A hydraulic hose bursts on the header of your combine for the third time this harvest. Replacing it each time does not tell you why it keeps bursting. On the example farm in this article, that hose alone accounts for 48 of the 100 hours of stoppage in one harvest (the hours are made up for the example).
Add up the hours your machines were down for each cause. Take the cause with the most hours and ask why about it up to five times, until you reach an answer you can change in how the farm works and check later. This is the five whys. Keep the record on a cause sheet, one page per failure. It holds the chain of whys, proof beside each answer (a part you kept, a photo, an invoice), the change and the date to add up the hours again.
The only machine case in the sources listed at the end comes from sugarcane, not from grain. It is the stoppage log of one sugarcane harvester over 107 days in Brazil. Three parts of that machine stopped it for 66.38, 29.70 and 27.35 hours. The first was the elevator that lifts the cut cane out of the machine. The second was the undercarriage it rides on (the tracks or wheels and their frame). The third was the base cutter that cuts the stalk at the ground. The two longest lines in the log, 112 and 75.52 hours, tie their hours to no specific part, so no why could start from them.
For the hose, the chain runs like this. The hose rubbed against the frame, because a clip was missing, because the clip was not refitted after the off-season service, because the service checklist has no step for it. The checklist is something you can change and check later, so the chain stops there. The full chain, with its proof, is shown further down.
This is not a promise of fewer hours of stoppage. The method comes from factories and is used in health care, and none of the sources tests it on a grain farm. The section on evidence below says what they do contain.
Which failure should get a sheet first?
The one with the most hours of stoppage in your own log, the notes of each stop with its date, hours and cause. Work on one failure at a time, so that the first sheet is finished before the next begins. Count hours, not stops, because a stop costs the time the machine is down.
Hours do not all weigh the same: an hour lost at the peak of planting or harvest costs more than one in the off-season. If that matters on your farm, also add up the peak weeks alone and see whether the order of the causes changes; keep the full count as the starting point.
A Pareto chart lists the causes from the most hours to the fewest. Each has its share of the total and a running total (its hours added to those of the causes above it). The Institute for Healthcare Improvement (IHI) is a US non-profit that publishes improvement guides for health care. It describes the Pareto principle as roughly 80 percent of an effect coming from 20 percent of the causes. It calls the causes at the top the “vital few”, the ones that matter most. That split is a pattern and not a law, so follow your own count and not the rule.
An undergraduate thesis by Pizzeco at the Federal University of Uberlândia, in Brazil, used the count on one sugarcane harvester at one mill. The harvester worked 107 days, from 19 July to 3 November 2018, against a target of availability above 85 percent (availability is the share of the time the machine is ready to work). Its stoppage log had two catch-all lines, “miscellaneous”, 112 hours, and “preventive service” (planned work, not a failure), 75.52 hours, and neither ties its hours to a specific part. Pizzeco set both aside. Three parts were left: the elevator with 66.38 hours, the undercarriage with 29.70 and the base cutter with 27.35.
A log line that names no part can hold the most hours and still give you nothing to ask why about. Pizzeco suggests that the work order carry the part and the hour-meter reading at the moment of the stop, and found stops with no hour-meter reading written down. The article on the factory service interval sets that reading against the manufacturer’s schedule.
On the example farm, one soybean harvest’s log adds up to 100 hours of stoppage. The hours are made up for the example.
| Cause | Hours stopped | Share of total | Running total |
|---|---|---|---|
| Hydraulic hose bursts on the header | 48 | 48% | 48% |
| Header drive-chain slips | 24 | 24% | 72% |
| Tire punctures | 12 | 12% | 84% |
| Sensor faults | 10 | 10% | 94% |
| Other, each under 4 hours | 6 | 6% | 100% |
The hose line gets the first sheet and the drive-chain line is next. The count needs only a log that gives the hours and the part for each stop. If yours does not, the section on what goes on the cause sheet, further down, shows what to record from the next stop on.
How do you ask why five times?
Write the failure in one precise sentence, then ask why. Serrat’s note for the Asian Development Bank, an international development bank, gives the steps. State the problem and agree on the statement with the people involved, write the answers to the first why and ask why again about each answer. Then look for causes in the way the work is set up, not in one person, and decide what to change. Serrat says Sakichi Toyoda developed the technique for Toyota Industries Corporation.
Five is “a good rule of thumb” in that note, and not a target. You may need fewer whys, or more. The chains in the thesis run from three to five answers.
Serrat also encourages the people involved to verify each answer on the spot before the next why. On the cause sheet you turn that into a column: beside every answer, write what proves it true, such as a part you kept, a photo, an invoice or an hour-meter reading. A reading is only as good as the meter, so if the meter and the fuel bought do not match, start with that gap. An answer with nothing beside it is a guess, and the next why stands on it.
