oversight

Need To Evaluate and Improve Postal Source Data System Before Further Expansion

Published by the Government Accountability Office on 1971-07-01.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

REPORT TO THE CONGRESS




Need To Evaluate And Improve
Postal Source Data System
Before Further Expansion1148,,,,
Post Office Department   ,




            t9 ?*4


BY THE COMPTROLLER GENERAL
OF THE UNITED STATES


                               JULLE
                                  CY1,COMP
                                       19 7 1
        FILE COPY - COMP GEN
       COMPTROLLER GENERAL OF THE UNITED STATES
                  WASHINGTON. D.C.   Z0548




B-114874




To the President of the Senate and the
Speaker of the House of Representatives

      This is our report on the need of the Post Office
Department to evaluate and improve the Postal Source
Data System before further expansion.

     Our review was made pursuant to the Budget and
Accounting Act, 1921 (31 U.S.C. 53), and the act of Sep-
tember 2, 1960 (39 U.S.C. 2206).

      Copies of this report are being sent to the Direc-
tor, Office of Management and Budget; the Postmaster
General; and the U.S. Postal Service Board of Governors.




                                        Comptroller General
                                        of the United States




            50TH ANNIVERSARY 1921-1971
.COMPTROLLER GENERAL'S              NEED TO EVALUATE AND IMPROVE POSTAL SOURCE
REPORT TO THE CONGRESS              DATA SYSTEM BEFORE FURTHER EXPANSION
                                    Post Office Department B-114874

DIGE ST

WHY THE REVIEW WAS MADE

     In 1966 the Post Office Department embarked on a multimillion-dollar
     program to install, nationwide, an automated data collection and proc-
     essing system to provide postal management with more timely and accu-
     rate information on employee time and attendance, labor-hour distribu-
     tion, and mail volumes processed than was possible under manual data
     systems.
      Substantial economies in operating costs were expected to result from
      this automation effort, called the Postal Source Data System.
     The system was to be installed initially in the 75 largest post of-
     fices. The Department later decided to expand it to include at least
     35 other large post offices.
      After the Department began installing the system, the General Account-
      ing Office (GAO) noted that implementation schedules were not being
      met, that installation costs were rapidly increasing, and that local
      postal management officials believed the information produced by the
      system would have little, if any, value. GAO noted also that the ap-
      propriations committees of the Congress were expressing interest in the
      schedule slippages and the cost and usefulness of the system.
      Therefore GAO conducted a review to ascertain (1)the effectiveness of
      the planning for the system, (2)the reasons for the rising costs of
      acquiring and operating the system, and (3)whether the system was sup-
      plying postal officials with useful data for improving postal opera-
      tions.
      This report deals with those aspects of the system for which the De-
      partment has responsibility and is not intended to be critical of the
      prime contractor or its performance under its contracts with the De-
      partment. Therefore the prime contractor was not requested to review
      or comment on the matters discussed in this report.

FINDINGS AND CONCLUSIONS

      In 1966 the Department estimated that the total acquisition cost of the
      system when fully installed would be $30.2 million and that it would

