Pareto chart in maintenance: find the 20% that stop the plant

In almost every plant, a handful of assets account for most of the downtime. The trouble is that, without looking at the numbers, the priority list follows memory: the machine that broke yesterday, the request from whoever complained the loudest. A Pareto chart swaps memory for history.

The idea comes from the economist Vilfredo Pareto and was brought into industrial quality by Joseph Juran: a few causes explain most of the effect. It's the famous 80/20 rule. In maintenance, it usually looks like this: about 20% of the assets or failure modes account for about 80% of the downtime hours. The exact split changes, but the concentration is almost always there.

Count downtime hours, not number of failures

The most common mistake is building the Pareto chart from the number of work orders. A conveyor that jams 22 times in six months, 40 minutes each time, looks like the worst asset in the plant. A compressor that failed 4 times, but stayed down 8 hours each time, barely shows up. For production, the compressor hurt a lot more.

Pick the measure based on what you want to reduce:

• Downtime hours: to tackle lost availability. It's the most useful measure to start with.
• Cost: to tackle spending on parts, contractors and overtime.
• Number of failures: to tackle short, frequent stops that annoy operations and hide lost production.

Example: six months of corrective work orders

A fictional beverage plant exported six months of corrective work: 55 failures on 23 assets, adding up to 120 downtime hours. Sorted by hours:

Equipment	Failures	Downtime hours	% of hours	Cumulative %
Line 2 filler	9	41	34.2%	34.2%
Compressor C-02	4	33	27.5%	61.7%
Conveyor E-05	22	18	15.0%	76.7%
Pump B-11	3	9	7.5%	84.2%
Boiler	2	6	5.0%	89.2%
Other 18 assets	15	13	10.8%	100%
Total	55	120

Three assets out of 23 explain 77% of the downtime. Notice that the conveyor, the leader in number of failures, drops to third when the measure is time. And the other 18 assets together add up to 13 hours: that's where it isn't worth spending energy right now.

The second level: inside the equipment

Knowing the filler is the problem still doesn't tell you what to do. Repeat the Pareto with only its 9 failures, now grouped by failure mode:

Failure mode (line 2 filler)	Failures	Downtime hours
Jam in the cap conveyor	4	19
Leak in the pneumatic system	2	16
Level sensor out of calibration	3	6

Now the target is concrete: the cap conveyor, with 19 of the 41 hours. This is where a root cause analysis comes in. The Pareto chart tells you where to look; it doesn't tell you why.

Mistakes that spoil the Pareto chart

• A huge "other" bar. If "other" is the biggest bar on the chart, the categories are poorly defined or the work order doesn't record the equipment. See what to record on a work order.
• Too many categories. Thirty failure modes with one occurrence each show nothing. Group them by family: bearing, seal, electrical, pneumatic, sensor.
• A short period. One month catches an isolated event and makes it look like a trend. Three to six months give a fairer picture.
• Mixing types of stops. Setup and preventive work aren't failures. If they go into a corrective Pareto, they distort the result.
• Doing it once and filing it away. Once you fix the top item, the second one becomes the top. Redo the chart every month and check whether your actions brought the bars down.

Pareto and "sick" equipment

A useful complement is tracking equipment with repeat failures over a short period, for example 3 or more failures in 30 days. The Pareto chart shows the six-month total; the repeat list shows what's getting worse right now. The % sick units calculator does this math for the whole fleet.

In taskfield, every corrective job is recorded with the equipment, downtime and failure mode. A Pareto by hours, cost or count comes straight out of the history in a few clicks, with no spreadsheet export.