Lorex Inc Case Analysis

Attachment previewI am using the case to make certain quantitative points, mostly used in TQM, Six Sigma and others.  And I can’t do that unless you have tried.  There are no concepts there that you did not already know but they are just hard to frame and apply.

Here we go:

In order to deal with these types of quality process issues (which organizations have to do a lot) you have to follow some framework for analysis.

Step 1:  What are we trying to decide?

We certainly want to know what fill target we should recommend.  The issues is that defects (bottles that are under 10oz will cost 20% of price).  But on the other hand if we target a high fill target, it costs more in ingredients.  So we need to understand the trade-offs.

Step 2:  How would you measure it?

In this case we should be concern about profit.  But it is not always the case because in operations, a lot of times we only worry about the cost.  But since we are given the price etc, we can use profit as what we want to measure in order to decide what target level we recommend.

Then one of the questions we need to address is the defective rate.  In order to understand this, we need to know if we start the machine rolling, what will the volume of the bottles be like?  That is why they sampled 144 bottles.  As you can see the weight of the bottles, based on the sample, varies.  For the sample, they targeted 10.2oz; the bottles were all over the map with 12 of the bottles less than 10oz (defective).  If you plot the 144 bottles, it kind of centered around 10.2oz (as expected or else something else is wrong with your machine).  Well, it does look kind of like a normal distribution (I know you don’t ever want me to speak those words in your face).  If you think it is good enough to be considered normally distributed, then life is relatively easy for you.  Because if you target 10.2oz, with a standard deviation of .16oz(measured in the sample), you can theoretically find out what percentage of your bottles will likely be under 10oz.  In this case, it is about 10%.  But you don’t have to assume that it is normally distributed.  You can always use 12 out of 144 as your defective rate (8%) as observed in the sample run.  Then you problem is then what about 10.3oz, 10.1oz.  For Lorex, let say we can assume that they are normally distributed.  IF you are real picky, you can use Chi-Square Statistical Test to validate that.  By the way in Excel, there are all kinds of functions that can calculate the % for you.  You can consult your statistics textbook (if you haven’t burn it by now).