
Ask a shift supervisor what the line's OEE was yesterday and you will usually get a number based on memory and a downtime register filled in at the end of the shift, not during it. That is not the supervisor's fault — writing down every stoppage as it happens is not realistic on a busy floor. But it means the number going into your monthly review is a guess dressed up as data.
Why the paper number is usually wrong
Short stoppages under a few minutes rarely get logged, even though they add up to real lost time across a shift. Reasons get written down after the fact, often generalised as “machine issue” when the actual cause varies. None of this is anyone lying — it is just what happens when data entry competes with running the line.
What changes with data straight from the PLC
- Every stop and start is timestamped automatically, including the short ones nobody writes down.
- Availability, performance and quality — the three parts of OEE — are calculated the same way every time, not estimated differently by each supervisor.
- You can see which specific fault code or changeover is costing the most time, not just a vague total.
- The dashboard updates through the shift, so a problem shows up while it can still be fixed that day.
What it takes to set up
Most PLCs already hold this data; the work is connecting to it safely, over OPC UA or the protocol your machines already speak, without touching how the line runs. Older machines without a PLC output can usually be fitted with a simple sensor to capture run and stop states. This is a few weeks of work per line, not a plant-wide shutdown.
Questions to ask any vendor
- Does it read from the PLC directly, or rely on someone still logging stops by hand?
- Can it connect to machines from different manufacturers on the same dashboard?
- Where does the data sit, and can our own team access the raw numbers, not just a summary screen?
- What happens to the connection if the network goes down mid-shift?
How we can help
We connect to your machines' PLCs and build OEE dashboards on real data, using Thudippu, our model for reading machine signals, to catch faults a simple stop-start log would miss. See our factory work, or talk to us to tell us about your line.


