MEASURE: Introduction

During this phase, you will measure and analyze the critical to quality output identified in the
Define phase. This will establish the baseline process performance with respect to the goal
of the process improvement project. You and your project team then need to understand
the hiking shoe production process in as much detail as possible by creating a detailed
process map. The process map is carefully reviewed in order to identify the parameters
which are likely to influence the critical to quality output. In an actual project, a data
collection plan would be developed and implemented for these parameters.

Now we will accomplish learning objective:

4: Establish the baseline process performance with respect to the customer

MEASURE: Customer Expectations

As you remember from Session 1, in the DEFINE part of our project, we met Rafik who is
planning to go on a hike. The trek for the hiking Rafik planned will be tough and the weather
forecast is calling for a heavy rain for the first three days. Rafik therefore put the attribute
“waterproof” as a higher priority compared to “style” and “comfort” for his quality expectation
for his new hiking shoes.

Now that we know Rafik cares about waterproof the most and now that you have seen how
hiking shoes are produced, let's look go to our factory where Diane is working.

The company where Diane works has a long tradition in hiking shoes production. Located in
Bavaria, south-eastern Germany, “Schuhmann” is known for their high quality products and
the wide range of hiking shoes styles for every day wear as well as for different sports. They
measure waterproof using two tests, as you can see on the next page.
MEASURE: Waterproof Tests

How waterproof a fabric is, is usually decided from the results of two tests:

The first is the Rain Room Endurance test (RRE) where a fabric will be placed in a room where
different climatic conditions are reproduced. Leakage is then recorded in relation to the
amount of time and pressure of simulated rainfall.

The second is the Static Column Water Resistance test (SCWR). In this test a column of water
is placed on top of a fabric and the level at which the water begins to seep through is
recorded.

The results for the SCWR test are measured in mm; as such a 5,000mm rating means that
the water has permeated through the fabric at the 5,000mm level so it is not as waterproof
as a fabric with a 10,000mm resistance.

The following table shows a standard waterproof rating method:

Images taken from www.evo.com

High quality waterproof fabrics are in the 20,000mm+ range. These are totally waterproof
even under serious pressure. These items will be able to withstand shallow-depth
submersion without leakage.
For this DMAIC Project, we will consider the results of the Static Column Water Resistance
test used on our hiking shoes.

MEASURE: Waterproof Testing Specifications

The waterproof specifications at Schuhmann are as follows:

• Target: 15,000mm
• Upper specification Limit (USL): 16,000mm
• Lower specification Limit (LSL): 14,000mm

Schuhmann targets that the shoe will be resistant, on average, to 15,000mm of water placed
in a column of water over the shoe, which the shoe placed on its side so that the integrity of
the sole is tested.*

The hiking shoes must be resistant up to a water level of at least 14,000mm (Lower
Specification Limit). If this limit is not reached, the hiking shoes should be rejected as poor
quality.

There is no rejection if the resistance is above 16,000mm.

In order to establish the baseline process performance for waterproof at Schuhmann, Diane
has collected waterproof data for 720 shoes at Schuhmann using the SCWR test over the
past 3 months (13 weeks):

* The authors would like to clarify that the SCWR test is used on fabrics but has been adapted here for the hiking shoe
scenario.

MEASURE: Waterproof Results at Schuhmann

As stated on the previous page, Diane has collected waterproof data at Schuhmann for 720
shoes using the SCWR test over the past 3 months (13 weeks):

Again, the specifications for waterproof at Schuhmann are as follows:

• Target: 15,000mm
• Upper specification Limit (USL): 16,000mm
• Lower specification Limit (LSL): 14,000mm

The results of this baseline data collection are very disturbing!

Instead of being on target and well above the lower specification limit (LSL) of 14,000mm
(which means that water permeates the shoe at only a level of 14,000 mm), the distribution
looks like this:
The data is normally distributed with mean μ = 14,700mm and standard deviation σ = 550mm.

Diane could calculate, using her knowledge of the Normal distribution, that this meant over
10% of the shoe production (based on this sample of 720 shoes) had waterproof that did not
meet the lower specification limit (LSL) of 14,000 mm!

z = (X – μ)/σ = (14,000 – 14,700)/550 = -1.27

From the Z-tables, we find the probability that Z ≤ -1.27 is 10.2%.

It is clear to Diane that the process for the hiking shoes is not acceptable in terms of
waterproof, according to the specifications (which we assume here are industry standards
and are the level of waterproof quality that Schuhmann wants to deliver).

We have presented this in the MEASURE section of our DMAIC project, because we are
measuring data to evaluate our now known customer expectations from DEFINE. In a real
process-improvement project, it may be that this type of information actually triggers the
entire project.

The baseline data may be available as part of the DEFINE process, or it may be that as we
continue to investigate our process and its parameters, that we have to go back and
measure a more appropriate parameter as a baseline.
This case study is a simple illustration of the general procedure, and here we have
established that the waterproof of the shoes is not of acceptable quality and have an idea,
using the data, of what percentage of our production is of poor waterproof quality.

You and Diane see two problems with the process:

1. The process is off target: the process mean (μ) = 14,700 mm instead of the target =
15,000 mm.
2. The variability of the process is too large: this can be seen in the standard deviation
(σ) = 550 mm.

At Schuhmann, they want to have waterproof level above 14,000 for over 99% of the shoes
(less than 1% of the shoes should have waterproof level less than 14,000)

Can you help Diane and identify which standard deviation acceptable if the process were on
target at μ =15,000mm for less than 1% of the shoes to have waterproof level below 14,000
mm?

1. From the standard normal tables, what is the maximum z value that corresponds to a
probability P(Z ≤ z)<1.0% ?

z = -1.27
z = 1.27
z = -2.32
z = -2.33
z = 2.33

2. Now that we know z, what should our standard deviation (σ) be in order to have less than
1.0% of waterproof data below the lower specification limit 14,000mm?

430 mm
550 mm
450 mm
400 mm

MEASURE: Summary

Now Diane knows that she should continue with the project to:

1. Bring the process on target: improve the process so that the mean of the waterproof
distribution increases from 14,700 mm to 15,000 mm.
 2. Reduce the variability of the process: improve the process so that the standard
deviation of the process is reduced from 550 mm to at least 430 mm.

This is shown in the picture below: the blue curve shows the baseline data again, with
mean 14,700 mm and standard deviation 550mm. The green curve shows the
process at target with mean 15,000 mm and a standard deviation of 430 mm, so that
only 1% of the shoes do not meet the waterproof quality lower specification limit.

You have now completed the first part of MEASURE. You and Diane have established the
baseline process performance with respect to the customer expectations for waterproof.

In the next section of MEASURE you will look more deeply into the hiking shoe production
process flow and identify the key parameters to be measured which are likely to influence
the critical to quality output.