Legislative Analyst's Office
State of California
June 16, 1994
Information technology (IT) is a key component of the
state's operational infrastructure. Each year the state spends
over $1 billion on IT $1 out of every $23 spent on state opera-
tions to carry out state programs. Without it, many state activi-
ties would be cost prohibitive or unable to serve California's
growing population.
Given the amount of money involved and the importance of IT
to governmental operations, it is critical that the state's IT
efforts be cost-effective and that departments have the tools to
use IT well. This policy brief reviews the state's IT infrastruc-
ture and presents our findings and recommendations. Previously,
we published a Supplemental Analysis of the 1994-95 Budget Bill
which contained our findings and recommendations regarding the
state's Office of Information Technology (OIT).
Findings
While there have been many significant advances in the
state's deployment and uses of IT since the state's first comput-
er was installed in the mid-1950s, there remain fundamental prob-
lems that prevent the state from realizing a better return on its
IT investment. When IT projects are in trouble they typically are
experiencing major cost overruns, significant implementation
delays, and solutions that do not work as effectively as planned.
These problems reflect, in many instances, planning, coordina-
tion, and oversight failures that have persisted for some time.
Recommendations
The recommendations presented here generally call for the
enactment of legislation or specific actions to be taken by the
administration to address the problems identified in our report.
These recommendations generally fall in two broad categories:
improving state oversight and coordination of IT projects, and
enabling departments to use IT more effectively.
HOW IMPORTANT IS INFORMATION TECHNOLOGY TO
STATE GOVERNMENT?
Since 1956, when the state's first computer was installed at
the Department of Employment, the use of IT has grown to the
extent that it is pervasive throughout state government, costing
about $1.3billion in 1994-95, including telecommunications. Since
1983-84, these expenditures have increased 158 percent.
Without IT, many state operations would be cost-prohibitive
or simply unable to meet current service demands. Thus, it is
appropriate to view IT as a key component of the state's opera-
tional infrastructure. In today's state government, computers
perform a wide variety of functions including aiding in the de-
sign of highway systems, providing local law enforcement officers
with criminal history record information, and providing citizens
with local employment opportunities.
The importance of IT has been recognized by the Legislature,
which has established in law both policies governing the state's
uses of IT and the OIT as the organization responsible for over-
seeing the use of such technology in the state.
WHAT IS THE STATE'S INFORMATION TECHNOLOGY ENVIRONMENT?
The state government computing environment is extensive with
a very large inventory of tools and workers, as shown below:
_________________________________________
FIGURE 2:
CURRENT STATE IT ENVIRONMENT IS EXTENSIVE
_________________________________________
* Several mainframe computers and many departmental
computers.
* Thousands of personal computers and computer
terminals.
* Hundreds of technical staff.
* Thousands of nontechnical end-users.
* Numerous databases containing much of the state's
information assets.
* A large number of separately maintained communica-
tions networks, many of them linked statewide and
to nonstate networks.
_________________________________________
The OIT Has Overall Responsibility.
Overall responsibility for the application of IT within the
executive branch is vested with the OIT in the Department of
Finance (DOF). Departments are required to apply IT consistently
with extensive and detailed policies contained in the State
Administrative Manual (SAM), with all their IT activities subject
to the review and approval of the OIT. While the OIT has control
over virtually all aspects of IT and some aspects of telecommuni-
cations (such as local area networks), the Department of General
Services (DGS) has primary control over certain aspects of
telecommunications (for example, communications lines and
equipment).
Departmental Computing.
The current computing environment in the executive branch is
one marked by many departmental computing systems, some relying
on personal computers (PCs), others relying on terminals linked
to a mainframe computer, and some employing combinations of PCs
networked to each other and to a departmental data center, which
is in turn linked to a state or outside data center. Many
permutations exist, and they are growing as the result of
developments which have facilitated the networking of computers
generally. In fact, the continued rampant growth in PCs has
created an entirely new class of computer user, the enduser,
who does not work in a technical classification, but who has
access to considerable computing power.
Data Centers.
While several data centers exist within the executive
branch, only two the Stephen P. Teale Data Center (TDC) and the
Health and Welfare Agency Data Center (HWDC) are very large,
general purpose centers serving multiple clients. The TDC
provides services to over 200 state agencies, while the HWDC
serves primarily the constituent departments and other units
comprising the Health and Welfare Agency. Both data centers
operate on a fully reimbursable basis through fees paid by client
departments.
Technical Staff.
The majority of technical staff employed to design, develop,
and maintain computerbased applications are, for the most part,
in departments using IT. While the data centers employ large
staffs, they do not typically develop computer application
programs other than for their own, internal needs.
Communications Networks.
The many communication networks (typically, communications
lines that hook together computers and users) in the executive
branch allow state agencies to communicate electronically with
each other, as well as to other governmental jurisdictions and
the private sector. For example, the HWDC and the TDC maintain
separate, extensive statewide networks. In addition, several
departments maintain separate networks, with some maintaining
more than one. The DGS has implemented CALNET, another statewide
network service which the DGS hopes will be increasingly used by
state agencies requiring network services. In addition to
allowing departments to communicate electronically with each
other, these networks also enable public access to government
files, an increasing area of interest facilitated by the
expansion of the Internet, which we discuss in a separate
section.
Information Stored on Various Media.
All of the infrastructure described above exists for the
sole purpose of facilitating the collection, maintenance,
security and dissemination of information. While much of the
state's information continues to exist on paper or microfilm, an
increasing amount is maintained in computerbased systems.
