Key takeaways
- A light MES starts from the machine: stops, speed and OEE. A full MES starts from the order: routing, dispatch and genealogy. They answer different first questions.
- Let your own data decide. Two weeks of stop and count data from one line shows where the minutes really go.
- When most of the lost time is stops and slow running, the money is at the machine, and that is the part a light MES covers first.
- Unit-level traceability, paper travellers and dispatching by phone are the signs that you need full MES functions as well.
Every MES conversation reaches the same fork sooner or later: the light version that watches the machines, or the full system that runs the orders. Vendors are happy to choose for you. The cheaper way is to let one line choose: two weeks of honest stop and count data shows whether your losses live at the machine or at the order. For the definition, what a light MES leaves out and a five-question test, read Fabrico's guide light MES: what it is and when it is enough. This page is the measuring part.
The two start from different questions
- Light MES: where did the time go, and why? It runs on machine signals plus the reasons operators give for each stop, and it shows OEE and a downtime Pareto by line and shift.
- Full MES: what runs next, and was it made right? It runs on orders, routings, operator confirmations and material lots, and it gives dispatch lists, electronic travellers and genealogy.
Neither is a cut-down copy of the other. A plant can need one, the other, or one now and the other later. The question is which problem costs you more minutes today.
The OEE math that decides it: one shift on one line
- Availability: 480 planned minutes, 96 minutes of stops, so the line ran 384 minutes. 384 ÷ 480 = 80.0%.
- Performance: with a 30-second ideal cycle, 384 minutes could make 768 parts. The line made 640, so 640 ÷ 768 = 83.3%, and 64 minutes went to short stops and slow running.
- Quality: 620 of the 640 parts were good first time, 96.9%. The 20 bad parts used 10 minutes of machine time.
- OEE: 620 good parts × 30 seconds = 310 minutes of good output, out of 480. 310 ÷ 480 = 64.6%.
Of the 170 minutes lost, 160 (94%) were lost at the machine, to stops and speed, and only 10 to quality. On this line the order side is not where the money is: the first system worth having is the one that sees every stop and its cause. Put your own shift into the OEE calculator to get the same split.
Now flip it. A line that runs well but ships parts built to the wrong revision, or a plant whose customer audit asks for the material lot of every unit, has a problem no stop log will fix. That is full MES territory, whatever the OEE says.
What to measure first: two weeks, one line
- Every stop with its start, end and reason. A clipboard catches the long ones; the short ones need the machine signal, or they vanish into the performance loss. The downtime log template and the free OEE tracker both work for a first pass.
- Total and good count per shift, so performance and quality are numbers, not opinions.
- The ideal cycle time, from the machine specification or the best rate the line has held (the cycle time calculator shows the gap).
- One changeover, timed from the last good part to the first good part.
- Every wait for an order, material or an instruction. These are the minutes a full MES would chase, so give them their own reason code.
After two weeks, sort the lost minutes into three piles: stops and speed, quality, and waiting for orders or material. The biggest pile tells you where to start, and the six big losses calculator turns the first two piles into minutes per loss.
Log a shift, get the OEE split and a downtime Pareto, free, in your browser.
Signs you need more than a light MES
- Regulators or customers require you to trace every unit or batch back to its materials and process data.
- Work instructions and travellers are on paper and change often, and the wrong version reaches the line.
- Jobs are dispatched by whiteboard and phone, and the wrong job runs often enough to show up in your waiting pile.
- Your ERP cannot hold routings and BOMs well enough to plan from.
And the signs that a light MES is the right place to start: most of the lost minutes are stops and slow running, the ERP runs orders well enough, and the daily question on the floor is "why did it stop?", not "what runs next?".
Common mistakes
- Choosing from a feature list. Two weeks of data from one line beats any demo script.
- Counting only the long stops. The short ones hide in the performance loss, which is why a stop log on paper always looks better than the machine.
- Treating ERP output as OEE. The ERP knows what was reported, not when the machine stopped or why.
- Going plant-wide on day one. Prove it on one line, then copy what worked.
Where Fabrico fits
Fabrico is a light MES: machine data collection, production monitoring, OEE and downtime analytics, and integration with SAP and other ERP systems. Job dispatching, BOMs, routings and genealogy stay in your ERP. It reads automated lines from the PLC, fits an IoT device with external sensors to older machines, and uses AI cameras at manual stations. The full scope is on Fabrico's MES and OEE software page.
See your own line's losses with a price tag
Fabrico logs each stop as it happens with its start time and length, operators tag the cause on a tablet at the line, and Fabrico shows the OEE split and the downtime Pareto per line and shift.
Book a Fabrico demoFAQ
What does light MES mean?
Light MES, MES lite and lightweight MES name the same idea: the part of a manufacturing execution system closest to the machine. It collects machine data, logs stops with their causes and turns them into OEE and downtime figures, while orders, BOMs and routings stay in the ERP.
Is a light MES the same as OEE software?
Close, but not quite. OEE software can be only the calculation. A light MES usually adds the plumbing around it: the machine connection, stop reasons from operators and a link to the ERP for orders and products.
How long should I measure before choosing?
Two weeks on one representative line is usually enough to see where the minutes go, because it covers several shifts, changeovers and the odd bad day. Measure longer if the line runs very different products from week to week.
Can I start with a light MES and add full MES functions later?
Yes, as long as the system you start with links to the orders in your ERP. The machine data and the loss model carry over, and order functions such as dispatching or genealogy can be added where the ERP falls short.
Do I need a PLC on every machine?
No. Automated lines can be read from the PLC, older machines can get an add-on sensor that sees running and stopped, and manual stations can be counted by a camera or by the operator.
Related: how to implement OEE · OEE data collection methods · how to track downtime · what is OEE · OEE calculator