|
               
|
By Dick Kuiper
Over the last 30 years, many pioneers have struggled down
the trail of manufacturing systems implementation. The road
is littered with the bones of software and hardware that
didn't work and manufacturing managers with arrows in their
chests. After all this time, many have yet to learn the
rules for not failing. Yes, I said "not failing." It differs
from succeeding. Not failing is about the things you
shouldn't do; these are not always the opposite of the
things you should do.
In analyzing hundreds of projects - from stellar
successes and dismal failures - we have itemized a list of
rules to help avoid failure. Following them won't guarantee
success but it will reduce the risk of failure.
#1: Don't Develop a Wish List
One of the biggest bugaboos in software selection is a
system requirements list that enumerates wishes instead of
needs. Develop a system requirements list for what it is
intended to be, a checklist to highlight the truly important
needs and to avoid forgetting anything. While the list may
include medium and even low priority needs, it must be
constructed in such a way that the high priority stuff
really stands out. If not, the critical requirements get
lost in a sea of nonessentials. The risk is then searching
for packages that are abundant in general functionality
rather than those that best address your most important
specifics.
You can get standard requirements lists from any number
of consultants. The only thing the lists have in common is
size - they're lengthy and overly comprehensive. But they do
provide a helpful starting point, and certainly are
preferable to reinventing the wheel. But, without proper
guidance, you are subject to the "kid in the candy store"
syndrome.
#2: Don't Let the Project Drag On Too Long
The longer it takes to install any system, the greater the
problems with commitment, continuity and consistency, and
hence the risk of failure. A year is the generally accepted
maximum span, from assigning the project team through
implementing the core applications. If your installation
takes longer than those in similar business environments,
you're either doing too much, moving too slowly, or have the
wrong software. Take the steps necessary to get done in a
year or less if possible.
#3: Don't Plan the All-Encompassing "Mother of All
Projects"
The smaller the project, the greater the likelihood of
success. Resist the urge to incorporate all of the ancillary
projects (e.g., econometric forecasting, CAD interfacing)
with the core manufacturing system into one huge
deliverable. It will collapse under its own weight. Add-on
modules and non-critical customization should be defined as
follow-on phases or separate projects with their own
schedules and team assignments.
#4: Don't Make It a Computer Project
System implementation projects are not exercises in
technological excellence; the goal is to provide an
information flow system that helps people run the
manufacturing business more effectively. The people running
the operating units of the business, not computer
programmers, are in the best position to decide the what,
when and where of information needs.
#5: Don't Modify the Core Software Package
Select a system that meets your basic functional needs
without software modifications. Many packages today are
better designed for external customization, alleviating the
need for internal modification. Parameter settings allow you
to activate/deactivate individual functions, custom reports
are readily developed by users, and even screen layouts can
be altered without changing the core programs. Rather than
charging ahead to make parts of the new system operate like
the existing one, concentrate on training people how to do
the old jobs with the new system.
#6: Don't Let Management's Commitment Flag
Because of the size and cost of most manufacturing systems
projects, there usually is a lot of fanfare at the start-up.
Top management pronounces its support, but later loses touch
with the project in favor of other priorities. Sparing
management from getting bogged down in details, yet keeping
them informed and involved through a judicious status update
process, is the key.
#7: Don't Skimp on Training
Up-front education generally is provided to key players,
often with the intention that they will train the others.
This trickle-down almost never happens unless it follows a
structured train-the-trainer format. Untrained users
represent one of the biggest risks: they will revert to old
procedures and lose most of the advantages offered by the
new system. Before too long, you will realize you have two
systems instead of one, and that can be worse than simply
staying with the old system.
#8: Don't Let Systems Intimidate You
An unfortunate view taken by some manufacturing managers is
that systems projects are so difficult that only the
technical whiz kids should be expected to comprehend them.
This is not true. While they are complex, if tasks are
structured into logical bite-size pieces and the project
team does not fall victim to buzzword-mania, most managers
should be able to get their arms around the project.
Without focusing too heavily on the negative, remember
that making a "to do" list for an implementation project
should be followed by also making a "not to do" list. It
will greatly reduce the risk of failure.
Kuiper is vice president of Expert Buying Systems Inc., and
can be contacted at (702) 363-4046. This article is the
first in an occasional series on the selection and
implementation of manufacturing applications software.
|