 Tear Sheet


                                                  JULY     1,1971
be fully operational in the 75 post offices by November 1968. The De-
partment predicted savings, averaging $7.2 million annually, during
each of the first 5 years of the system's operation.
Despite the lack of sufficient data to indicate the feasibility of the
proposed system and the types and quantities of equipment needed and
contrary to the recommendations of its own study groups, the Depart-
ment awarded a contract in the amount of $22.7 million for the pur-
chase and nationwide installation of equipment for the system.
The premature award of the contract resulted in the acquisition of un-
needed electronic data collection equipment costing about $1.2 million
and of other equipment that was not used for substantial periods of
time. By December 31, 1970, 31 amendments had increased the amount of
the contract to $49.7 million, more than double the original amount.
(See pp. 10 to 29.)
Acquisition costs incurred by the Department both in-house and under
contracts totaled $60.5 million by February 1971. Of this amount, at
least $44.5 million was attributed to the initially planned 75 instal-
lations--an increase of $14.3 million, or 47 percent, over the Depart-
ment's estimated costs of $30.2 million for these post offices: (See
p. 10.)
As of November 30, 1970, a fully operational system had not been im-
plemented at nine of the initial 75 post offices, including two of the
largest in the country--New York, N.Y. and Washington, D.C.--although
more than 2 years had passed beyond the original November 1968 target
date. (See pp. 7 and 10.)
Annual employee costs, as of October 1970, for authorized system posi-
tions exceeded the Department's original estimate of $5.5 million 'by
at least $14.1 million, or 256 percent. In addition, other postal em-
ployees had been diverted from their regular post office duties to op-
erate the system. The number of such employees and the related costs
were not readily available. (See pp.30 and 31.)
Automated payroll operations were more costly than under the former
manual system (see p. 31), and the reports generated by the system were
of little value to postal management. (See pp. 33 to 38.)
The Department initiated a multimillion-dollar expansion of the system
to an additional 35 post offices (see p. 39) although (1)the system
was not fully operational at the initially planned 75 post offices
(see p. 10), (2)the information being generated was of limited use to
postal management (see pp. 33 to 38), and (3)numerous problems were
being encountered with that portion of the system that had been installed
(see pp. 34 to 38).




                                 2
      Most of the deficiencies and the increased costs resulted from inade-
      quate study and insufficient testing of this complex data collection
      and processing system prior to nationwide installation. (See p. 41.)

RECOMMENDATIONS OR SUGGESTIONS

      The Postmaster General should
         --suspend the expansion program pending a comprehensive evaluation
           and cost-benefit study of the system and
         --curtail the procurement of equipment and software for the system
           and keep operational costs to a minimum pending the outcom6 of the
           evaluation and cost-benefit study. (See p. 44.)

AGENCY ACTION AND UNRESOLVED ISSUES

      The Postmaster General said that the results of GAO's review confirmed
      many of the findings previously reported by the Department's internal
      auditors and that the difficulties, delays, and added costs involved in
      establishing the system were a matter of concern to the Department.
      Although recognizing that some mistakes were made, the Postmaster Gen-
      eral disagreed that all of them could have been avoided.
      The Postmaster General stated that over the past 12 months considerable
      improvements had been introduced into the system and that these efforts,
      in his opinion, had overcome, to a significant degree, the problems
      cited by GAO.
      He stated also that the Department concurred generally with GAO's sug-
      gestion that further expansion of the system be suspended pending an eval-
      uation and cost-benefit study; however, he believed that a limited further
      expansion to the 110 post offices must be completed to avoid further ad-
      verse effects on the overall objective of the system. He advised GAO
      that commitments had been made and obligations had been incurred for most
      of the equipment involved in the limited further expansion to the 110
      post offices. (See p. 42.)

       GAO recognizes that certain improvements have been made in the system
       but believes that it is neither desirable nor a judicious investment of
       funds to expand the system to any additional post offices, including
       the planned limited expansion to 110 offices, when its effectiveness
       and economy have not been demonstrated at the facilities where it has
       been installed.

       GAO noted that, notwithstanding the Department's existing commitments
       and obligations for equipment to expand the system to the 110 post offices,

Tear Sheet



                                       3
     future expenditures of millions of dollars still could be avoided if
     the current expansion were suspended. (See pp. 43 and 44.)

MATTERS FOR CONSIDERATION BY THE CONGRESS

    This report demonstrates the need for the Post Office Department to
    critically evaluate and improve the economy and effectiveness of the
    Postal Source Data System before investing more funds to expand it to
    additional post offices.




                                   4
                       Contents


DIGEST                                                    1

CHAPTER

  1       INTRODUCTION                                    5
              Chronology of acquisition of PSDS           6
              Status of PSDS implementation               7

  2       INSTALLATION OF PSDS BEHIND SCHEDULE AND
          COSTS ARE SPIRALING UPWARD                     10
              Limited testing at Minneapolis and Mil-
                waukee post offices                '     11
              Inadequate feasibility study               13
              System specifications not prepared prior
                to selection and purchase of PSDS
                equipment                                17
              Site surveys not made prior to contract
                award                                    21
              Change from Work Measurement System to
                Work Load Recording System               25
                  Overprocurement of mail weighing
                    equipment                            26
                  Relocation of electronic scale
                    equipment                            27 /
              Underutilization of costly computer
                equipment                                28

