Variance… Is it always a bad thing?

In Six Sigma we’re taught reducing variance is a good thing but is this really always the case?

Take currency for example. In the US all paper currency is the same size and the same color. By reducing variance I’m sure the US Mint has saved costs by having a standardized ink color, standardized cutting machines, etc. However, how does this benefit me as a user? When looking at a $5 and $10 bill twenty feet away, most people can not tell the difference. How many times have you had to thumb through your wallet to find the correct bill or worried you gave someone an incorrect note (or been accused of giving back incorrect change if you’ve been on the receiving end)?

Now compare the US currency with Australian currency. Each denomination of paper money is a slightly different size and color (note: while there are extra upfront costs to the process, they are greatly offset by reduced printing demand as the result of an applied plastic coating.).Essentially the country has error proofed (i.e. Lean manufacturing) its currency to the end user.

The point I’m trying to make is that sometimes in Six Sigma we focus so hard on reducing variance to cut costs that we overlook characteristics deemed critical to quality by the end user. Even though costs have been reduced, underlying problems still exist, leaving the customer to view the product service of sub par quality.By introducing variance as a possible solution, one may be able greatly improve customer satisfaction.

Comments 3

  1. JConsidine

    I agree wholeheartedly – I always grimace when I hear someone quote Neutron Jack Welch’s famous line "Variation is Evil" – not just because it seems to ascribe a sinister moral overtone to the notion, but because variation is everywhere and in everything.

    I find that there’s still a lot of misunderstanding about variation in general – otherwise there wouldn’t be so much "tinkering" with things when one datapoint "seems" to look off.

    There’s also the systems notion to consider – another process may cease to work if variation is eliminated in an upstream process or input.

  2. Eugene Koh

    You had mixed the concept of ’variance’ with ’diversity’. Variance in six sigma refers to those things that are INTENDED AND EXPECTED TO BE THE SAME but in reality they ended up different. If it is INTENDED to be different ,just like the case you mentioned about coins and currency, it is diversity.

    These are 2 different subjects. So there is nothing wrong about reducing variation, it’s just a matter of how much do you want to reduce the variance till, which is a function of cost and effort.

    Eugene Koh

  3. Goutham B

    My understanding of Six SIgma is the process of reducing variance from requirement. In your example, I believe your gripe is with the fact that the requirement (Customer specs) didnt build in / include diversity. Six SIgma like every other process tunes based on provided parameters. It is not the role of the optimization process to question the parameters (i.e. The customer says this; but they dont know any better; they should be asking for that). So I would back up Jack Welch’s comment. However the caveat remains that any process can only be a reflection of perfection as percieved by the customer. I think it is wrong for one to ask for a tuning process to change the customer’s thought process to make a paradigm shift in ones perception of perfection. This BTW is one of the documented drawbacks of Six Sigma that the organization has been trying to address of late

Leave a Reply