Page 98 - The Toyota Way Fieldbook

Chapter 4. Create Initial Process Stability 75

A common example of external variability is product demand, or model mix.
The reality in today’s manufacturing environment is a shift from high volume,
low variety (HVLV) to low volume, high variety (LVHV). This creates many
challenges since the different products require varying amounts of time and/or
processing steps to manufacture. Balancing the resources (people, machines,
and material) with this inherent product variation is virtually impossible with-
out employing the concept of isolating variability. If you are unable to control the
variability, the next best option is to isolate it, which reduces the impact on
the whole. In the last chapter, where we discussed value stream mapping, we
brought up the concept of a product family. In fact, separating products into
“similar families” that belong to a common value stream is an example of iso-
lating variation.
In considering methods for isolating variation, it’s important to think about
future steps. These will include the creation of flow and pull, as well as stan-
dardizing. The value stream mapping process is a useful tool for developing an
understanding of the relationship of the different processing steps and times,
and the effect on creating balanced flow later.

The 80/20 Rule
The 80/20 rule is useful when considering divisions in products that will iso-
late variation. The time required to complete the product at each operation is
the critical element for the creation of future flow, so look at the products to
determine where the variation occurs relative to time. To reduce variability in
processing time we consolidate similar products based on the required pro-
cessing time. Time is also the important factor in establishing the alignment of
resources.
In fact, some operations are not affected by product variability. (We call
these operations “flow through” processes, because all products flow through
without any change in time required.) For example, a washing or cleaning oper-
ation is not affected by the variation-in-part complexity, or model type, and thus
requires the same amount of time regardless of what is produced. We are looking
for the operations that are most affected by the product variation, especially if
they create a bottleneck.
The tricky thing about variation is that 20 percent (the minority) of the
product often provides 80 percent of the total variation. This may be difficult to
see, because the ripples of prior variation create new ripples. A great deal of
variation can seemingly be “removed” from the overall results of an operation
by simply isolating this minority—“seemingly” because the variation is in fact
not removed at all, but the magnifying effects are reduced, providing greater
consistency.