Although a significant amount of this computerbased information
is directly accessible by those connected to the particular
system housing the information, there are relatively few systems
that link information among departments. Consequently, the bulk
of the state's corporate database is not easily accessible
among state agencies.
Procurements.
A separate law governs the acquisition of IT products and
services because of the highly technical nature of these
purchases. Because these procurements frequently result in
highvalue contracts, formal protests to proposed awards are not
uncommon. The DGS, which has overall IT procurement authority,
often delegates limited procurement authority to departments it
determines to be qualified.
WHAT ARE THE TRENDS IN THE STATE GOVERNMENT'S
INFORMATION TECHNOLOGY?
The overall trend in state IT is toward expanding the number
of endusers as access to PCbased systems continues to grow. While
the administration does not maintain records on the extent of
this proliferation of PC systems, the number of PCs can be
counted in the thousands. For example, since 198586, the
California Computer Source, a private sector-managed store,
operated under contract with the DGS, has sold state agencies
$245 million worth of desktop computing equipment.
WHAT ARE THE PROBLEMS LIMITING THE STATE'S EFFECTIVE USES OF
INFORMATION TECHNOLOGY?
While there have been many significant advances in the
state's deployment and uses of IT since that first computer was
installed in 1956, there remain fundamental problems which
prevent the state from realizing a better return on its
investment in IT. These problems also contribute to repeated
failed efforts to develop computerbased systems on time, within
cost, and which produce anticipated benefits. These problems are
important not only because the state's annual investment is high,
but because they represent delayed or missed opportunities for
improving governmental services.
Problems are Not New.
In both 1967 and 1973 we reported on the state's uses of
computing technology and in each of those reports described
problems which needed to be resolved in order for the state to be
able to apply this technology in the most cost-effective manner.
Subsequently, in April 1983 we published The Utilization and
Management of Information Processing Technology in California
State Government, in which we identified several specific
problems which we believed were at that time inhibiting the
state's appropriate and effective uses of IT. These problems
reflect, in many instances, planning, coordination, and oversight
problems cited in the 1967 and 1973 reports.
In our Analysis of the 1994-95 Budget Bill, we identify
several examples of current problems departments have experienced
in applying IT the most noticeable being the DMV's failure to
implement a new database system despite an expenditure of over
$40 million. (Please see our analysis of the DMV, page A54; the
Department of Social Services, page C113; the Department of
Corrections, page D55; the Board of Equalization, page H71; and
the Stephen P. Teale Data Center, page H110.)
Problems Fall Into Four Primary Categories.
The problems inhibiting the state from making the most
effective uses of IT fall into four primary categories:
* Statewide leadership.
* Statewide oversight.
* Statewide coordination.
* Effective uses of IT.
Figure 5 summarizes these problems.
___________________________________
FIGURE 5
MAJOR STATE INFORMATION TECHNOLOGY PROBLEMS
* Statewide leadership
There is no centralized, effective leadership to chart
and guide the state's course for its growing reliance
on information technology.
There is no statewide plan for information technology.
Statewide standards do not exist in specific, key
areas.
* Statewide oversight
There is a redundancy of data maintained in separate
computer systems.
Costly database management systems proliferate and are
replicated at various data centers.
Noncompatible computing systems continue to
proliferate.
* Statewide coordination
There is no centralized, effective coordination of the
state's many information technology activities.
The proliferation of separatelymaintained computer
networks continues.
There is inadequate coordination of the activities of
major data centers.
* Effective uses of information technology
Despite the expenditure of billions of dollars to
implement information technology, neither the
executive, judicial, or legislative branches of
government can easily access the mountain of data
stored in the state's computer files and convert it to
useful information.
Departments which are not sufficiently skilled in the
uses of information technology are not provided
adequate oversight, guidance or help in their efforts
to apply information technology.
There is an insufficient base of state technical staff,
and contractual efforts to supplement this staff are
inconsistent.
Governor's Executive Order Acknowledges Problems.
On May 12, 1994, the Governor issued Executive Order W8894,
establishing the Governor's Task Force on Government Technology
Policy and Procurement. According to the executive order, the
task force will, . . . conduct a review of the state's informa-
tion and technology procurement practices, and the manner in
which the state plans for, manages and oversees the development
of information systems. The task force, which will be comprised
of individuals who have experience in technology and information
system management and acquisition in private and nonprofit
organizations, is required to submit a progress report to the
Governor not later than August 15, 1994. At the time this report
was prepared, we were unable to obtain from the administration
details regarding the task force. Consequently, it is not known
when appointments will be made and when the task force will be in
a position to address the state's IT problems.
The Bottom Line: A Questionable Return on Investment.
The net effect of these problems is an annual expenditure
for IT which is not producing an optimum return on the state's
investment. In many instances, it's not even producing a
reasonably good return on the investment. As a result, funds
which could be used to develop new applications are used instead
to pay for duplication and costly implementations. The remainder
of this report discusses these problems and offers potential
solutions.
INFORMATION TECHNOLOGY LEADERSHIP
Role of the OIT Missed Opportunities
As we indicated earlier, the OIT has overall responsibility
for the application of IT in state government. When OIT was
established in 1983, the Legislature stated its intent that the
office:
* Identify new applications for IT
* Improve productivity and service to clients
* Assist agencies in designing and implementing uses of IT
* Ensure the appropriate compatibility of systems and
interchange of data and information
* Facilitate the attainment of such goals as the onetime
collection of data, the minimum duplication of records,
and the maximum availability of information at the
lowest overall cost.