   3      PREDICTED SAVINGS MAY NOT BE REALIZED          30

   4      LIMITED USEFULNESS OF PSDS INFORMATION         33
              Employee time and attendance reports       33
              Work Load Recording System reports         37

   5      PSDS EXPANSION AND MODIFICATION PROGRAM        39

   6      CONCLUSIONS, AGENCY COMMENTS AND OUR EVALU-
          ATION, AND RECOMMENDATIONS                     41
              Conclusions                                41
              Agency comments and our evaluation         41
              Recommendations to the Postmaster
                General                                  44
CHAPTER                                                   Page

      7    SCOPE OF REVIEW                                45

APPENDIX

       I   PSDS post offices                              49

      II   Basic components of Postal Source Data Sys-
             tem                                           50

  III      Pictures of selected PSDS equipment compo-
             nents                                         52

      IV   Letter dated September 9, 1970, from the
             Postmaster General to the General Account-
               ing Office                                 62

       V   Principal management officials of the Post
             Office Department responsible for admin-
             istration of Postal Source Data System       65

                         ABBREVIATIONS

ADP        automatic data processing

ADPC       Automatic Data Processing Center

GAO        General Accounting Office

PSDS       Postal Source Data System
COMPTROLLER GENERAL'S             NEED TO EVALUATE AND IMPROVE POSTAL SOURCE
REPORT TO THE CONGRESS            DATA SYSTEM BEFORE FURTHER EXPANSION
                                  Post Office Department B-114874

DIGEST

WHY THE REVIEW WAS MADE

    In 1966 the Post Office Department embarked on a multimillion-dollar
    program to install, nationwide, an automated data collection and proc-
    essing system to provide postal management with more timely and accu-
    rate information on employee time and attendance, labor-hour distribu-
    tion, and mail volumes processed than was possible under manual data
    systems.
    Substantial economies in operating costs were expected to result from
    this automation effort, called the Postal Source Data System.
    The system was to be installed initially in the 75 largest post of-
    fices. The Department later decided to expand it to include at least
    35 other large post offices.
    After the Department began installing the system, the General Account-
    ing Office (GAO) noted that implementation schedules were not being
    met, that installation costs were rapidly increasing, and that local
    postal management officials believed the information produced by the
    system would have little, if any, value. GAO noted also that the ap-
    propriations committees of the Congress were expressing interest in the
    schedule slippages and the cost and usefulness of the system.
    Therefore GAO conducted a review to ascertain (1)the effectiveness of
    the planning for the system, (2)the reasons for the rising costs of
    acquiring and operating the system, and (3)whether the system was sup-
    plying postal officials with useful data for improving postal opera-
    tions.
    This report deals with those aspects of the system for which the De-
    partment has responsibility and is not intended to be critical of the
    prime contractor or its performance under its contracts with the De-
    partment. Therefore the prime contractor was not requested to review
    or comment on the matters discussed in this report.

