Daily Production Board

What this tool does

CM8-247 is an hour-by-hour production board. One row per period, per line: what the period was planned to produce, what it actually produced, how many failed, how many minutes were lost, why, and what was done about it inside the period. From that it works out attainment, variance, first-pass quality and a downtime Pareto, and prints a board you can hand over at the end of the shift. Everything runs inside this single file — no account, no upload, no network request of any kind.

A board is a trigger, not a report

This is the whole idea, and almost every failed board fails on it. The value of an hourly board is not the total at the bottom; that total is available anyway, from the count at shift end. What the board adds is time: it makes a shortfall visible while there are still hours left in which to act on it.

An hour that finishes thirty units short is a question asked at ten past the hour, by the person who was standing there, to somebody who can act. If that conversation happens the board has done its job, even if the row is never read again. If it does not, the row is decoration.

Planning the periods honestly

The plan quantity is where most boards quietly go wrong. A flat rate applied to every hour of the shift — 110, 110, 110, all day — is a fiction whenever the shift contains something that is not steady running: a changeover, a scheduled meeting, planned maintenance, a shortened first or last hour.

The damage is behavioural, not arithmetic. Everyone on the line knows the first hour cannot make 110 because there is a twenty-minute changeover in it. When the board shows red for a reason everybody already understood and nobody could have avoided, it stops meaning anything — and the red that matters, three hours later, gets the same shrug. A flat plan teaches people to ignore the board.

Plan each period for what it can actually do: an hour carrying a twenty-minute changeover at a rate of 110 an hour is planned at about 73. Take a scheduled meeting out of the plan and record it as a planned stop when it happens. A period with no production planned is legitimate — set the plan to zero and the tool shows a dash. The test is simple: on a good day with no problems, every period should land at or very near 100 %. If your good days come out at 78 %, the plan is wrong, not the line.

Recording the loss in the period

Fill the row in as the period ends, not at the end of the shift. Memory at shift end is fiction: minutes get rounded to comfortable numbers, two separate stops merge into one, the reason drifts towards whichever category is least awkward to write down, and the small losses — the four short stops on the screw feeder — vanish because nobody remembers them. What you are left with is a plausible story that happens to be wrong, and every Pareto and improvement decision built on it inherits the error.

Loss reason discipline

This tool offers eight reasons plus "none", and that is deliberate. With forty categories, two people watching the same stoppage file it differently, the counts fragment across near-duplicates, and nothing ever reaches the top of the Pareto. Fewer, clearer categories used consistently beat a long list used approximately.

• Breakdown — the equipment failed and had to be repaired.
• Changeover / setup — last good part of one product to first good part of the next.
• Material shortage — nothing to run, or the wrong thing delivered.
• Quality issue / rework — production stopped or slowed to deal with defects.
• Labour — absence, an unplanned break, waiting for an operator or a skill.
• Running slow — the line ran, but under rate: micro-stops, jams cleared.
• Planned stop — a meeting, planned maintenance, a scheduled clean.
• Other — nothing else fits. If "other" tops your Pareto, your categories need a conversation, not a new option.

Where a period had two losses, put the larger one in the reason and describe both in the detail line. If that is common, your periods are too long. The tool refuses two combinations: a period that produced something, came in under plan and claims no loss reason — the units went somewhere — and a loss reason with no downtime and no shortfall, which either had an effect worth recording or was not a loss.

The escalation trigger

Escalation only works if it is defined in advance. "Escalate if it looks bad" leaves the judgement with the operator, and on a busy line nothing gets raised until it is far too late. Write the rule down and put it on the board. The tool ships with the two-part rule most lines converge on, both halves adjustable in Settings:

• Any single stop longer than 15 minutes — long enough that the period cannot be recovered without help.
• Two consecutive periods below plan — one bad period is noise; two in a row is a trend, and the shift will not catch up on its own.

