Governing Magazine/March 1995

TECHNOLOGY COLUMN

THE BENEFITS OF `STRATEGIC' COMPUTING

By Jerry Mechling

Jerry Mechling is director of the Program on Strategic Computing and
Telecommunications in the Public Sector at Harvard University's
Kennedy School of Government. To reach him, call 617-495-3036 or send
e-mail to jerrym@ksgrsch.harvard.edu.

In 1967, I took my first job in government, as an aide to Mayor John
V. Lindsay of New York City. At that time, the mayor--along with
nearly all public- and private-sector executives--thought computing
was something to be delegated to technical experts. Since then, the
cost-effectiveness of computers has improved by a factor of more than
2,000. Now, senior executives are being told that computing is
`strategic' and that they should get personally involved. If so, how
is strategic computing different from what we had before? How do you
evaluate or measure its potential benefits?
 Relative to manual methods, computers have high fixed or set-up costs
(programming) and low marginal costs (operations). Computing has thus
traditionally focused on high-volume work routines, such as payroll
processing; the automation of routine work was practically the only
kind of computing that made sense in 1967.
 But computer processing costs have fallen 25 percent a year since
then, primarily due to continuing improvements in chip fabrication.
This has opened the way for new applications. Where once restricted to
incremental improvements in high-volume routines, computing can now be
cost-effectively applied to low-volume work and also to the creation
of completely new routines. In short, computing can now be used for
innovation, not just automation. Therein lies its strategic potential.
 Are the benefits worth it? Strategic computing promises to shift
work far from the status quo. The goal is not a 5 percent productivity
improvement, but more like 50 percent to 100 percent. Unfortunately,
any change so large is, by its very nature, clouded in uncertainty.
Return-on-investment analyses in this context rest on assumptions so
arbitrary and uncertain that the calculations often are completely
worthless. The truth is, you can't always know in advance the
magnitude of benefits strategic computing will yield. You must instead
invest to learn if the benefits are worth it, and this learning must
be based heavily on doing, not just on planning and analysis.
 The kinds of activities that need to be supported are essentially
research and development, and include assessing the needs of
customers, benchmarking against outside organizations, and
experimenting with new technologies and processes. Granted, it is
extremely hard to fund such investments in government, especially when
times are tight. Nevertheless, private firms know they can't kill R&D,
even as they downsize; otherwise they will soon have no new products
and revenues. Government without innovation is also a losing
proposition. We must find ways to get something like 5 percent or 10
percent of our information technology budgets into strategic,
innovation-oriented investments.
 How do you reap the benefits? To begin, senior managers can no longer
safely delegate computing to technologists. To reengineer work across
program or organizational or jurisdictional boundaries, we need
leadership, not just management. We need people who can mobilize
support for new solutions, not merely those who can manage within the
status quo. We need people who are committed and skillful enough to
succeed in high-conflict settings.
 One model for strategic computing leadership is available from the
private sector and depends on top management power and leveraging
crises to quickly implement new computer-based work procedures. Jack
Welch of GE, the late Mike Walsh of Cummins Engine and Tenneco, and,
more recently, Lou Gerstner of IBM are some of those who come
immediately to mind. There are important lessons here for public
leaders, but of limited application. Significant--i.e., painful--
change requires significant agreement on the need for change, and a
crisis may help in gaining such agreement. Nevertheless, requirements
for openness and due process will inevitably slow down and complicate
public-sector strategic computing. Moving fast and from the top can be
helpful, but it risks ignoring the importance of the many stakeholders
in the public arena, especially the senior civil service, oversight
authorities and external customers.
 Strategic computing, in sum, is risky. Many projects will fall short,
at least as measured against the promises made to gain support for
them. Even projects with strong strategic potential may need to be
scaled back. While strategic computing is rapidly becoming much more
feasible throughout government than it was in John Lindsay's time,
success with specific and risky projects will inevitably require up-
front and continuing investments in learning, leadership and
disciplined judgment--not just risk-taking.

----------------------------------------------------------------------

Copyright 1995, Congressional Quarterly, Inc. Reproduction in any form
without the written permission of the publisher is prohibited.
Governing, City & State and Governing.com are registered trademarks of
Congressional Quarterly, Inc.
http://governing.com