* Develop plans and policies for the uses of IT as a means
of saving money, increasing worker productivity
and improving services to the public.
* Approve proposed expenditures for IT projects only if
published policies and procedures have been followed
and met.
* Develop coordinated plans and polices regarding the data
centers, IT personnel and office automation, including
the use of personal computing and electronic mail.
In our May 3, 1994 "Supplemental Analysis of the 1994-95
Budget Bill", we concluded that OIT has essentially failed to
carry out the statutory mission established for it by the Legis-
lature, our findings are summarized in the appendix to this
brief.
INFORMATION TECHNOLOGY OVERSIGHT AND COORDINATION
The rapid pace of significant developments in the IT
industry offers an increasingly varied array of products and
services to state agencies. While this array is frequently
appealing, it presents a real challenge in terms of oversight and
coordination of the state's IT activities. Addressing this
challenge is necessary in order to ensure that the state's
substantial investment in this technology provides an appropriate
return on investment and builds an increasingly solid base for
meeting the needs of state programs and their clients. In this
section we discuss areas where we believe oversight and
coordination need to be strengthened in order to ensure a good
return on investment.
PERSONAL COMPUTERS LACK OF AN INTEGRATED APPROACH
Proliferation of Personal Computer Systems.
The state has not been effective in overseeing its
implementation of PC systems, which have proliferated in the
absence of an overall plan. Consequently, many incompatible
systems have been installed over the years, often within the same
organization. This has tended to inhibit state departments from
developing an integrated approach to the management of
information. Moreover, a lack of standards has resulted in
duplication, as evidenced by departments maintaining multiple
versions of database, spreadsheet, and wordprocessing software.
This multiplicity of hardware and software systems ends up not
only costing money, but results in retraining of staff when an
organization either shifts staff to other systems or attempts to
standardize its systems. While there are good reasons for having
some diversity of equipment and software, it makes no sense to
allow such diversity to be determined in the absence of an
overall state plan which will ensure that diversity results in
costeffective computer systems.
Hidden Costs for PCs.
Another problem associated with the proliferation of PC
systems is that acquisition has often occurred on the basis of
cost estimates which do not accurately reflect the actual cost of
the systems when all relevant factors are considered. While PC
systems may be acquired based on an estimated multiyear life
cycle and to perform specific tasks, users frequently want to add
hardware and software capability, and these costs are typically
excluded from initial equipment justifications. Additional
software acquired in response to user demand not only costs more
initially, but much software has an annual license fee. Moreover,
most users want to upgrade systems to faster and more powerful
ones, thereby hastening obsolescence.
One of the most important components of PC systems,
technical support, tends to be either omitted or grossly under-
estimated when implementing these systems. Where this occurs, PC
systems either perform at a suboptimum level, or technical
resources need to be redirected to provide the necessary level of
support to ensure that the PC systems are used in the most cost-
effective manner. Either way, there is a cost to the state. When
all the foregoing factors are considered, the assumed cost of a
new PC system should be viewed as only a small part of the total,
for there will typically be significant added costs. The
magnitude of such hidden costs was explored recently by a
national consulting group, which estimated that the fiveyear cost
of a single PC, when all costs were considered, was a little over
$40,000. Although the administration has developed policy
regarding the justification and uses of PCs, it has not
adequately overseen the implementation of its policies to ensure
that PC systems are justified on the basis of complete cost
analyses. Figure 6 illustrates the hidden cost of PCs, reflecting
fiveyear costs for those categories estimated in the consultant
group's study.
_____________________________________________
FIGURE 6: INITIAL COST OF A DESKTOP COMPUTER
* Hardware Upgrades
* Software Licenses and Upgrades
* Technical Support
* Administration
* End User Operations
___________________________________________
Are Benefits Equivalent to Costs?
According to the consultant study referenced above, while
the cost of hardware has dropped steadily, the costs to support
and use these systems have increased dramatically. Given this
context, it is reasonable to ask whether the state is receiving
benefits equivalent to its multimillion dollar investment in PC
systems. Unfortunately, the answer is not known. The
administration does not track IT infrastructure investment, nor
does it assess the return on investment or determine specific
ways to reduce the cost of those aspects of desktop computing
which can be made more efficient.
Cost of Ownership Can be Reduced.
Experts believe the cost of PC systems ownership can be
reduced by 45 percent or more through techniques such as
(1)improvements in administrative management, (2)automating
certain functions, such as the help desk and software
distribution, and (3)upgrading software to make the systems
easier to use. We therefore recommend that the Legislature direct
the administration to determine specific methods for reducing the
costs associated with PC systems, and to develop and implement a
plan which will ensure that the recommended methods are followed
by state agencies.
Data Centers
In recent years the most significant growth in computing
power has occurred at the departmental level as a result of the
proliferation of PC-based systems. Despite this phenomenal
growth, state data centers remain formidable and important
components of the state's IT infrastructure. The data centers
tend to attract some of the state's best IT talent, because their
size and complexity tend to offer better promotional and career
development opportunities for IT specialists than can be obtained
in most state departments. Consequently, they represent valuable
investments of equipment and personnel resources. At the same
time, continued changes in IT maintain pressure on the data
centers to constantly reduce operating costs and offer new
services in order to remain competitive with departmental-based
systems. This situation makes it imperative that data center
activities be managed from a statewide perspective; however, such
management has not been the rule.
Data Center Activities Should Be Coordinated.
