A shop-floor dashboard is designed well when a worker standing thirty feet away can find its most important number and tell whether the line is winning or losing in about three seconds. You get there with size, hierarchy, and disciplined color, not by adding more data. A board that has to be studied is a report in the wrong place.
This is a design problem, not a data problem. Deciding which numbers a plant should track is covered in building a production dashboard; this post is about the other half, making the board readable from across a noisy floor by a person who is walking, standing, and glancing, not sitting and reading. Get the design wrong and the best metrics in the world go unseen.
What makes a shop-floor dashboard different from a report?
The viewer, the distance, and the time budget. A report is read by one person at a desk, up close, with minutes to spend. A floor board is read by dozens of people in motion, from ten to forty feet away, with about three seconds of attention before the next task pulls them back. Everything about the design follows from that: the numbers must be large, the layout must have an obvious top-to-bottom priority, and color must carry meaning a person can decode without stopping to think.
The test to hold every board to is simple. Stand where the operators actually stand, not at the screen, at the workstation, and give yourself three seconds. If you cannot answer "are we ahead or behind right now?" in that time, the board has failed, no matter how accurate the data behind it is. This is the discipline of visual management: information that works at a glance, or it does not work.
How big does the text need to be?
Use one inch of character height for every ten feet of viewing distance as a floor, and go larger for the number that matters most. That rule of thumb traces back to legibility research the National Bureau of Standards (now NIST) published on letter size as a function of viewing distance, and it has held up for decades of signage and display work. A board read from thirty feet needs three-inch digits on its hero metric, bigger than most people guess, and the single most common reason floor boards fail.
Three details make characters actually legible at distance, not just large. Stroke thickness should sit around 10 to 20 percent of character height, so numbers do not go spindly on a bright screen. Line spacing should run about 135 to 170 percent of character height, so rows do not smear together. And contrast has to be high, dark figures on light, or light on dark, because a floor is a glare environment with skylights, wash-downs, and moving equipment between the eye and the screen.
| Viewing distance | Minimum character height | Suggested hero-metric height |
|---|---|---|
| 10 ft | 1 in | 2 in |
| 20 ft | 2 in | 3–4 in |
| 30 ft | 3 in | 5–6 in |
| 40 ft | 4 in | 6–8 in |
| 50 ft | 5 in | 8–10 in |
What belongs on the board, and what doesn't?
One hero metric, a thin band of context, and a short exception list, in that order, top to bottom. The hero is the one number the shift is judged on, usually pace to plan or shift OEE. Below it sits just enough context to act: units versus plan, the current top loss, the next changeover. At the bottom, the exceptions, the two or three things going wrong right now that someone can do something about. Everything else belongs in the report the supervisor reads at the desk, not on the wall.
The failure mode is the "airplane cockpit" board, twenty tiles, six charts, and a scrolling ticker, all the same size. When everything is emphasized, nothing is, and the viewer's eye has no path to follow. Cutting a crowded board down to one hero and three supporting items almost always makes it more useful, because it finally answers the only question a passing operator is asking: are we okay, and if not, what is hurting us?
How should you design the board, step by step?
Build it in priority order, and test it from the floor before you hang it:
- Name the one question. Decide the single thing the board must answer at a glance, almost always "are we ahead or behind plan right now?" Everything else is secondary by definition.
- Pick the hero metric. Choose the one number that answers that question and commit to making it the biggest thing on the screen. Pair it with its target so "72%" instantly reads as behind, not just as a number.
- Set the size from the distance. Measure how far away people actually stand, apply the inch-per-ten-feet rule, and size the hero above that minimum. Do not trust how it looks on your laptop.
- Add only the context that drives action. Units to plan, current top loss, next changeover. If a tile does not change what someone does in the next hour, it does not belong.
- Give color one job. Reserve your alert color for "needs attention" and use it nowhere else. A board where five things are red on a normal shift has trained everyone to ignore red.
- Walk it back and test. Stand at the workstation, glance for three seconds, and see if you can call the state of the line. Fix whatever you could not read. Re-test every time you add a tile.
Two standards worth designing to:
- Legibility research on letter size versus viewing distance, the basis for the inch-per-ten-feet rule of thumb, was published by the U.S. National Bureau of Standards (now the National Institute of Standards and Technology). The technical note "Size of letters required for visibility as a function of viewing distance" is the primary source behind the sizing tables signage vendors still use.
- Safety-critical color and format on the floor follow the ANSI Z535 family of standards, which define the meanings of safety colors and the layout of signs and labels. The current safety-colors standard, ANSI Z535.1-2022, published through the ANSI webstore with the National Electrical Manufacturers Association (NEMA) as secretariat, defines those color meanings. Aligning your board's color logic with the color meanings workers already know reduces the decoding time to near zero.
How should you use color?
Give every color exactly one meaning, and use your alert color sparingly. The strongest boards run mostly neutral, dark text on light, or light on dark, and reserve a single accent for "this needs attention right now." When that accent shows up, it means one thing and the eye jumps to it. The moment you use the same color for a heading, a logo, and an alarm, the alarm stops working.
Most plants map to the traffic-light convention workers already carry in their heads: green for on-target, red for off-target, amber for at-risk. That is fine, just keep the discipline that a color's presence always means the same thing, and that a normal shift is mostly calm, not a wall of red. If your board lights up red when the line is running fine, the thresholds are wrong, and you have taught the floor that the board cries wolf.
How does the board turn into action?
A board only earns its wall space if it changes behavior, and the design decides whether it does. The hero metric feeds the daily management huddle; the exception list is the agenda. When a loss crosses a threshold, the board should tie into the plant's andon response so the right person is pulled in, not just informed after the fact. The board is the trigger; the huddle and the andon are the response.
Behind the glass, the numbers have to be live and honest, or the design does not matter. A beautifully laid-out board fed by yesterday's spreadsheet trains people to ignore it within a week. This is where the design connects back to the data: pull the hero and the losses from real-time signals off the line the way Harmony's manufacturing analytics does (see the platform), and tie the metric definitions to your broader manufacturing KPIs so the wall, the huddle, and the monthly review all tell the same story. You can price what closing the top loss on the board is worth with the OEE calculator. A dashboard that is readable, honest, and wired to a response is a management tool. Everything short of that is wall art.