restricted to C-level roles, sits above the per-asset figures. KPI targets are set per asset, and the dashboards export to CSV, Excel and PDF.
Machine connectivity without middleware

ProAlert EdgeSense reads GPIO and digital I/O and speaks Modbus TCP/RTU to PLCs (Programmable Logic Controllers) directly, counting cycles and detecting production state, at roughly $500 per unit against industrial middleware that typically runs $20,000 or more per line. See EdgeSense.

The Part a Standalone Product Cannot Do

A Low Score Is a Question, Not an Answer

This is the argument for running performance analytics inside an operations platform rather than buying an OEE product on its own. Every OEE tool can tell you a press ran at 61 percent. Recovering that 61 percent is maintenance work, parts, quality investigation and scheduling... and if those live in other systems, the dashboard is a report nobody can act on without a week of cross-referencing.

Availability points at maintenance

The stoppages behind the Availability loss are already work orders, failure records and ISO 14224 (Equipment Reliability Data Standard) classifications on the same asset, with the labor and parts they consumed attached. The Pareto of what is actually costing you uptime builds itself. Maintenance management.

Quality points at the nonconformance

The Quality component is fed by scrap captured at source with per-source thresholds and an approval queue, and by the nonconformance register behind it, with defect heat mapping by physical location on the part. Quality and nonconformance.

The gap gets an owner

A miss against plan becomes a tracked gap on the A-FMDS daily board, promotes to a problem with an owner and a due date, and escalates on its own if nothing moves. That is the difference between measuring OEE and improving it. ProAlert A-FMDS.

The loss has footage

Event-triggered capture from any RTSP (Real Time Streaming Protocol) camera records the seconds either side of a stoppage and correlates the clip to the record, so an unexplained Performance loss can be watched rather than guessed at. Video management.

The loss carries a cost

Labor, parts, downtime and the rate of the crew running at the time land on asset profitability, which turns a percentage into a number a finance director will act on. Performance analytics.

The plan is the denominator

Shift schedules, planned downtime and the production plan the asset was measured against are configured in the same system that scores it, so the denominator cannot drift away from what the floor actually ran. Production planning.

Keeping It In Proportion

The Other Thirteen Disciplines

Performance analytics is discipline eight of fourteen. Here is the rest of the platform it sits inside.

For the full per-asset feature list see live OEE in the feature library. If you are here because your plant runs an OEE tool, a maintenance system and a quality system that disagree with each other, the convergence case is the page to read next.

Common Questions

Questions Buyers Actually Ask

What is OEE?

OEE stands for Overall Equipment Effectiveness. It is a single percentage describing how much of the production a machine could theoretically have delivered during its planned running time it actually did deliver, expressed as the product of three components: Availability, the share of planned time the machine was actually running; Performance, how close its output rate came to its designed rate while it was running; and Quality, the share of what it produced that was good the first time. An OEE of 100 percent would mean producing only good parts, as fast as the machine is designed to run, with no stoppages during planned production time.

How is OEE calculated?

OEE is the product of its three components. Availability is run time divided by planned production time. Performance is actual output divided by the output the machine should have produced at its designed cycle rate over the time it was running. Quality is good units divided by total units produced. Multiply the three together and express the result as a percentage. The definitions look simple, but the numbers are only comparable if everyone agrees what counts as planned production time and what counts as a stoppage, which is where most OEE programmes actually succeed or fail.

What is OEE software?

OEE software captures machine state and output automatically rather than from clipboards and shift-end recollection, applies one consistent definition of planned time and downtime across every asset, and calculates the three components continuously instead of in a month-end batch. The value is not the number itself, which a spreadsheet can also produce. It is that the number is calculated the same way everywhere, arrives while the shift can still be recovered, and carries the underlying events so a low score can be traced to the specific stoppages that caused it.

Is ProAlert an OEE system?

No. ProAlert is a single-database operations platform covering fourteen operational disciplines for asset-intensive manufacturers, of which performance analytics, including OEE, is one. The other thirteen cover maintenance management, procurement and supply chain, quality and nonconformance, asset lifecycle and compliance, A-FMDS daily management, workforce and labor costing, production planning, video management, industrial IoT, mobile, safety broadcast, the platform foundation and shop-floor response coordination. The distinction matters commercially, because a low OEE score is only useful if you can act on it, and acting on it means maintenance, parts and quality... which here are the same database.

What is a good OEE score?

The figures most often quoted are 85 percent as world class for discrete manufacturing and around 60 percent as typical, but these should be treated with caution. OEE is highly sensitive to how a plant defines planned production time, so two plants reporting 70 percent and 85 percent may be performing identically and simply counting differently. The number is far more useful compared against the same asset's own history under a fixed definition than against an industry benchmark.

Why is OEE alone not enough?

OEE tells you that an asset underperformed. It does not tell you what to do about it, and standalone OEE products stop at the number. Recovering the loss requires the maintenance history that explains the stoppages, the parts and labor to fix the cause, the nonconformance record behind the quality component, and the production plan the asset was measured against. When those live in separate systems, the OEE dashboard becomes a report nobody can act on without a week of cross-referencing.

See live OEE on your own machine types.

A 30-minute demo covering how planned time is defined, how the three components are captured, and what happens to a low score once it exists.

Schedule a Demo