In addition to the two primary state data centers the TDC
and the HWDC mainframebased data centers are also maintained by
the Franchise Tax Board, the Public Employees' Retirement System
and the Department of Justice. At present, the only coordination
which exists among these centers is whatever level they may
choose. Consequently, there is a significant level of duplication
in terms of hardware and software acquisitions, and the hiring
and development of the highly skilled staff necessary to make the
hardware and software perform effectively. In some instances,
data centers will attempt to compete with one another by trying
to develop a service that another developed first. While
competition can produce benefits, in an era of dwindling
resources it is not good business to allow data centers to act as
totally independent entities.
Moreover, as most data centers are now linked
electronically, they should be viewed at least in part as a
statewide resource, and their activities should be coordinated in
such a manner as to ensure the minimum expenditure of resources
to establish a necessary service. For example, the TDC should not
be allowed to establish a computerbased kiosk type program if the
HWDC has already invested state funds to develop such a program.
Similarly, the HWDC need not have investigated establishing a
Geographic Information System (GIS) service a few years ago,
shortly after the TDC had expended significant effort and funds
to establish a statewide GIS service. Competition can be healthy,
but it should be conducted based on a statewide, coordinated
perspective.
In Which Directions Should Data Centers Grow?
In our recent analysis of the TDC, we point out that the
data center has focused on mainframe computer growth, and we
question whether such continued emphasis is in the state's long
term interests. A review of national trends regarding the use of
mainframe computers indicates that this technology is in a period
of transition. While experts disagree as to precisely how this
transition will play out, and predictions of the death of the
mainframe have proved to be premature, there is general agreement
that the role of mainframes is changing, and that the future is
somewhat uncertain. Given the state's very heavy reliance on
mainframe computers, especially at the state's two primary data
centers the HWDC and the TDC it would be preferable to have some
external review of data center growth plans, and an overall plan,
to ensure that the directions they take are consistent with the
state's long term interests. At present, there is no meaningful
oversight of data center directions in a statewide context, as
there is no statewide plan for the data centers.
Therefore, we recommend enactment of legislation regarding
the coordination of data center activities to specifically
include a requirement that (1)data center activities are
coordinated in a manner which will minimize the duplication
of effort and expenditure of resources and (2)the
administration develop a plan to ensure that major data
center developments are consistent with the state's longterm
interests in maximizing the usefulness of its information
systems.
COMMUNICATIONS ISSUES
Too Many Independent Networks.
The proliferation of separate departmental communications
networks has been a longstanding issue in the state's uses of IT.
When we first raised this issue in 1967, we were concerned about
centralized coordination of networks proposed by individual
departments. Now there is a situation where not only has the
total number of independent networks increased, but a number of
departments maintain or buy services from several networks (for
example, the Department of Education and the TDC). While in some
instances it may make sense to use more than one network, there
is still no effective central coordination to minimize the growth
in separate networks or contracts with network service providers.
Nor is there any meaningful central plan to consolidate the many
networks which have sprung up in order to avoid costly
duplication. While the state has invested heavily in CALNET,
there is no aggressive program or incentives to get state
agencies to move their data communications to CALNET. It makes no
sense to invest heavily in a system with statewide capability and
continue to invest in many separate systems at the same time,
with no firm plan to consolidate these systems. On that basis, we
recommend that the Legislature direct the administration to
develop a plan to reduce the number of separately maintained
communications networks to eliminate duplication and achieve net
savings.
Communications Management Itself Should Be Consolidated.
Currently, the responsibility for oversight of IT and
communications is split statutorily between the DOF and the DGS,
respectively. While this distribution of authority was appropri-
ate when IT and communications technologies were more separate,
it should be reexamined in view of the continuing integration of
these two technologies. The OIT has acknowledged that this is an
issue, but has not taken effective action to resolve it. The
increasing critical nature of telecommunications, and the merging
of communications and information technologies, argue strongly
for a consolidated state approach. For this reason, we recommend
the enactment of legislation to consolidate oversight of state
telecommunications with the oversight of the state's information
technology programs.
ELECTRONIC MAIL (EMAIL)
E-Mail Systems Should be Linked.
While many state agencies make extensive use of e-mail
systems to improve communication and worker effectiveness, there
has been no overall plan for such systems. Consequently, many
agencies maintain separate systems, some of which are not able to
communicate effectively with the larger body of email users.
Considering the power of email to eliminate telephone tag and
thereby speed the delivery of information, and that over 40,000
state employees are accessible through the most widely used email
system, including key staff in many agencies, it only makes sense
to facilitate communication on a statewide basis, so that anyone
with a desktop computer or terminal can communicate
electronically with anyone else having an email access. We
believe that there is also merit to facilitating electronic
communication with local agencies, some of which are already
linked to state networks.
No Consistency in Individual E-Mail Address Codes and Addressing
Schemes.
Because state e-mail systems are currently maintained
independently of one another, there is no consistency in the
identification codes (that is, their electronic address)
necessary to send someone e-mail. This frequently results in
time-wasting efforts to determine someone's e-mail address.
Inconsistencies in addressing schemes can make it extremely
difficult to pass e-mail messages from one system to another.
Moreover, these inconsistencies are not easily or inexpensively
reconciled. To address the individual code problem, some
entities, such as the HWDC , have established a consistent method
for users of their system which enables someone outside of the
HWDC to determine the proper code if they know how to spell the
name of the person they wish to communicate with. Other
organizations have no standards and their users have codes which
are difficult to use even when they are known. We believe that
there needs to be a uniform method for assigning e-mail
addresses, and also a uniform and simple method for looking them
up electronically when that is necessary (for example, when the
name of the person might be known, but the code for the destina-
tion computer is not). There also needs to be an effort to
resolve the problems associated with inconsistent addressing
schemes, because the present situation prevents the effective use
of the state's e-mail network.
