Build one dashboard per audience. The operator needs a line board: state, count against plan and the current stop. The supervisor needs the shift view with the top stops. The manager needs trends and a Pareto. Always show OEE with its three factors and the top losses, never OEE alone.
One dashboard per audience
A dashboard that tries to serve everyone serves no one. The operator on the line, the supervisor and the plant manager ask different questions.
Build one view for each.
| Audience | Question | Time window | Best layout |
|---|---|---|---|
| Operator | Are we on target this hour, and what is stopped now? | Now and this shift | Line board |
| Supervisor | How did the shift go and what hurt most? | Shift and day | Shift view with Pareto |
| Manager | Are we getting better, and where? | Week and month | Trend and comparison |
Example 1: the line board
This is the screen at the machine. It must be readable from a few meters, so it shows few things: state, count against plan, and the current stop with a timer.
Notice the third tile. It shows OEE and its three factors together.
A single OEE number says little. The factors say where to look. See availability, performance and quality.
Example 2: hour by hour
The hour-by-hour board compares output with plan for each hour of the shift. It is the best way to see when the day went wrong.
Example: the plan is 1,200 pieces an hour. Bars turn amber below 95% of plan and red below 80%.
Attainment: 8,510 / 9,600 = 88.6%
Biggest gap: 08:00, 560 pieces short of plan
The next step is to link the red hour to its stop. See hour-by-hour production board for how to run it.
Example 3: the downtime Pareto
A Pareto chart ranks stop causes by total minutes. It shows the few causes that cause most of the loss.
Always show the count of stops next to the minutes, as above. 58 minutes from 52 short misfeeds needs a different fix from one long breakdown. Read how to build a downtime Pareto chart.
Example 4: the weekly trend
For managers, the question is direction. A line chart of OEE by week, with the three factors beneath it, shows whether the work is paying off.
Always mark the changes you made, such as a new feeder or a new reason list, so the trend has a story.
Example: four weeks on one line, with the change marked in the week it happened.
| Week | A | P | Q | OEE | Note |
|---|---|---|---|---|---|
| 40 | 80% | 82% | 96% | 63.0% | |
| 41 | 81% | 83% | 96% | 64.5% | |
| 42 | 84% | 84% | 96% | 67.7% | New label feeder fitted |
| 43 | 86% | 86% | 97% | 71.7% |
The marked week explains the step up. Without the note, a manager would see 63% to 72% and not know which change to copy to the next line.
Weight weeks by scheduled time when you average them, as Overall Equipment Effectiveness (R. C. Hansen, 2001) advises, so a short week does not count the same as a full one.
Example 5: the shift view
The supervisor view lists each line with its OEE and factors, the top three stops of the shift, and the open actions from yesterday. It supports the daily meeting. See daily production meeting.
| Line | OEE | A / P / Q | Top stop | Open action |
|---|---|---|---|---|
| Line 1 | 71.8% | 88 / 85 / 96 | Label misfeed, 41 min | Reset label sensor gap |
| Line 2 | 69.7% | 79 / 90 / 98 | Changeover, 55 min | Changeover review Thursday |
| Line 3 | 69.2% | 91 / 80 / 95 | Slow running at the capper | Check capper speed setpoint |
Three lines with almost the same OEE, and three different problems. That is why the factors sit next to the number.
Example 6: the plant comparison
A table of lines side by side is tempting, and risky. A line that runs short, hard products looks worse than one that runs long, easy ones.
Compare each line with its own past, and with a target per product family, not with each other. See line OEE vs machine OEE.
Common mistakes
- OEE as a single big number. It hides which factor is weak. Always show A, P and Q.
- Averages across products. A plain average of lines or shifts misleads. Weight by scheduled time.
- Charts with no action. If a tile never changes a decision, remove it.
- Fancy 3D and gauges. They look good in a demo and read badly on the floor. Bars and numbers win.
- Stale data. A board that shows yesterday as if it were now is worse than no board.
- No owner for the board. Name one person who keeps it right and one who acts on it.
How to start
Start with the line board for one line. Ask the operators what they want to see.
State, count against plan, current stop, shift OEE with factors, top stop so far.
A target per product family. Use your own best accredited rate, not a guess. See ideal cycle time.
Review with the crew every day. Remove what nobody uses.
Then the shift view, then the weekly trend.
Design rules that hold up
- Show the target next to the number. A number with no target is trivia.
- Never rely on color alone. Add a word or an icon. Some people cannot tell red from green.
- Big text for the floor. If it cannot be read from the walkway, remove tiles until it can.
- Five tiles at most per screen. Each must lead to an action.
- Show losses, not just OEE. The top three stops are worth more than the third decimal of OEE. See the six big losses.
- Use one definition everywhere. If two screens give two OEE numbers for the same shift, nobody trusts either.
If you want to build your own, start with free OEE dashboard software or compare tools in best manufacturing KPI dashboard software.
Fabrico is a light MES for OEE. It shows OEE per line and shift, with stops, speed and counts live from the machines and the cause tagged by operators at the line. See OEE per line and shift in Fabrico.
What people ask next
What should an OEE dashboard show?
OEE with availability, performance and quality, the top stop causes, output against plan, and the current state. Each tile should lead to an action.
How many metrics are too many?
More than about five tiles on one screen. If the floor cannot read it from the walkway, there are too many. Move the rest to a second view.
Should I compare lines on one dashboard?
Carefully. Different products and speeds make raw comparison unfair. Compare each line with its own history and target. See how to average OEE across machines.
How often should the dashboard update?
Often enough to act in the shift. For a line board, every few seconds to a minute is typical. For a weekly trend, once a day is enough.
Can I build one in a spreadsheet?
Yes, for a start. The limit is data entry, which is late and rounded. See the OEE tracking template and the OEE tracker.
Related guides: Plant Manager OEE KPIs: What to See Each Week · OEE Dashboard in Power BI: Tables and DAX · OEE Dashboard in Grafana with SQL and Node-RED