OEE monitoring
OEE software that measures the line, not the logbook
OEE software calculates overall equipment effectiveness from live machine data: availability (was it running), performance (at what speed) and quality (how many good parts). It replaces end of shift estimates with a number every supervisor can trust, and it shows exactly which of the three is costing you output.
Free scoping call. We start with one line, never change your machine logic, and you own the code and the data.
- material movements tracked a day, integrated steel plant
- 720,000+
- fewer dispatch errors with RFID and barcode dispatch
- 90%
- items an hour through RFID warehouse automation
- 16,800
Company reported project figures, not independently audited, not a promise for another plant. Samsung Business Insights: our RFID warehouse automation story (opens in a new tab). Forbes India Brand Connect (sponsored): founder profile (opens in a new tab).
Sound familiar?
- OEE is calculated in a spreadsheet at the end of the month, if at all.
- Everyone agrees the line "loses time", but nobody can say how many minutes, where, or why.
- Supervisors write down the long breakdowns and never the two minute stops.
- Two plants report OEE, and the numbers cannot be compared.
Each of these is output, margin or time you have already paid for and are not getting back.
What you get
A trustworthy OEE number
Availability, performance and quality from machine signals, by machine, shift and product.
Your top five losses
Ranked by minutes lost on the bottleneck, reviewed with production and maintenance.
Reason capture at the line
A short reason list on a tablet, in the words your operators use.
Code and data you own
Documented, in your repository, connected to your ERP.
What we have built
Our QR based maintenance and video SOP systems support blast furnaces, coke ovens, rolling mills and processing lines, with more than ₹80,000 crore (about US$9.6 billion) of machinery under maintenance.
Figures are company reported from past projects. They are not independently audited benchmarks and are not a promise for another plant. Read the case studies.
How we work
What we build
We build OEE monitoring as part of a manufacturing execution system or as a standalone first step: machine connectivity, reason capture at the line, loss dashboards and shift reports, integrated with your ERP. You own the code and the data. If you already run an MES or SCADA, we connect to it instead of replacing it.
How an engagement starts
- 1Pick the bottleneck. One machine or line whose output limits the plant.
- 2Connect and baseline. Capture state and counts for two to four weeks, with reasons entered at the line.
- 3Agree the losses. Review the top losses with production and maintenance, and fix the definitions.
- 4Act and extend. Run improvement on the top loss, then extend to the next lines on the same data model.
How it works
What separates useful OEE software from a dashboard
- Automatic state and count capture from the PLC, a machine signal or a retrofit sensor, so the number does not depend on memory.
- A downtime reason tree the operators actually use: short, specific to the machine, entered on a tablet at the line.
- Ideal cycle times held per product, not one number per machine, or performance will look wrong every time the product changes.
- Micro stop handling: stops shorter than a threshold count as performance loss, longer ones prompt for a reason.
- Loss views that answer questions: top five losses this week, by shift, by product.
- Links to action: a maintenance request, a quality hold or a shift note from the same screen.
How we connect OEE to your machines
Modern machines expose state and counts over OPC UA, Modbus TCP or the controller vendor protocol. Older machines usually have a cycle complete signal and a running contact we can read through a small edge I/O module without touching the machine logic. Where there is no signal at all, a current sensor on the main motor tells run from stop reliably.
The edge device buffers data if the network drops, so no shift is lost. Data moves over MQTT into the OEE application, and production orders come from your ERP so every count lands against the right product and order.
Mistakes that make OEE numbers meaningless
- Excluding planned stops so generously that OEE looks high while output stays flat.
- Using nameplate speed as the ideal cycle time for a machine that has never run at it, which makes performance look permanently poor.
- A downtime reason list with 80 options, most of which never get picked, and "others" as the top reason.
- Comparing OEE between plants that define it differently.
- Chasing a benchmark instead of your own trend on your own bottleneck.
Background
What OEE software does
Overall equipment effectiveness compares what a machine produced with what it could have produced in the time it was scheduled to run. OEE software collects the three inputs automatically: run and stop signals, part counts and reject counts. It then calculates OEE by machine, line, shift, product and operator, and shows where the gap comes from.