Better E-mail Framework Needed.
Finally, we believe that it is important for the state to
plan now to replace the electronic mail system which is the
primary e-mail system used by state workers. Replacement is
important, because the current system is becoming increasingly
obsolete when compared to modern e-mail systems. As such, it is
not a good foundation on which to build a statewide e-mail
capability to improve efficiency and resolve current problems
impeding electronic communication.
In view of the problems identified above, we recommend the
enactment of legislation to require the administration to
develop a plan which will (1)allow the state's various
electronic mail systems to be linked in a manner which will
facilitate statewide electronic communication among all
levels of government, (2)provide for uniformity in the
individual codes assigned to persons using electronic mail
and an effective method for looking up electronicmail ad-
dresses, and (3)determine an electronic mail approach for
the state which will provide for the replacement of the
system currently in primary use.
STATE EXPERTISE NOT FULLY UTILIZED
A Helpful Forum Eliminated.
In establishing the statutory authority regarding the
management of IT, the Legislature authorized the creation of a
committee of user departments to assist the OIT in fulfilling its
leadership and coordinative responsibilities. This body, known as
the California Forum on Information Technology (CFIT), met
periodically and helped develop some meaningful programs to
further the effective uses of IT. These programs included the
establishment of the Data Processing Managers Training Academy to
develop future highlevel IT managers, and the Executive Institute
where state program and IT managers could gather to exchange
ideas and learn about developments and trends in the world of IT.
In 1993, the CFIT was abandoned based on the DOF's determination
that it was no longer needed.
New Body Meets but Has Yet to Deliver a Product.
At about the time the fate of the CFIT was being determined,
the Governor issued Executive Order W3792 establishing the
MultiAgency Information Management Authority (MAIMA), based on a
recommendation from the Director of the OIT, to:
* Provide leadership for the identification, initiation and
implementation of information management capabilities
that are of a multiagency scope.
* Assess the merits of potential multiagency projects and
initiate those that promise the greatest cost savings
and operational efficiency.
* Identify resources to be contributed to multiagency
projects and assign responsibility for the successful
completion of projects.
* Assign responsibility for implementing information systems
established as the result of multiagency projects.
The MAIMA, comprised of the undersecretaries of all cabinet
level agencies and the Chief Deputy Director of the Governor's
Office of Planning and Research, was required by the Governor's
Executive Order to report its assessment of the Executive Order,
including recommendations, in October 1993. To date, however, the
MAIMA has produced no report, and meetings have been infrequent.
A Meaningful Advisory Group Needs to be Re-established.
In retrospect, the administration's decision to eliminate
the CFIT has not been beneficial, because no meaningful
substitute capability was established in its place. Clearly, the
state needs to improve its IT infrastructure, and a committee of
state technology experts can provide valuable assistance in this
regard.
Consequently, to ensure that significant statewide IT
problems are addressed we recommend that the administration
establish an information technology advisory group to assist
in the identification and resolution of significant problems
inhibiting the state's costeffective application of
information technology. We recommend that this committee be
comprised of representatives of various state agencies, with
the chairperson selected by the committee.
ENABLING DEPARTMENTS TO USE INFORMATION
TECHNOLOGY EFFECTIVELY
Departmental Problems In Implementing IT Projects.
The extent to which the state receives a return on its $1
billion dollar plus annual investment in IT is related directly
to the ability of state agencies to apply the technology in
carrying out their programs. This ability varies widely among
state agencies. At one extreme are those agencies possessing
considerable IT expertise and resources for its application, and
at the other extreme are those agencies which either lack IT
expertise or resources, or both. While agencies with expertise
and resources are generally in a better position to apply IT
effectively, even they can experience problems.
The types of problems experienced by departments are not new.
Figure 7 portrays the typical characteristics of troubled IT
projects which have come to be well-known in both the private and
public sectors.
___________________________________________
FIGURE 7:
TYPICAL CHARACTERISTICS OF IT PROJECTS IN TROUBLE
* Major cost overruns.
* Significant implementation delays.
* Solutions which do not work as effectively as
planned.
_______________________________________________
Based on our review, Figure 8 lists relatively recent and
current projects which we have identified as experiencing one or
more of the characteristics of projects which have had
significant problems. The figure shows that the problems are
found in a broad range of departments, both those with and
without significant IT expertise.
___________________________________________
FIGURE 8
PROJECTS WHICH HAVE EXPERIENCED SIGNIFICANT PROBLEMS
* Department of Motor Vehicles Database Re-design -
$40 million spent and little to show.
* Department of Corrections Corrections Management
Information System - Continued schedule slippage
and cost increases ($101million is the latest
estimate of project cost).
* Department of Social Services Statewide Automated
Welfare System - Cost increases, delay and reduced
net benefits (project cost now estimated at $800
million, to be implemented over 12 years).
* Department of Social Services Child Welfare System -
Three years behind schedule with implementation
difficulties anticipated to result in a change in
project scope and/or a significant cost increase.
*Department of Social Services Statewide Automated
Child Support System - Cost increase (from
$140.8million to $152.2million) and significant
schedule slippage.
* Student Aid Commission Financial Aid Processing
System - Cost increases and contract management
problems.
* Board of Equalization Conversion to State Data Center
- Cost increases and delays.
* Department of Health Services Vital Records
Improvement Project - Implementation delays
related in part to cost concerns.
