In many companies, and in particular in many Chinese companies, the response to problems is often “let’s invest in technology”. Unfortunately, spending a lot of resourcing on engineering is seldom the best solution.
I just read a great story that illustrates this.
A toothpaste manufacturer had a regular quality issue that was generating bad publicity: some boxes didn’t contain a tube of toothpaste. It was due to the way the production lines were setup.
They spent 8 million dollars on a 6-months project led by an external engineering company that developed high-precision scales that would fit perfectly on the assembly lines. They were designed to ring an alarm every time an empty box was about to get packed and shipped.
A few months later, the CEO was happy because this quality issue has disappeared. But he was intrigued by a number in the report he was reading: in the past month, zero incidence of empty box had been flagged by the high-precision scales.
An investigation showed that the production staff was tired of walking to the scale and remove an empty box multiple times a day. So they found their own solution: they placed a 20 dollar fan just before the scales. Any empty box was blown away. Problem solved!
Note the superiority of the fan: not only is it much cheaper and faster to put in place, but also it segregates bad pieces (by blowing them into a container by the side of the line) instead of stopping the line and requiring an operator to walk each time.
So, what should that company have done? Here is a simple approach I’d recommend:
- Gather data about the occurrence of this issue — maybe they’d need to have someone checking 100% of a line’s output for 1 or 2 days.
- Analyze the data. For example, look for trends or patterns and interpret them (e.g. it might occur mostly in the 10 min following the restart of the line after the lunch break.)
- Discuss with the people working in this process all day long. Maybe they know the source of the issue.
- Discuss with the engineers who developed and tweaked the process. Maybe they have insights, too.
- Based on the above data, conduct a root cause analysis and follow the Plan-Do-Check-Adjust approach. Common root causes are the product design, the process design, the competency of operators, a wrong choice of supplier, etc. It is NOT a “whose fault is this” exercise.
- If corrective actions can’t be found, or if they are expensive, look for a mistake-proofing system. I described the approach to implement corrective actions and maximize the chances that the problem won’t come back in the future in this article.
There are many types of mistake-proofing systems, and some are better than others. Here are a few of them:
(The above carton comes from Poka-Yoke: Improving Product Quality by Preventing Defects).
Making the problem impossible (e.g. by blowing the empty box away with a fan) is superior to detecting the problem (e.g. by having a scale sound an alarm) since the second solution adds to operators’ workload and can be bypassed when operators are in a rush to complete a job.
However, in China, engineers love mistake-proofing devices based on sensors. Problems can be detected, and sometimes automatically corrected. But it’s never as good as a fan, or a stick, or any other really simple device. Mechanisms based on sensors need to be calibrated and maintained regularly — unfortunately this is often skipped.
And the very best? Addressing the root cause(s), naturally. This way, the same issues don’t come back again and again.