The escalation tile compares what your rule would have flagged against what was actually ticked, and reports how many periods met the trigger and were not raised — the clearest measure of whether the board is used or merely completed. Escalating is not an accusation: it passes a problem to somebody with the authority to solve it.

Countermeasures written at the time

The countermeasure records what was actually done, by whom and when, while the period was still running or immediately after. "Maintenance called 11:08, belt changed, running 11:43, spare now held at the line" is a countermeasure. "Improve maintenance" is a wish. Put times in it, and separate what recovered the period from what was changed so it does not recur. A board full of recoveries and empty of changes explains why the same reason tops the Pareto month after month.

Reading the board

The board chart shows where the day was lost. Each bar carries good units, then rejects, then the gap to plan, so any bar with a visible top block is a period that did not deliver. Two large blocks and six clean bars is a day with two events to investigate; eight medium blocks is a plan that is wrong, or a line that cannot hold rate.

The attainment line shows the day's shape against your target. Dips that recover are a line being managed; a step down that never comes back is the signature of an unescalated problem.

The downtime Pareto is your improvement backlog, in order. Two or three reasons usually account for most of the minutes, and fixing those removes more lost output than working evenly down the list. Filter to a single date and it becomes that day's Pareto; leave the filter open and it becomes the list worth spending money on.

The loss analysis table adds the units below plan attributed to each reason — the sum of the shortfalls on the rows carrying it. It does not net off periods that ran ahead, so it will normally exceed the day's net variance. That is deliberate: recovery later is real, but it does not undo the loss, and pretending it does hides the event you wanted to fix.

The formulas

Attainment is shown per period and again for the whole day, where it is the day's total actual divided by the day's total plan — not the average of the period percentages, which would weight a period planned at ten units the same as one planned at two hundred. Where the plan is zero the tool prints a dash.

First-pass quality counts rejects against what was produced, not against what was planned: a unit reworked successfully still counts as a reject, because the rework cost was real.

Where this sits alongside other boards

Three tools in this range look alike and answer different questions. This one is the hour-by-hour plan-versus-actual board: rate against plan within the shift, so a shortfall is caught while the shift is still running. The SQCDP daily performance board records one colour per letter per day for the tier meeting — a day-level status across safety, quality, cost, delivery and people, not a rate. The machine downtime and loss log records individual stoppages against a machine with start and end times and produces reliability figures. Use this board on the line during the shift, the SQCDP board at the morning meeting, the downtime log for one machine's history.

FAQ

Do the periods have to be hours? No — clock hours, "Hour 1" to "Hour 8", or two-hour blocks on a slow-cycle line. Keep the wording identical every day so the periods line up, and remember that downtime minutes are only comparable between periods of the same length.

Should rejects come out of the actual quantity? No. Enter everything the period produced as the actual and the failures separately as rejects. Netting them off by hand hides the quality loss inside the output figure.

Is this OEE? No. Overall equipment effectiveness multiplies availability, performance and quality against a theoretical rate and needs loading time and ideal cycle time defined. This board compares actual output with your plan — a different and usually more useful question on the shop floor — but do not present attainment as OEE; they are not the same measure and will not agree.

Saving your work

Board rows, settings and the report header are written to this browser's local storage as you type. That storage belongs to one browser on one computer: another browser, a private window, a second machine or a clean-up tool that clears site data will not have it.

Treat Export .json as the real save — one file containing everything, which Import .json restores anywhere. Export CSV gives you every filtered row for spreadsheet work. Reset asks twice, then erases everything. There is no undo.

Accuracy & disclaimer

The arithmetic here is deliberately simple and the tool does it faithfully. Everything that matters sits underneath it: whether the plan was honest, whether the count at the end of the period was accurate, whether the reason chosen was the real one, and whether any of it was written down while it was still true.

A board filled in at the end of the shift measures memory, not production. The totals still add up, the charts still draw and the Pareto still ranks — which is exactly what makes a reconstructed board dangerous. This is a shop-floor record-keeping and prioritisation aid, not a measurement system, not an OEE calculation, and not a substitute for your system of record for what was made.