A CASE
STUDY—STRANGE BEDFELLOWS
Dan River, Inc., is
a Virginia-based textile firm with 100+ years of history. In the
early 1990s, management decided the existing data systems were an
impediment to the growth of the corporation and set forth to replace
them. In doing so, they chose to work with a firm called Datatex,
from Israel, but now with domestic operations headquartered in
Atlanta. Datatex's AS/400-based system, marketed under the acronym
TIM (Textile Integrated Manufacturing) has been written
specifically for the textile industry. Recent enhancements make it
suitable for apparel manufacturers as well. When first purchased,
TIM's European roots made implementing a challenge, but willingness
on the part of Datatex to add modules that aligned with U.S.
accounting and marketing practices overcame those obstacles. As Dan
River Apparel Fabrics began to enter more global markets, TIM's
international character became a strong asset. The software has
been in constant use since the implementation began in 1992 and has
seen five upgrades via new releases. The use of TIM was recently
reaffirmed through its implementation in newly acquired greige and
finishing operations.
The strengths of
TIM are many and include its industry focus, an ability to customize
via tables that control both functionality and data format,
enhancements gained through regular maintenance releases, and the
maturity resulting from an established global community. TIM's major
opportunities for improvement have to do with planning and detailed
machine scheduling. Apparel Fabrics has helped to overcome the first
by working with Datatex to develop a closed-loop master scheduling
tool by integrating their order simulation module, COBRA, with the
existing Production Planning module. Through this enhancement, the
user will be able to insert a production plan and explode it onto
critical resources. As real orders are booked and exploded, the
forecast, and its resource requirements, will be returned and a
specific product's requirements will be inserted. Thus, the
feasibility of the production plan will be tested and converted to
real orders as they are entered and released.
Datatex offered
detailed machine scheduling, through a third-party product, but the
functionality and package compatibility were left wanting. Because
of this, and the need to better schedule and control newly acquired
manufacturing facilities, Apparel Fabrics began to review options to
secure a scheduling tool for its weaving operations in Sevierville,
Tennessee. It was decided to use the Factory Planner module within
i2, a decision that was made easier because Dan River had purchased
it for use in the Home Fashions Division. Factory Planner is the
most mature portion of the i2 suite of products and focuses on
constraint balancing and schedule optimization, the characteristics
of an APS system. Apparel Fabrics decided to proceed and began to
develop a strategy for combining these two products.
What made the
integration especially challenging was the merging of a table-driven
ERP system, with its innate flexibility, and a PC-based APS system
that had totally customizable files and many different functional
attributes based on file and default structures. In fact, defining
data structures and stabilizing outputs soon became the biggest
challenge. As a result, Apparel Fabrics compromised package
optimization for the speed of implementation and simplicity of
system outputs. This occurred chiefly in the area of routing or path
selection.
Theoretically,
Factory Planner should be allowed to select the best pathway and
resources for the product being scheduled, based on the constraints
in place at the time the order is scheduled. To accomplish this, i2
requires a very specific hierarchy of alternate resources for each
routing. The problem confronting
Apparel Fabrics is
that optimal flows depend on both bottlenecks and order quantities.
As cost and margins are significant to the success of a textile
firm, process routings will change from order to order based on lot
quantities and operations off different production orders will
routinely be merged and, subsequently, split to gain manufacturing
economies. Thus, rules embedded in Factory Planners would not
accurately model the best alternatives based on different
quantities. As such, the flow had to be modeled in TIM first, which
supports and accurately costs alternatives, and sent to i2 as the
routing for this entity.
Another unique
attribute of the Apparel Fabrics installation is the use of pseudo
production orders that break greige manufacturing orders into
subsets. These allow more accurate modeling of shop floor activities
without increasing the overhead required to create and track the
orders. These sublevel orders allow for the accurate depiction of a
production order's critical path. In this manner, the flexibility
lost by creating a routing for each order was regained by breaking
the activities into substeps that enabled i2 to accurately
prioritize events. Thus, careful analysis of the business needs,
reviewed in concert with the attributes of the software tool,
allowed Apparel Fabrics to achieve the needed benefit swiftly and
with a minimum of disruption.
LESSONS LEARNED
After reviewing the
implementation, the important factors that facilitated the
expeditious integration of Apparel Fabrics' ERP system to an APS
tool were
• defining a vision...and sticking to it
• identifying the required data and its significance
• assuring the accuracy and efficacy of that data
• maintaining the focus and limiting the scope
• simplifying wherever and whenever possible
• involving the users from start to finish.
Easy, right? Wrong!
A number of obstacles were confronted and handled, some elegantly
and some less so, but by remembering the parameters shown above,
progress continued unchecked. The biggest challenge came from
outside Datatex and i2 and serves as a key lesson. Apparel Fabrics
added the installation of a third system (data collection) to this
mix, a beta version no less, and it has created the majority of the
headaches. It should be noted that the simultaneous implementation
of multiple systems can be painful and should be avoided if
possible. Again, a bite at a time makes for a more enjoyable chew.
EVOLUTION, NOT
REVOLUTION
Most firms will
find themselves using a suite of tools as they progress in their
systems' journey, because of the desire to use application specific
packages or because of the layering of established and functional
systems. If this description fits your situation, you need not
panic, because technology will enable you to integrate different
packages even more seamlessly in the future. Therefore, unless your
strategic objectives indicate one supplier be used for all
applications, continue the course. Refuse to accept mediocrity or
ineptitude on the part of your software, but realize that changing
out systems simply to install the "latest and greatest" will too
often become a fool's errand.