* Secretary of State Imaging Technology - New system
failed and was abandoned.
* Department of Housing and Community
Development Mobile Home Registration and
Titling - Repeated difficulties over several years
in efforts to implement an effective system.
* Department of Transportation New Database Structure -
Delays and difficulties implementing a new
database structure for departmental applications.
______________________________________________
Why Do Departments Have Problems Implementing IT Projects
We believe there are several basic reasons why departments
frequently experience difficulties in implementing computer-based
systems, and that unless the approach to IT projects is changed,
departments can expect to experience the same costly problems
repeatedly. Projects fail to adhere to original cost and schedule
estimates, or fail to perform as anticipated, due to a lack of
one or more of the following:
* Sufficiently experienced project managers
* An adequate base of skilled technical staff
* Qualified contract managers
* The ability to determine the most costeffective and/or
practical solution
* The ability to select qualified contractors.
Lack of Sufficiently Experienced Project Management Staff.
Most of the IT projects listed in Figure 8 are fairly
complex undertakings which require experienced project managers.
Yet, we believe there is a general lack of state staff with
demonstrated project management skills for large, complex
systems, especially regarding the completion of major projects on
time and within budget. This problem is made worse when a
department is attempting to implement a major system, but does
not have an adequate base of skilled technical staff to perform
the systems analysis, design, coding and testing necessary before
a new system can be put into production.
Limited Technical Capability.
In some instances, neither the department proposing an IT
project, nor the OIT staff reviewing the proposal, has sufficient
technical ability to ensure that the selected implementation
alternative is in fact the most feasible and costeffective
solution. This generally results from the fact that applying IT
to meet state needs is sufficiently complex to make it difficult
for state agencies to maintain a level of expertise which will
tend to ensure a high degree of success. The long list of
troubled and failed projects over the past 20 years offers mute
testimony to this fact.
Lack of Skilled Contract Management.
The inability of governmental agencies to properly manage
contracts with the private sector is a universal problem. Where
contract management is not adequate, cost overruns and schedule
delays can be more frequent, with the state generally ending up
paying more than it would have had an experienced contract
management team been assigned the project. Although the SAM
requires that contract managers be experienced, there is no teeth
to this requirement, and the state has paid the price for this
oversight. One method of alleviating this situation is to have
less skilled contract managers work with the more experienced
contract managers on major projects.
Difficulties in Determining CostEffectiveness and the Practical
Solution.
Under current practice, an IT project is approved on the
basis of a feasibility study report (FSR) which, while addressing
the feasibility of alternatives, does not address the actual
ability of the department to implement the project successfully.
Moreover, FSRs do not necessarily result in the identification of
the most feasible and costeffective solution.
Difficulties in Selecting a Qualified Contractor.
An additional significant problem occurs when a department
realizes that it does not have sufficient expertise to ensure the
successful implementation of an IT project, but it is ill-
equipped to select a qualified contractor to provide assistance.
State records of actual contractor performance on state IT
projects, both superior and deficient, are virtually nonexistent.
That poor and exemplary performance are not documented and made
available to all agencies makes it more difficult for a
department to select a truly qualified contractor to perform its
critical IT work.
How to Increase the Likelihood of Project Success
Figure 9 summarizes the methods we believe can help to
increase the chances that the state will achieve a higher level
of success in its efforts to apply IT.
_____________________________________________
FIGURE 9:
HOW TO ENSURE BETTER INFORMATION TECHNOLOGY PROJECT
SUCCESS
* Certify departments as to their ability to implement
a proposed IT project.
* Train and certify project and contract managers.
* Document poor and superior contractor performance and
maintain a central, online computer file regarding
such contracts accessible by state agencies.
* Require independent, qualified review of complex
projects, including FSR review, bid proposals, the
selection of contractors, and project and contract
management assistance.
* Require that FSRs identify the need for outside
assistance and include the associated costs in the
FSR.
* Where practical, fund major projects only through a
pilot or prototype phase, so that full
implementation costs, schedules and benefits can
be more accurately projected.
________________________________________________
Certify Departments.
Departments should not be allowed to embark on the
implementation of an IT project when there is no assurance that
they have the ability to ensure a successful implementation. It
makes no sense for the OIT to approve an FSR and then leave to
chance the matter of whether the project will be successful. It
does make sense for the OIT to ensure that a department is
properly equipped, whether with its own staff or outside staff,
or a combination, such that project success can be better
assured. Certifying departments as to their ability to implement
a project should end the current practice of allowing departments
to put themselves at risk when implementing projects.
Certification could be done by the OIT, a committee of state
experts, or an outside contractor.
Identify Need for Outside Assistance.
Whether they admit it or not, departments often need
assistance in the areas of selecting the best technical solution,
writing technical specifications, hiring contractors, and
managing projects and contractors. As noted above, the OIT cannot
be expected to meet these needs with its own staff. Consequently,
there should be an upfront acknowledgement in the FSR of the need
for outside assistance, with an identification of the associated
costs and a mechanism should be put in place to ensure that such
assistance can be obtained. One way to facilitate providing such
assistance is through the establishment of master agreements with
qualified outside contractors.
Certify Project and Contract Managers.
No significant state IT project should be implemented
without proper project and contract management oversight, and
state managers should be trained and certified before being
assigned these tasks.
Identify Poor and Superior Performing Contractors.
