The article entitled ADelays,
Bugs, and Cost Overruns Plague PeopleSoft=s
Services@ in the September 24th
edition of The Chronicle of Higher Education reported on the frustrations
that administrators at a number of institutions have experienced, and are
continuing to experience, in implementing a new generation of administrative
software from PeopleSoft. Some, but not all, of the difficulties reported in the
article have been encountered by those of us in the California community-college
system who are now implementing not only PeopleSoft but also other such
software.
As the article noted, a number of structural features of that
class of software can lead to cost overruns and dissatisfaction with the results
of its implementation. These features include the complexity of the software,
the difficulty of deciding about initial configuration options, the dangers of
choosing to customize the software, and the realities of coping with bug fixes
and updates to the software. However, the article did not give details about how
to avoid the problems it described.
Even worse, without more information, the article might lead an
administrator new to dealing with this class of software to the incorrect
conclusion that he or she must simply choose the right softwareCi.e.,
something other than PeopleSoftCto
avoid the problems described. That conclusion would be unfair to PeopleSoft, a
company that has worked closely with higher education in recent years to produce
high-quality, innovative products for our administrative purposes.
More importantly, the notion that there is a product from any
vendor that will work right out of the box with few dangers of cost overruns or
of dissatisfaction with the results is simply mistaken. Administrators who reach
that conclusion will make bad decisions for their institutions. Many of the
dangers described in the article are inherent in the nature of this genre of
software. Installing a complex administrative-software system that is sufficient
to meet the needs of a modern college or university requires more-sophisticated
knowledge and tactical decisions from us than have been required of college
administrators in the past.
While it would be nice if we could depend on software vendors to
educate us about the prerequisites and ramifications of choosing, implementing,
maintaining, and using their products, we cannot. Salespeople say what we want
to hear. For example, they often emphasize the very features of the systemsC
e.g., ease of customizing the products to the potential buyer=s
current business practicesCthat will
lead to serious problems down the road. They know we want to hear that we really
don=t have to change our ways to use
their software.
Nor can we depend on the consulting firms that make their living
implementing administrative software to give us the complete story about
deploying the systems. After all, by far the biggest cost in any such
implementation is often the exorbitant fees we pay consulting firms to help us
set up and customize the software. Although they can be valuable partners, their
interests do not always coincide with ours.
Ultimately, we have to depend on ourselves to know as much as we
can about the strengths and weaknesses of the software we are considering.
There are a number of good software systems, including
PeopleSoft, that institutions can purchase. The problem is that it is all too
easy to implement any of them using strategies that will lead to both short- and
long-term problems. To prevent the problems, administrators need to understand
the complexity of the systems, the dangers of customization, and the critical
nature of documentation.
First, the software packages are complex systems. Changes made
to computer code or data-base structure in one part of the system can affect
other parts. That is both good and bad news. It makes fixing some problems easy.
One community college solved performance problems throughout its PeopleSoft
system by making fairly simple changes in the programming commands that put data
in and retrieved data from the underlying Oracle database. But changing code to
fix one part of the system can produce problems in another part.
Knowledge of that aspect of large-scale software should result
in some concrete administrative strategies. Changes to the software must be made
serially, be heavily documented, and be tested carefully for unexpected
consequences. If the staff of a consulting firm or the institution=s
own information technology staff is allowed to operate in any other wayCe.g.,
to make many changes at the same time, or to fail to document changes carefullyC
unexpected problems and cost overruns are likely to occur.
Second, implementation decisions must be made with future
maintenance in mind. Failure to understand that fact is the most serious mistake
administrators can make in implementing such software. For example, PeopleSoft
allows the customer to customize its software or to create new applications to
use alongside those that the company has developed.
However, customizing a commercial application creates
significant difficulties when the vendor releases a new version. That version
might contain new features that conflict with the changes you have made, or that
remove structures you have depended on in your customization of the product. If
you have customized the software, your information technology staff will have to
spend a great deal of time evaluating the relationships between your
customizations and the vendor=s
changes before you can proceed with any upgrade. The more changes you make, the
more time it will take to go through the process every time you upgrade your
software.
Even though PeopleSoft provides sophisticated tools to help with
the process of comparing your software with the new version, upgrades of
extensively customized systems can take months. That is true for all of
PeopleSoft=s competitors, too. If you
don=t have enough computing staff
members to perform the upgrades, you will have to pay for high-priced consulting
help.
Here again, knowledge of the details should lead to concrete
strategies. You should change your business practices to match your software,
instead of customizing the product. That is going to be uncomfortable for many
staff members, but not as uncomfortable as not being able to upgrade or patch
the software because you don=t have
the resources to update a customized product.
Any good software will include ways to tailor it to your needs
that don=t involve customization. For
example, with PeopleSoft, you can write your own self-contained subsystems that
don=t cause the difficulties described
above at upgrade time. In any system, you can use the report-writing tools to
create custom reports that extract just the information you need without
customizing the software.
Third, documentation of set-up decisions and changes is critical
to a successful implementation. That may sound obvious, but the reality is that
neither consultants nor information technology staff members like documenting,
so it seldom gets done well, if at all. Poorly documenting the implementation of
a complex administrative software system leaves the institution at the mercy of
information technology staff members, who are notoriously difficult to retain
these days. Even worse, failing to create clear, usable, comprehensive
documentation means that the software cannot be upgraded without figuring out
how it was set up in the first placeC
which takes time and money.
When software implementations go bad, the temptation is to blame
the software. Some faults do indeed lie there, but many others lie in our
administrative decisions. We expect complex software systems to work right out
of the box. We fail to arm ourselves with an understanding of the details of the
systems we have chosen. We train our staffs insufficiently or incorrectly. We
choose the comfort of customizing software to the way we=ve
always done things over the difficulties of using the basic, easily upgradable
product. We let staff members get away with poor documentation. We turn too many
tasks over to consultants, so that our staff members are lost when the
consultants leave.
The technological sophistication required to implement
administrative software is greater than that to which academic administrators
are accustomed. But no piece of shrink-wrapped software alone can provide the
functionality we need to serve students who live in the information age. We have
to develop the more-complex strategies required to implement and manage the
tools of this age.
Joel M. Smith is dean of academic information services at Mira Costa
College. This article is based on a letter written originally to The
Chronicle of Higher Education.
Top of article |