The chain for the hose line has five answers, each with its proof, and all of it is made up for the example. The first four describe what happened to the hose and its clip. The fifth, a step missing from the service checklist (the list of steps the workshop follows at each service), is something the farm can change and later check.
Where does the chain stop?
At an answer the farm can change and check. IHI’s own tip for its cause-and-effect diagram is the same test: keep asking why until the cause is specific enough to test a change and measure its effects. IHI’s page on the five whys adds that what looks like the cause is sometimes just another symptom.
Two kinds of answer fail the test. One is a person’s name, or a word like careless. The RCA2 guide is a 2016 manual on finding why harm happened in hospitals, written by the National Patient Safety Foundation, a US patient-safety group. It says a human error always has a cause before it, and so does a procedure that was not followed. So neither is a root cause, the cause at the bottom of the chain. The other kind is a vague word such as poor, inadequate or wrong, which RCA2 tells analysts to replace with what actually happened. RCA2 was written for hospitals, so this article borrows only its rules for wording an answer.
The thesis ran five whys on the wear of the base-cutter blades, shown below. The blades rub the soil. The operators avoid the height regulator, the control that sets how high the base cutter rides, because the machine sways as it follows the ground. The chain ends on the base-cutter pressure (how hard the cutter is pressed down), which is poorly calibrated when the regulator is used.
The author concludes that the operators lacked information on how to calibrate that pressure. But the last answer says only “poor”, a word of the kind RCA2 says to replace. The question under the figure, which pressure and who sets it, is this article’s own and not the thesis’s. On your own sheet you would write the answer to it.
The thesis also ran five whys on the elevator, the part that lifts the cut cane out of the machine. In two of those chains the last answer is that the mill’s service truck (the workshop on wheels that goes out to the machines) lubricated the part at irregular times. The author’s own reading of those chains is that the mill had set no fixed lubrication interval. That is the kind of answer the RCA2 rule above asks for: the irregular lubrication is not blamed on the people, because a missing interval stands behind it.
What does the fishbone add?
A chain follows one answer downward. The fishbone (also called the Ishikawa diagram after its creator) is a diagram that lays out the branches of causes side by side. IHI describes it as a diagram with the effect in a box at one end and the causes grouped under Materials, Methods, Equipment, Environment and People, in branches that resemble the bones of a fish.
On the cause sheet you use it as a check and not as a drawing. Label each answer of the chain with one of the five (Equipment means the machines and tools), and see which nobody has asked about. On the example farm the chain covers Equipment and Methods, and three branches are still open. RCA2 asks teams to find several contributing factors and not to stop when one turns up.
IHI calls these five categories the classic set, the usual starting point, and lets a team choose five or six that fit its problem. Open a second sheet for a branch when the answer you found there is itself a failure that repeats in your log, or when the hours behind it justify the work.
How much evidence stands behind five whys?
The method is described in a note from a development bank, in health-care improvement guides and in a patient-safety guide. The only study on a farm machine among the sources listed at the end is the undergraduate thesis on one sugarcane harvester. It measured the stoppage hours and proposed a maintenance plan, and did not measure whether the hours fell afterwards. The sources hold no test on a grain farm and no comparison of farms that keep the sheet with farms that do not. That does not mean the method fails. It means nobody has measured it on a farm like yours.
Serrat lists the criticisms. Whoever investigates stops at symptoms, cannot go beyond what they already know and gets little help in asking the right questions. Different groups of people also reach different causes for the same problem. IHI’s page says the same of people who see different parts of the work. For that reason, whoever fills in the sheet writes proof beside each answer, so that the next person can check it.
What goes on the cause sheet?
One page per failure, with the seven fields in the table. Copy them onto a blank page. Start from the stoppage log of the last harvest. Use the dealer’s repair history for each machine to fill what the log lacks, such as the part or the date (the article on what repairs cost per machine shows where to ask). If the log names no part on its lines, you cannot start a sheet yet. For the next stops, write the part and the hour-meter reading on every line, and start the first sheet when you have enough stops to add up.
| Field | What you write |
|---|---|
| Failure | The failure in one precise sentence, the machine and the dates |
| Hours stopped | The hours from your log and their share of the total |
| Whys | Up to five lines: each answer, with its proof beside it |
| Fishbone check | Materials, Methods, Equipment, Environment, People: answered or not asked yet |
| Cause to change | The last answer, the one the farm can change and later check |
| Action | The change, who makes it and by what date |
| Count again | The date when you will count the hours stopped by this cause again |
The figure below is the same page filled in for the hose line.
The same page takes any failure that repeats, not only a stopped machine. Harvest loss above your limit also repeats. Split the loss between the header and the inside of the machine, as the article on the harvest-loss limit shows, and start the chain with the larger part. Plan about two hours for the first sheet.
The cause you end on is usually a missing step or a task nobody owns, and the change is to add the step, name who does it and count again. That is farm management at work; see farm management, and the production articles for other production records. Related: Where is each tool and part kept in your farm shop?