A downtime reason code is a short label that says why a line stopped. A good list has a few top-level groups that split planned from unplanned and machine from people from materials, and a short list of causes under each. Every minute gets exactly one code, and there is never a code called "other".
What a reason code list is for
A stop with no reason is only a gap in the clock. A stop with a reason can be counted, ranked and fixed.
The codes are the base of every downtime report. They feed the downtime Pareto chart, the OEE loss tree and the weekly review.
Bad codes give bad charts. A list with 80 choices gets random picks. A list with three choices tells you nothing.
The codes also map to Nakajima's six big losses, so your numbers can be read next to any OEE material. For how to collect the stops, see how to track downtime.
A code tree you can start from
The groups below follow the classification in Overall Equipment Effectiveness (R. C. Hansen, 2001). The causes under each group are examples. Adapt them to your line.
STO, STI, DTT, DTO, DTQ, S and W are the loss codes used in Overall Equipment Effectiveness (R. C. Hansen, 2001). Excluded time is not a loss code, but every minute on the calendar must still land somewhere.
Speed loss and waste are not stops. They are measured from counts, not logged by hand. See speed losses.
How to pick a code in three questions
Operators should not read a manual. Give them a path they can follow in a few seconds.
- 1Was it planned? If the stop was planned as part of normal running (changeover, test, planned loading), use an STO code.
- 2Did something outside the machine cause it? No material, no people, no information, a blocked or starved line. Use an STI code.
- 3If not, it is unplanned downtime. Equipment failure is DTT. People or procedure is DTO. Material or process quality is DTQ.
If the answer to all three is "I don't know", the code is "cause not yet known, call the mechanic". Do not make "other" available.
Rules for a good list
- No "other" and no "miscellaneous". Overall Equipment Effectiveness (R. C. Hansen, 2001) says to categorize every minute. A bucket for the unknown becomes the biggest bar in every chart.
- Two levels at most. About six groups at the top, a short list under each. Keep it detailed enough to guide priorities, but not so detailed that people drown in data.
- One code per minute. If a changeover is followed by a jam, that is two stops with two codes.
- Name what the operator can see. "Labeller jam" works. "Mechanical root cause" does not.
- Separate planned from unplanned. The two have different owners and different fixes. See planned vs unplanned downtime.
- Each code has an owner. Someone who gets the Pareto bar and fixes it.
Worked example
Example: a packaging line has 960 scheduled minutes in a day. The operators log 200 minutes of stops with the codes above.
80 + 50 + 40 + 20 + 10 = 200 minutes. The shares add up to 100%.
Runtime is 960 - 200 = 760 minutes. Availability is 760 / 960 = 79.2%.
Now the numbers answer a question. The planned stops (STO) are 40 of 200 minutes. The other 160 minutes are unplanned and are the work list.
The biggest bar, DTT, is the place to drill down. Open it by the code under it, such as machine failure or sensor fault, and rank those causes. That is the method in the Pareto guide.
Level two for one machine
The lower level is where the list becomes specific. Example: a labeller on a bottling line.
| Cause the operator sees | Group |
|---|---|
| Label web break | DTT |
| Label sensor misreads the gap | DTT |
| Label roll change | STO |
| Waiting for label rolls | STI |
| Wrong label roll loaded | DTO |
| Label stock out of tolerance | DTQ |
Six causes are enough for the first month. Each one lands in exactly one group, and each group has an owner.
When a cause becomes the top bar, dig into it with root cause analysis. Do not add ten more codes to the list.
Roll the list out
- Draft with the crews. Operators know the real stops. Test the list on paper for a week.
- Count how often each code is used. Merge codes that are never picked. Split a code that holds a third of all minutes.
- Review every quarter. Lines change, and so do the causes.
- Use the same top level across lines. Then you can add the lines together. Keep the lower level per machine.
For how operators actually enter the codes, read operator downtime logging. For a sheet to start on paper, use the downtime log template.
Common mistakes
- Writing the list in the office and never testing it at the line.
- Mixing planned and unplanned in one group.
- Coding by the operator's guess about the root cause, when the visible symptom was enough.
- Letting each shift add its own codes.
- Using codes as blame. If people fear the code, they pick the safest one.
A code list only works if it is quick to use. Fabrico takes the stop from the machine signal and the operator tags the cause at the line, so each stop has its code and its clock time. See how Fabrico tags the cause at the line.
What people ask next
How many downtime reason codes should I have?
Keep about six groups at the top and a short list under each, so an operator can choose in a few seconds. Review which codes are never used and cut them.
Should I have an "other" code?
No. It turns into the largest bar and hides the real causes. Use a code like "cause not yet known" that triggers a follow-up, and review it every day.
Are reason codes the same as work order failure codes?
Not exactly. Downtime codes say why production stopped. Work order codes say what failed and how it was fixed. They can share names, but the first is for operators in seconds.
Who should choose the code, the operator or the maintainer?
The operator at the time of the stop, because they see it first. The maintainer can refine the cause after the repair. See operator downtime logging.
Do I need different codes for each machine?
Use the same top level for every machine so lines can be compared. Keep the lower-level causes specific to each machine type.
Related guides: Machine Downtime Categories for OEE Reporting · Unplanned Downtime Examples by Type and Machine