Although current law provides for identifying contractors
who fail to meet contractual requirements, state agencies seldom
volunteer documentation pertaining to poor contractor
performance. Both poor and exemplary performance by any
contractor should be documented and made available to any agency
which needs contractor assistance, so that an agency can be
informed as to a potential contractor's performance record. This
information should be maintained in a computerbased file, so as
to facilitate access and reduce manual search efforts.
Do Pilot Testing.
Given the frequent underestimating of project costs,
schedules and benefits, it would be preferable to have a process
which favors pilot or prototype projects, thereby enabling
refinement of full implementation costs, schedule and benefits.
This approach, with appropriately designed and conducted
evaluations, should help to end the repeated instances of
projects which are sold to the Legislature on the basis of an
economic justification, but which change quickly and drastically
as implementation occurs.
In view of the above, we recommend that the Legislature
direct the administration to develop and implement a policy
which will provide for:
(1)certification of departments as to their ability to
implement IT technology,
(2)training and certification of project and contract
managers,
(3)documentation and reporting of poor contractor
performance, and maintenance of this information
on a computer-based file made accessible to state
agencies,
(4)requiring that Feasibility Study Reports identify
and include costs for any needed outside
assistance,
(5)requiring that, wherever possible, project cost,
schedule and benefit estimates be refined through
a pilot or prototype project, and
(6)ensure that departments are able to obtain competent
outside assistance wherever needed to implement IT
projects successfully.
OTHER INFORMATION TECHNOLOGY ISSUES
Facilitating Legislative Review of Information Technology
Projects
At present, FSRs and Special Project Reports (SPRs)
concerning major IT projects generally do not facilitate legisla-
tive review because important cost, benefit, and schedule
information tend to be buried in these documents. In addition,
information regarding problems with the project is often
presented in such vague and general terms that the reader has to
read between the lines to get any indication that a project is
experiencing difficulties. The problem is especially acute
regarding SPRs, which are supposed to inform the reader as to
significant changes in approved projects. Instead of providing an
up front summary, in a consistent format for all SPRs, critical
information is often spread throughout the document, or provided
in such vague terms that it is difficult, if not impossible for a
reviewer to draw an accurate conclusion regarding the status of
the project.
Legislative review of FSRs and SPRs would be greatly
facilitated if these documents had concise tables in consistent
formats placed at the front of the document. In the case of an
FSR, the costs, benefits and schedule of considered alternatives
should be displayed, with the recommended alternative clearly
identified. For SPRs, the summary should disclose, in a chart
format, the original project cost, benefits and schedule, the
last revisions to each of these, and the latest revision proposed
in the SPR. Also, the SPR needs to clearly identify and discuss
current problems encountered by the project. Government Code
Section 11734 requires the Director of the OIT to establish
procedures facilitating legislative review of state projects.
Consequently, we recommend that the administration modify
Feasibility Study Report and Special Project Report formats to
contain consistent formats, with original and revised cost,
benefit, and schedule summaries at the front of these documents,
so as to facilitate review and oversight.
Procurement Practices
State Can Better Leverage Its Purchasing Power. State
agencies spend significant sums of money annually to acquire IT
equipment, services and software. Our review of state IT
practices indicates that there is significant room for
improvement in terms of leveraging the state's purchasing power
so as to reduce the total amount spent by various agencies for
the same type of good or service, thereby maximizing the return
on the state's total investment. For example, until the HWDC
negotiated a master agreement for services provided by a major
consulting group, departments within the Health and Welfare
Agency had negotiated their own, separate contracts. By
negotiating one master agreement, the HWDC was able to obtain a
better return on the agency's investment (that is, more
services). We believe that this kind of procurement should be
looked at from a statewide perspective, as other departments
outside of the Health and Welfare Agency continue to buy the same
equipment services and software on an individual departmental
basis. We believe improved coordination can be accomplished
without slowing the acquisition of goods or services appreciably,
particularly if the administration establishes a computer-based
method of coordination.
Consequently, we recommend that the Legislature direct the
administration to establish a process to ensure that state
information technology equipment, services and software
purchases are coordinated so as to maximize the return on
the state's expenditures.
New Method for Resolving Procurement Protests Is Needed.
IT-related procurements tend to be time-consuming and
costly, and often result in formal protests filed by competing
vendors. According to information provided by the DGS, thirty-six
formal protests were filed in the 18-month period from June 1,
1992 through November 15, 1993. It took an average of days to
resolve protests filed in 199293, and 76 days for protests filed
in the first six months of 199394. Protests not only consume time
and money, they also deny a department the goods or services it
needs to perform its work.
The current method of resolving procurement protests
includes a hearing before an administrative hearing officer, who
then makes a recommendation to the Board of Control for final
decision. Two of the three departments which have seats on the
Board of Control, the DOF, and the DGS, are the same departments
which approve procurements resulting in a formal protest. We
believe that it is inappropriate to ask departments which approve
an action to rule on a protest which challenges that action. A
more appropriate method of resolving protests would be to
establish a process whereby bidders could elect to submit a
dispute to binding arbitration, conducted by a hearing officer or
panel mutually acceptable to the state and the bidders. To
discourage frivolous protests, which may occur now in some
instances, it would be reasonable to require that a losing party
pay the cost of the arbitration. A bidder who did not elect
arbitration could pursue the matter in the courts, an option they
currently have, but under present policy can do so only after
they have gone through the current protest process, which,
according to DGS data, they are likely to lose.
Consequently, we recommend the enactment of legislation to
establish a new process for resolving formal protests of
proposed information technology awards, to provide for
voluntary, binding, impartial arbitration or such other
method of resolving protests as the Legislature finds
preferable to the current method. We further recommend that
any new procedure specifically preclude from the
decisionmaking process any department which played a role in
approving the award being protested.