FINDINGS AND CONCLUSIONS

    In 1966 the Department estimated that the total acquisition cost of the
    system when fully installed would be $30.2 million and that it would
be fully operational in the 75 post offices by November 1968. The De-
partment predicted savings, averaging $7.2 million annually, during
each of the first 5 years of the system's operation.
Despite the lack of sufficient data to indicate the feasibility of the
proposed system and the types and quantities of equipment needed and
contrary to the recommendations of its own study groups, the Depart-
ment awarded a contract in the amount of $22.7 million for the pur-
chase and nationwide installation of equipment for the system.
The premature award of the contract resulted in the acquisition of un-
needed electronic data collection equipment costing about $1.2 million
and of other equipment that was not used for substantial periods of
time. By December 31, 1970, 31 amendments had increased the amount of
the contract to $49.7 million, more than double the original amount.
(See pp. 10 to 29.)
Acquisition costs incurred by the Department both in-house and under
contracts totaled $60.5 million by February 1971. Of this amount, at
least $44.5 million was attributed to the initially planned 75 instal-
lations--an increase of $14.3 million, or 47 percent, over the Depart-
ment's estimated costs of $30.2 million for these post offices. (See
p. 10.)
As of November 30, 1970, a fully operational system had not been im-
plemented at nine of the initial 75 post offices, including two of the
largest in the country--New York, N.Y. and Washington, D.C.--although
more than 2 years had passed beyond the original November 1968 target
date. (See pp. 7 and 10.)
Annual employee costs, as of October 1970, for authorized system posi-
tions exceeded the Department's original estimate of $5.5 million by
at least $14.1 million, or 256 percent. In addition, other postal em-
ployees had been diverted from their regular post office duties to op-
erate the system. The number of such employees and the related costs
were not readily available. (See pp.30 and 31.)
Automated payroll operations were more costly than under the former
manual system (see p. 31), and the reports generated by the system were
of little value to postal management. (See pp. 33 to 38.)
The Department initiated a multimillion-dollar expansion of the system
to an additional 35 post offices (see p. 39) although (1)the system
was not fully operational at the initially planned 75 post offices
(see p. 10), (2)the information being generated was of limited use to
postal management (see pp. 33 to 38), and (3)numerous problems were
being encountered with that portion of the system that had been installed
(see pp. 34 to 38).
    Most of the deficiencies and the increased costs resulted from inade-
    quate study and insufficient testing of this complex data collection
    and processing system prior to nationwide installation. (See p. 41.)

RECOMVENDATIONS OR SUGGESTIONS

     The Postmaster General should
       --suspend the expansion program pending a comprehensive evaluation
         and cost-benefit study of the system and
       --curtail the procurement of equipment and software for the system
         and keep operational costs to a minimum pending the outcome of the
         evaluation and cost-benefit study. (See p. 44.)

AGENCY ACTION AND UNRESOLVED ISSUES

    The Postmaster General said that the results of GAO's review confirmed
    many of the findings previously reported by the Department's internal
    auditors and that the difficulties, delays, and added costs involved in
    establishing the system were a matter of concern to the Department.
    Although recognizing that some mistakes were made, the Postmaster Gen-
    eral disagreed that all of them could have been avoided.
     The Postmaster General stated that over the past 12 months considerable
     improvements had been introduced into the system and that these efforts,
     in his opinion, had overcome, to a significant degree, the problems
     cited by GAO.
     He stated also that the Department concurred generally with GAO's sug-
     gestion that further expansion of the system be suspended pending an eval-
     uation and cost-benefit study; however, he believed that a limited further
     expansion to the 110 post offices must be completed to avoid further ad-
     verse effects on the overall objective of the system. He advised GAO
     that commitments had been made and obligations had been incurred for most
     of the equipment involved in the limited further expansion to the 110
     post offices. (See p. 42.)
     GAO recognizes that certain improvements have been made in the system
     but believes that it is neither desirable nor a judicious investment of
     funds to expand the system to any additional post offices, including
     the planned limited expansion to 110 offices, when its effectiveness
     and economy have not been demonstrated at the facilities where it has
     been installed.
     GAO noted that, notwithstanding the Department's existing commitments
     and obligations for equipment to expand the system to the 110 post offices,




                                      3
     future expenditures of millions of dollars still could be avoided if
     the current expansion were suspended. (See pp. 43 and 44.)

MATTERS FOR CONSIDERATION BY THE CONGRESS

     This report demonstrates the need for the Post Office Department to
     critically evaluate and improve the economy and effectiveness of the
     Postal Source Data System before investing more funds to expand it to
     additional post offices.




                                    4
                         CHAPTER 1

                       INTRODUCTION

     The Postal Source Data System (PSDS) was designed to
collect data at its source in selected post offices concern-
ing (1) employee time and attendance, (2) labor-hour dis-
tribution, and (3) mail volume, and to forward this data to
centralized high-speed computers for processing. The pro-
cessed data was to be electronically transmitted to the
originating post offices and other organizations within the
postal service in the form of reports to be used as aids to
management. PSDS has been described by the Post Office De-
partment as being the largest system of its kind ever devel-
oped and an important first step in the Department's devel-
opment of a Postal Management Information System.

