Shop floor visibility means you can tell what each line is doing now, why, and how far it is from plan, in minutes and without walking over or asking. Measure it with three numbers: how long until you learn of a stop, how much stop time has no known cause, and how much of the shift your records cannot explain. Fix the largest gap first.
Three things that must all be true
A manager can have plenty of data and still be blind. Visibility needs three things at once, and most plants are missing at least one.
- Fresh: the information arrives while you can still act on it.
- Specific: it names the line, the stop and the cause, not only a plant average.
- Trusted: people believe it enough to act without first walking over to check.
Yesterday's report may be specific and trusted, but it is not fresh. A live screen with a blinking number and no cause is fresh, but it is not specific.
The test: time your own answers
Take a stopwatch and try to answer these questions for one line, right now. Ask from your desk and use only what you normally have.
| Question | Example: today | Good answer needs |
|---|---|---|
| Is line 3 running now? | 4 min, phone call | Live machine state |
| Are we ahead of or behind plan? | 15 min, walk to the board | Count against target by hour |
| What stopped us last, and for how long? | Next morning | Stop log with time stamps |
| What cost the most time this week? | 2 hours in a spreadsheet | Stops ranked by minutes |
| Which order is on the line, and how far along? | 10 min, ask the lead | Order linked to the count |
| Why was the night shift lower? | Not at all | Causes per shift |
Any answer that takes more than a few minutes is a blind spot. Any answer that starts "I think" is one too, however fast it came.
The biggest blind spot usually sits inside "running"
Long stops get noticed, because someone shouts. The losses nobody sees are the short stops and slow running that hide inside time the log calls running.
Published work points the same way. Performance losses, meaning short stops and slow running, dominated in about 20 cases studied by Ljungberg (1998), and operational efficiency losses had the largest impact on OEE across 98 Swedish companies (Ylipää et al., 2017).
In one study of 884 machines in 23 Swedish companies, unclassified losses were about 19% of scheduled production time (Hedman et al., 2016). That was with automatic measurement, so a signal alone does not give you the reason.
Measure your gap in ten minutes
You can size this with the numbers on one shift sheet. All you need is the scheduled time, the logged stops, the total count and the ideal rate.
Example: a line has 480 scheduled minutes, 45 logged stop minutes, an ideal rate of 50 units a minute and 18,500 units made.
The sum is 480 - 45 - 370 = 65 minutes. The log says the line was running then, but the count says it was not producing at full rate.
So the known stops are 45 minutes and the hidden loss is 65, which means more than half the lost time has no name. If you fix only what the log shows, you are working on the smaller part of the problem.
Three numbers to track
Detection delay tells you whether news moves fast enough. Unknown share tells you whether causes are being captured, and unknown downtime shows how to shrink it.
The unexplained share tells you how much the records are missing altogether. The checks in downtime data accuracy will confirm it against a second source.
What to fix, in order
Do not start by buying a screen. Match each gap to the cheapest fix that closes it.
- News is slow: add a call signal and an escalation rule first. Andon does this with lights and a named responder.
- Line state is unknown: add a machine signal. Machine monitoring explains the basics, and connecting old machines covers equipment with no usable PLC.
- Causes are missing: shorten the reason list and capture it at the line, as in operator downtime logging.
- The numbers come too late: change the routine before the tool. See shop floor reporting for how to shorten the chain.
- People do not trust it: test a shift against a stopwatch and show the crew the result.
Where visibility stops paying
You do not need live data on every machine in the building. A slow manual station that never limits the line can wait.
Begin at the constraint, since an hour lost there is an hour lost for the whole line. The guide to bottleneck analysis helps you find it if you are not sure.
The other limit is attention. If no one will act on a signal within the hour, it does not need to reach them within the minute.
Mistakes when plants chase visibility
- Treating screens as the goal. A screen shows data, and it does not assign anyone to respond.
- Connecting everything at once. A long rollout delays the first useful answer by months.
- Skipping the reason. A state with no cause tells you there is a problem but not what to do.
- Using the data to police people. Crews who feel watched learn to work around the system, so tell them what it is for.
- Never checking accuracy. One false number in front of a crew can cost weeks of trust.
What to do this week
Run the stopwatch test for your most important line, then do the ten-minute sum on three shift sheets. You will end up with a ranked list of blind spots that took under an hour to make.
If the top item needs a system, use the guide to judging a production line monitoring system before you talk to any vendor.
The hidden minutes are the hardest to name. Where cameras are fitted, Fabrico shows video of a stop, streamed on request next to its machine data, and operators tag the cause at the line. See how Fabrico uses cameras for OEE.
What people ask next
What is shop floor visibility?
It is the ability to see what each line is doing, why, and how far it is from plan, soon enough to act and accurately enough to believe. It rests on machine data for time and counts and on people for causes. See what machine monitoring is for the signal layer.
How do you improve shop floor visibility without new software?
Start with a call light, an hourly target and actual board, and a short reason list at the line. Those three shrink the wait and the unknowns. The hour-by-hour board is the simplest starting point.
Does shop floor visibility need real-time data?
Only where someone can react in that time. A stop at the constraint deserves a signal within a minute or so, while a weekly trend needs none of that. Match the speed of the data to the speed of the decision.
What is the difference between visibility and monitoring?
Monitoring is a way to collect machine signals. Visibility is the result, which also needs causes, order context and someone who looks and acts. A plant can monitor every machine and still not know why output fell.
Which lines should get visibility first?
Start with the line that limits output or the one the plan misses most. Measure there first and prove that a decision changed because of the data. See bottleneck analysis for how to find it.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.