Current Procurement Practice Does Not Ensure Good Business
Decisions.
Current law requires that IT procurements be conducted so as
to provide the most costeffective solution to the state's
requirements. The law also provides that the determination of the
most costeffective solution be based on evaluation criteria which
are specified in the state's Invitation for Bid or Request for
Proposal. Administrative policy also requires that all costs to
implement a solution be considered in the evaluation. In our
199495 Analysis of the TDC, we discuss the procurement of a
mainframe computer where the data center has paid too much. Yet,
the Board of Control, responding to a formal protest by a
competing vendor, upheld the state's procurement decision,
primarily because the data center's evaluation criteria did not
include certain items of cost the protesting bidder argued should
be included.
In enacting procurement law, the Legislature never intended
that the state end up with a costlier computer solution simply
because the bid evaluation criteria failed to include certain
costs. This situation needs to be corrected to prevent the state
from making poor business decisions when procuring IT solutions.
We therefore recommend the enactment of legislation to
clarify legislative intent regarding information technology
procurements so as to ensure the most costeffective solution
to the state's requirements.
Limitation on Master Service Agreements for Technical Services
Should be Removed.
The DGS has negotiated Master Service Agreements (MSAs) with
private sector firms for technical personnel needed by state
agencies for their IT efforts. While the MSAs have been well-
received by state agencies, DGS currently limits the state's use
of a specific individual under an MSA to nine months within a 12
month period. This causes problems for state agencies which need
help extending beyond nine months.
By contrast, departments can obtain greater than nine months
of help from specific individuals through separate contracts,
such as those frequently let with consulting organizations.
Accordingly, we see no reason why the DGS should not find a means
to allow the same level of service via the MSAs.
Consequently, we recommend that the Legislature direct the
DGS to ensure that Master Service Agreements for information
technology personal services not include a ninemonth
limitation on the use of such services.
FEASIBILITY STUDY REPORTS
FSRs Should Address Future Costs.
State policy requires the approval of an FSR prior to the
implementation of an IT project. Although an FSR must provide
cost and benefit information over the life of the project for the
alternative implementation methods considered, there is no
requirement that an FSR address the future cost implications to
upgrade or replace the solution recommended by the FSR. We
believe that this is a significant oversight, because what may
appear to be the most costeffective alternative now may end up
being more costly over time. Information technology changes
continuously, and it is important that trends be considered when
evaluating alternative approaches. For example, in evaluating
which mainframe to acquire, a data center's FSR should address
the likely relative costs to upgrade or replace each mainframe
being considered, because some will provide a less costly and
more effective path to future upgrades than others. While it may
not always be possible to develop an accurate assessment of
future costs, we believe that this could be done in many
situations. The determination as to whether to assess such costs
could be made by the administration as part of the normal project
proposal process. Requiring that, where practical, FSRs address
future cost will help to ensure that the state realizes the best
return on its IT investments over the long term.
We therefore recommend that the Legislature direct the
administration to establish requirements to ensure that
Feasibility Study Reports for information technology
projects address, where practical, the future cost
implications of each alternative evaluated.
Proposed FSR Solutions Need to be Independently Validated.
A problem with the current Feasibility Study Report process
is that there is no requirement for an independent validation of
the recommended solution. In some cases, the OIT does not have
sufficient staff expertise to validate that the recommended
solution is either technically viable or the best available
solution. In many cases, the department proposing the solution is
least capable of ensuring that the best solution has been
selected. While some departments have contracted with a qualified
contractor for an independent validation of their recommended
solution, this method is not employed for all major projects. We
believe that requiring an independent validation up front, for
projects determined by the administration as warranting such
validation, while admittedly costing additional funds, will help
to prevent more costly corrective measures during the
implementation period. In this regard, an investment in an
independent validation should be viewed as a relatively minor and
reasonable insurance premium.
Consequently, we recommend that the administration adopt
procedures requiring an independent validation of
recommended solutions contained in Feasibility Study Reports
for projects determined by the administration to warrant
such validation.
________________________________________________
________________________________________________
Appendix
PROBLEMS WITH OIT'S HANDLING OF
THE STATE'S INFORMATION TECHNOLOGY
* Failure to implement information systems effectively
* Several departments have had major, costly difficulties
in implementing IT systems.
* Failure to lead the way with emerging technologies
* The impetus for applying emerging technology has come
primarily from individual departments.
* Focus on procedures and policies, while more difficult
challenges remain unresolved
* No strategic plan, with an implementation component,
for state's uses of IT.
* Standards for state's use of IT not established
* E-mail systems are disjointed.
* Costly geographic information systems are established
independently.
* No plan for the Internet
* Despite cost and security implications, state usage is
growing in the absence of a specific policy and
guidelines.
* Inadequate Access to Statewide Data
* Despite the investment of billions of dollars over the
past 20 years, neither the executive, judicial, or
legislative branches of government are able to tap into
a state corporate data base of information via a
computer system.
* Inadequate Oversight of Projects
* OIT approved the Teale Data Center's acquisition of
computing equipment to support a major DMV project
without understanding the real costs of the project to
the data center and its other clients.
* Lack of Assistance in Finding Solutions
* Departments failing to obtain OIT approval for a proposed
technological solution to their needs are frequently
left with no solution because OIT does not help to
implement solutions.
* Oversight Requires Intervention
* OIT has not intervened in questionable information
technology activities it is aware of because
intervention was not specifically provided for in
policies it established.
[END]
.