The value is not the percentage. It is the loss breakdown behind it. A line at 55 percent OEE with most of its loss in changeovers needs a different plan from a line at 55 percent that loses output to slow running or scrap.
How OEE is calculated
Availability is run time divided by planned production time. Performance is ideal cycle time multiplied by total count, divided by run time. Quality is good count divided by total count. OEE is the product of the three.
Example: a press is planned for 420 minutes and stops for 60, so availability is 360 ÷ 420 = 85.7 percent. At an ideal cycle of 1 minute it makes 300 parts in 360 minutes, so performance is 300 ÷ 360 = 83.3 percent. 285 are good, so quality is 95 percent. OEE is 0.857 × 0.833 × 0.95 = 67.8 percent.
| Factor | Losses it captures | Typical fix |
|---|---|---|
| Availability | Breakdowns, changeovers, material or operator waiting | Downtime reason analysis, SMED, maintenance planning |
| Performance | Small stops, reduced speed, idling | Micro stop capture, speed loss by product |
| Quality | Scrap and rework, startup rejects | Defect by cause and shift, inline inspection |
When we are not the right choice
- You want an off the shelf OEE app with a monthly subscription and no integration. Several good ones exist.
- Nobody on the floor has time to act on the losses. The number alone will not change output.
Common concerns, answered
Will this disrupt production?
No. We connect read only first, install during planned stops, and edge devices buffer data so a network problem never stops a machine. Nothing writes to a PLC without your controls team and a test.
Our machines are old. Do we need to replace them?
No. Old machines are the usual case. A running signal, a cycle signal or a current sensor on the motor is enough to start, without touching the machine program.
We tried a system before and nobody used it.
That is the most common failure, so we design with the operators and supervisors who will use it, start with one report they already fill in by hand, and only extend once they rely on it.
What about IT security?
Read only by default, a segmented OT network, and a review with your IT team before anything connects, following the ISA/IEC 62443 zone approach.
We will not replace SAP or Oracle. Do we have to?
No. We connect to the ERP you already run through its standard interfaces and add what the shop floor is missing. Your ERP stays the system of record.
Do your engineers understand a heavy industrial floor?
Our systems run in steel plants, warehouses and process lines. We start every project by walking the line with your operators and supervisors, and the first release is tested by them, not only by IT.
Are we locked in?
No. You own the source code and the data, documented, in your own repository. Another team can maintain it.
Frequently asked questions
What is a good OEE score?
There is no universal good score, because OEE depends on the process, the product mix and how planned time is defined. Many practitioners quote 85 percent as world class for discrete manufacturing, but the more useful target is steady improvement on your own bottleneck using a consistent definition.
Can OEE software work on old machines?
Yes. Old machines usually have a running contact or a cycle signal that an edge I/O module can read without changing the machine program. Where there is no usable signal, a current sensor on the motor distinguishes running from stopped.
Do operators still need to enter anything?
Operators enter the reason for a stop and, where needed, rejects. Run time, stop time and counts are captured automatically. Keeping the reason list short and specific is what makes the data usable.
Is OEE software the same as an MES?
No. OEE monitoring is one function of an MES. An MES also manages work orders, material, quality, genealogy and ERP integration. Many plants start with OEE and grow into MES on the same data.
How long does it take to get a first OEE number?
For one connected line, a first reliable baseline usually takes two to four weeks after connection, because reason codes and ideal cycle times need to be corrected with the people who run the line. The exact time depends on machine access and data quality.
Related reading
References
Start with one line and one number
We scope one line, agree the number that should move, and show it moving before anything is scaled. You keep the code and the data.
Step 1
An engineer reads it
Not a sales bot. Our team replies within 24 hours with questions or times for a call.
Step 2
30 minute call
Your process, machines, systems and the number you want to move. We tell you honestly if we are a fit.
Step 3
A written scope for one line
What we would connect, what you would see, how long it takes. You decide; nothing starts without your approval.
Reviewed by Amey Kadle, Founder, Ajinkya Technologies. Last reviewed: .