PlantOps BI

Downtime Pareto

Across the three demo plants, 5 of 7 unplanned reasons account for 85.7% of unplanned downtime minutes. Pick a plant to see its own order.

All plants

834,560 unplanned minutes, plus 372,674 planned (changeover, maintenance, meetings) left out of the Pareto. The marked rows reach 80% of the minutes.

ReasonMinutesBarShareCumulative
EQP Equipment failure 219,848 26.3% 26.3%
MAT Material shortage 208,579 25.0% 51.3%
QAH Quality hold 112,940 13.5% 64.9%
OPR Operator unavailable 89,745 10.8% 75.6%
OTH Other 84,379 10.1% 85.7%
UPS Upstream starvation 80,504 9.6% 95.4%
UTL Power / utilities 38,565 4.6% 100.0%

Portland Plant

344,558 unplanned minutes, plus 152,076 planned (changeover, maintenance, meetings) left out of the Pareto. The marked rows reach 80% of the minutes.

ReasonMinutesBarShareCumulative
MAT Material shortage 89,893 26.1% 26.1%
EQP Equipment failure 79,496 23.1% 49.2%
QAH Quality hold 48,320 14.0% 63.2%
OPR Operator unavailable 37,910 11.0% 74.2%
OTH Other 37,087 10.8% 85.0%
UPS Upstream starvation 35,686 10.4% 95.3%
UTL Power / utilities 16,166 4.7% 100.0%

Boise Plant

273,787 unplanned minutes, plus 125,629 planned (changeover, maintenance, meetings) left out of the Pareto. The marked rows reach 80% of the minutes.

ReasonMinutesBarShareCumulative
EQP Equipment failure 88,527 32.3% 32.3%
MAT Material shortage 63,432 23.2% 55.5%
QAH Quality hold 32,901 12.0% 67.5%
OPR Operator unavailable 26,723 9.8% 77.3%
OTH Other 25,910 9.5% 86.7%
UPS Upstream starvation 24,850 9.1% 95.8%
UTL Power / utilities 11,444 4.2% 100.0%

Reno Plant

216,215 unplanned minutes, plus 94,969 planned (changeover, maintenance, meetings) left out of the Pareto. The marked rows reach 80% of the minutes.

ReasonMinutesBarShareCumulative
MAT Material shortage 55,254 25.6% 25.6%
EQP Equipment failure 51,825 24.0% 49.5%
QAH Quality hold 31,719 14.7% 64.2%
OPR Operator unavailable 25,112 11.6% 75.8%
OTH Other 21,382 9.9% 85.7%
UPS Upstream starvation 19,968 9.2% 94.9%
UTL Power / utilities 10,955 5.1% 100.0%

What to Read Into It

The top of the list is where a maintenance and materials plan starts. Just as telling is "Other" in 5th place with 10.1% of unplanned minutes. A catch-all that big means operators can't find a code that fits, and the minutes inside it are invisible to the Pareto. Splitting it is often the cheapest improvement on the list.

Every downtime minute in the data carries a reason, so the bars add up to the plant's real downtime. In live data, check that first: hours with minutes and no reason fall out of a reason Pareto silently.

The DAX

A cumulative share needs each reason's minutes compared against all the others in the same filter context. The measure builds that list once with ADDCOLUMNS, then sums every reason at least as large as the current one.

Downtime Minutes =
SUM ( fact_production_hourly[downtime_minutes] )

Unplanned Downtime Minutes =
CALCULATE (
    [Downtime Minutes],
    KEEPFILTERS ( dim_downtime_reason[category] = "Unplanned" )
)

-- For a visual with one row per reason, on any reason column (code or
-- description). Slicers on plant and date still apply, because ALLSELECTED
-- clears only the rows the visual itself is iterating.
Cumulative Unplanned % =
VAR ThisReason = [Unplanned Downtime Minutes]
VAR Reasons =
    ADDCOLUMNS (
        ALLSELECTED ( dim_downtime_reason ),
        "@Minutes", [Unplanned Downtime Minutes]
    )
VAR Running =
    SUMX ( FILTER ( Reasons, [@Minutes] >= ThisReason ), [@Minutes] )
RETURN
    IF (
        NOT ISBLANK ( ThisReason ),
        DIVIDE ( Running, SUMX ( Reasons, [@Minutes] ) )
    )

Three details matter. ALLSELECTED takes the whole reason table, not one column, so the measure works whether the axis shows the code or the description; on a single column, a filter from the other column would survive and every row would read 100%. KEEPFILTERS makes a filter on category set to Planned return blank instead of being replaced by Unplanned. And comparing with >= means two reasons with the same minutes get the same cumulative share, so the result does not depend on how the visual happens to sort ties. The tables above use the same rule, and the site's tests check it.

Unplanned downtime KPI · OEE across lines · About the demo dataset