Last Updated on 13 September 2023
When you have an aim such as making a product, you won’t be able to make it perfectly every time. If you aim to make a 1000mm bar a million times, some will be 1001mm, and some will be 999mm.
It isn’t always worth your effort to achieve perfection, as there isn’t always a benefit to your organization. Getting your delivery to your customer exaclty on the 2 week mark is often no worse than getting it a day earlier or a day later. Your product may be of equal value as long as it is within their allowable bounds.
You can’t accept all variation though, as if you sent them a 2000mm bar, it probably wouldn’t be any use to them at all. Variation is therefore split into two two categories: common cause variation and special cause variation.
- Common cause variation is small differences from every-day causes that your customer will allow. These causes are often unavoidable, such as from the environment or factors outside our control
- Special cause variation is larger differences from relatively unusual causes that will cause losses to your customer, and therefore to you. These are usually avoidable, and come from your actions.
It is usually the aim of six sigma projects to remove the special cause variation from the system. Common cause variation is tolerated as it is usually outside our control and not causing issues to your customer.
Taguchi Loss function
The Taguchi Loss function is an alternative to the pass / fail mentality that usually occurs in Six Sigma, and which lets you get away with ignoring the common cause variation. The function says that as soon as you deviate from perfection, there is a cost either to your customer, to you or a connected party.
It follows this curve:
As you can see, the vast majority of the cost still comes from the larger special cause variation, so these are still the ones you want to eliminate first. It teaches though that there is still a cost from the special cause variation, so if you can investigate and minimize (or even remove) this variation, there will still be a benefit to you and/or your customer.