You Cannot Fix Everything at Once
A production line stops for dozens of reasons. A plant that tries to work on all of them makes little progress on any. A plant that works on the loudest one, usually whatever went wrong yesterday, makes progress on the wrong thing. The Pareto chart is a simple tool for finding the few causes that account for most of the loss, so that limited engineering and maintenance time goes where it will return the most.
This guide explains what a Pareto chart is, how to build one from the downtime data you already have, how to read it and the mistakes that make it misleading.
What a Pareto Chart Is
A Pareto chart is a bar chart in which the categories are sorted from largest to smallest, with a line added to show the running cumulative percentage. The bars show how much each cause contributes. The line shows how quickly the causes add up to the total.
It is named after the Italian economist Vilfredo Pareto, who observed that a large share of wealth was held by a small share of the population. The quality pioneer Joseph Juran later applied the idea to manufacturing problems and described the vital few and the trivial many. The popular shorthand is the 80/20 rule: roughly 80 per cent of the effect comes from roughly 20 per cent of the causes. The ratio is not a law. The pattern, a steep drop from a few large causes to a long tail of small ones, turns up in most downtime data.
Building One for Downtime
Step one: collect the data. You need a list of stoppages over a period, each with a cause and a duration. A month of data is usually enough to start. It can come from a line monitoring system, a maintenance system or a shift log filled in by hand.
Step two: group by cause. Give each stoppage a category. Use specific categories. Labeller jam is useful. Mechanical fault is not.
Step three: total each category. Add up the minutes lost for each cause.
Step four: sort. Put the largest first.
Step five: calculate cumulative percentages. For each cause, work out its share of the total and add it to the shares above it.
Step six: draw it. Bars for the minutes, a line for the cumulative percentage. Any spreadsheet will do this.
A Worked Example
A canning line lost 2,000 minutes to unplanned stops in a month. Sorted by cause, the losses were as follows.
Labeller jams: 620 minutes, which is 31 per cent of the total. Filler valve faults: 480 minutes, 24 per cent, bringing the cumulative figure to 55 per cent. Changeover overruns: 410 minutes, 20.5 per cent, cumulative 75.5 per cent. Conveyor sensor faults: 190 minutes, 9.5 per cent, cumulative 85 per cent. Capper torque faults: 120 minutes, 6 per cent, cumulative 91 per cent. Palletiser faults: 90 minutes, 4.5 per cent, cumulative 95.5 per cent. All other causes: 90 minutes, 4.5 per cent.
Three causes account for just over three quarters of the lost time. If the plant could halve only those three, it would recover 755 minutes a month, more than it would gain by eliminating every other cause on the list entirely.
That is the message of a Pareto chart. The capper and the palletiser may be the faults people complain about, but they are not where the hours are.
Count Stops as Well as Minutes
A chart of minutes tells you where the time went. A chart of the number of stops can tell a different story, and it is worth building both.
In the same month, suppose the conveyor sensor faults occurred 140 times, at a little over a minute each, while the filler valve faults occurred 12 times at 40 minutes each. By duration the valve is the bigger problem. By frequency the sensor is, and frequent short stops do damage that minutes alone do not show. They break the line’s rhythm, cause product to back up, wear on operators and are often under recorded because each one seems too small to log.
The two charts point to different kinds of fix. Long, infrequent stops usually call for better parts, better spares holding or faster diagnosis. Short, frequent stops usually call for an engineering change at the source.
Drill Down
The first chart tells you where to look. It rarely tells you what to do. The next step is to take the largest bar and build a second Pareto inside it.
Labeller jams at 620 minutes is a category, not a cause. Break it down by label type, by product, by shift or by which station on the machine jammed. You may find that most of the jams occur on one label size, or in the first hour after a changeover, or on one supplier’s reels. Each level narrows the problem until it is specific enough for root cause analysis. A Pareto chart selects the problem. Techniques such as the 5 Whys and the fishbone diagram then find the cause.
Six Mistakes That Mislead
A large other category. If other or unknown is one of your biggest bars, the data is not yet good enough to act on. Fix the recording first.
Categories that are too broad. Electrical, mechanical and operator are departments, not causes. They start arguments and identify nothing.
Mixing planned and unplanned time. Changeovers and cleaning are planned. Keep them separate from breakdowns, or count only the overrun beyond the planned time, as in the example above.
Too short a period. One bad day can dominate a week of data. Use long enough to smooth out the exceptional event, or note it.
Measuring only time. Some stoppages cost far more than their duration suggests because they scrap product or force a clean. Where that matters, build the chart on cost.
Making the chart and stopping. A Pareto on the wall that never changes is decoration. Its value is in choosing a problem, fixing it and redrawing the chart.
Getting Reliable Data
The chart is only as good as what goes into it. Manual logs tend to capture the long stops and miss the short ones, and the cause recorded is often the operator’s best guess. That is still enough to start. As the plant matures, automatic capture of stop times from the line’s control system removes the guesswork about duration, leaving operators to select a reason from a short list.
Keep the reason list short and specific, review it every few months, and add a category whenever other starts to grow.
Using It Month After Month
Review the chart with production and maintenance together. Choose the top one or two causes and assign each an owner. Investigate, fix and record what was done. Redraw the chart the following month. If the fix worked, the bar shrinks, a different cause moves to the top and the process repeats. Over a year this routine does more for line performance than most capital projects.
Where BevTech Fits
BevTech is a family owned engineering business in Richlands, Brisbane, with more than 25 years in beverage and FMCG manufacturing. Plants usually call us once the chart has shown them where the hours are going and the fix needs engineering: a control change to stop a recurring fault, a machined part to end a jam, a sensor relocated out of the washdown, or a line integration that stops product backing up. Our industrial electricians, automation engineers, mechanical trades and machine shop work as one team on Allen Bradley and Siemens controlled lines.
If your top downtime cause has been the same for too long, use the contact form at bevtech.com.au/contact.