     PSDS was to be initially installed in the 75 largest
post offices in the continental United States, involving
about 60 percent of the total mail volume and 60 percent of
the total number of post office employees, and was to be
fully operational in these 75 offices by November 1968.
PSDS was subsequently expanded to include an additional 35
large post offices. Although PSDS has become fully opera-
tional in some of the 35 expansion offices, it has yet to
be implemented in several of the initial 75 offices. We
have identified PSDS acquisition costs of about $60 million
as of February 1971. The 110 post offices currently desig-
nated forPSDS installation are listed in appendix I.

     Source data entering PSDS through electronic input de-
vices at individual post offices is transmitted over long-
distance telephone lines to computers at one of five collec-
tion centers (teleconcentratQr sites), in Paramus, New
Jersey; Washington, D.C.; Chicago, Illinois; San Francisco,
California; and Oklahoma City, Oklahoma. The computers at
these sites continuously collect data from the post offices
they serve and forward the data automatically to large-
scale computers at one of two automatic data processing
centers (ADPCs) in either Wilkes-Barre, Pennsylvania (this
center was formerly located in Paramus) or St. Louis,
Missouri. The ADPCs, :after processing the data, transmit it
to high-speed printers in the post offices to provide local
postal management with print-out reports on labor-hours
worked and mail volumes.
                              5
     The flow of data within the system is illustrated on
page 8. A map showing the PSDS communications network de-
sign, as of August 1970, is on page 9.

     The more than 12,000 input, processing, and output de-
vices that constitute PSDS are connected by over 115,000
miles of transmission lines leased from telephone companies.
Alternate lines and dual equipment have been designed into
PSDS to provide protection against possible loss of data
due to equipment or transmission line failure.

     The basic components of PSDS are described in appen-
dix II and pictures of selected PSDS equipment components
are included in appendix III. The pictures in appendix III,
pages 53 to 57 were obtained from the Post Office Department;
the remaining pictures were taken by our representatives
during our field review.

CHRONOLOGY OF ACQUISITION OF PSDS

     Between 1962 and 1965 a prototype source data collec-
tion system was tested in the Minneapolis, Minnesota, and
Milwaukee, Wisconsin, post offices. On the basis of these
tests, a feasibility study was made and completed in 1965.
The study report concluded that a nationwide source data
collection system was feasible, that the cost of the system--
estimated at $33 million--could be completely amortized in
the first 5 years of operation, and that the system would
provide the Department with savings of $36 million in the
same period. On the basis of this study, a line item of
$33 million was included in the Department's budget for fis-
cal year 1967 for acquisition of such a system.

     In January 1966 a request for proposals for a Postal
Source Data System was released to 59 vendors. Five of the
59 vendors responded to the request and four submitted re-
sponsive proposals. After a period of evaluation, a success-
ful vendor was selected, and on June 30, 1966, the Depart-
ment entered into a negotiated contract with the successful
vendor (hereinafter referred to as the prime contractor),
in the amount of about $23 million (subsequently increased
to about $50 million) for the acquisition and installation
of PSDS as a fully operational system in the 75 (now 110)
largest post offices.

                             6
     This report deals with those aspects of PSDS for which
the Department, rather than the prime contractor, had full
or primary responsibility and is not intended to be in any
way critical of the prime contractor.

STATUS OF PSDS IMPLEMENTATION

     As of November 30, 1970, all three major system
applications--employee time and attendance, labor-hour dis-
tribution, and mail volume--were operational at 86 post of-
fices, including 66 of the initial 75 largest offices, and
the employee time and attendance application only was oper-
ational at 12 additional post offices. The Department ex-
pects the three major system applications to be fully opera-
tional at the 110 post offices in PSDS by August 1971.




                                7
           o                                     t.i

                    IL




     O:E
     LLI                                                                                  -


                                                                           0




                                                                           ao
     4<             U    _                                ELU



                                                                           "',v


                                                                           ut
                                                                           U.
               · mii     U                                                 ir
                                                                           U.~
                                                                ,T                       z


                                                                     /zz
                                                                 ESI
                                                                  r~~"~    O


           Q   O                                                           v,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~




                                O   s164531~~~~-$i~~~81
                                             m                                           z~
                         gow




dB                 ~~~~~~~~~~
      U-~    ~     t   ·

'4




CD
CL




CD

I-

2c

I-4




U                             3     Y
L&I
             I-a
            '4                                                         1
C-D          '4~       ~           ~     .i
            I-~ ~ ~   ~ ~ ~~~:
3E
CD             CD~   ~          ~~~~                       ~       P
                    SM~~~~~~~~~~~~~~~~~~
                          u ~~~~~~~~~~~~~~~~~~~~~~~~~~~s
                                       ~    ~ ~     ~      ~   ~   ~
               CD~~P
'4
                                         ~5
'43
Ca

CZ

CD
C-
                                   5              :=,~

'4
CO)



CD
                        CHAPTER 2

          INSTALLATION OF PSDS BEHIND SCHEDULE
             AND COSTS ARE SPIRALING UPWARD

     In February 1966 when justifying the need for funds for
PSDS before the Treasury, Post Office and Executive Office
Subcommittee, House Committee on Appropriations, Department
officials stated that (1) the total cost of PSDS would be
about $33 million (equipment, software, and installation,
$30.2 million; first year's supplies, communication line
rental, and maintenance, $2.8 million) and (2) the employee
time and attendance, labor-hour distribution, and mail vol-
ume recording functions of PSDS would be fully installed in
the 75 largest post offices by November 1968.

     By February 1971, more than 2 years after the target
date for a fully operational system in the initial 75 post
offices, PSDS acquisition costs related to the initial 75
post offices had risen to at least $44.5 million (an in-
crease of $14.3 million or 47 percent); but PSDS was not
operational in several of these post offices including those
in New York City and Washington.

     Also by February 1971, PSDS acquisition costs--both
in-house and under contracts--at 110 post offices (the ini-
tial 75 plus 35 expansion offices) totaled about $60.5 mil-
lion. The target date for full installation of the three
planned applications of PSDS at the last of the initial 75
largest offices and the last of the 35 expansion offices is
August 1971, more than 2-1/2 years beyond the initial target
date of November 1968 for having PSDS fully operational.

     We believe that the Department's decision to award
contracts for the purchase and installation of equipment
for PSDS in 75 post offices was premature and resulted in
excessive costs. In our opinion, most of the deficiencies
and spiraling costs associated with the installation of the
system resulted from inadequate study and testing before
nationwide installation. At the time the initial contract
for the acquisition and installation of the PSDS was
awarded:



                             10
     1. Only limited testing had been made of a prototype
        source data collection system and that system was
        significantly different from the one contracted.

     2. A study made by the Department prior to the award
        of the contract had not demonstrated adequately the
        feasibility of the proposed nationwide source data
        collection system.

     3. Sufficiently detailed system specifications had not
        been prepared.

     4. Site surveys to enable the Department to make real-
        istic estimates of such items as equipment and space
        needs had not been made at most of the 75 post of-
        fices where the system was to be installed.

Delays and added costs resulted also from various changes
in specifications and procedures for a Work Load Recording
System encompassing the labor-hour distribution and mail
volume system applications.

LIMITED TESTING AT MINNEAPOLIS AND
MILWAUKEE POST OFFICES

     In September 1962 the Department began testing a proto-
type automated source data collection system at the Minne-
apolis Post Office. In June 1964 system testing was ex-
panded to the Milwaukee Post Office. This system was de-
signed to capture (on punched paper tape in a form suitable
for computer processing) payroll data, labor-hours used,
mail volume recordings, and motor vehicle statistics at post
office work locations. The information on the tape was to
be transmitted over a communication line to a central com-
puter in the Minneapolis Postal Data Center for processing.
This prototype system was significantly different from the
PSDS the Department contracted for in 1966 (see p. 15 for
a description of the system acquired).

     Department records indicated that in a January 1965
meeting a Department study team, engaged in reviewing the
source data collection system, expressed concern that al-
though only timekeeping functions were being performed on
the system, computer capacity already had been reached. A
Bureau of Finance (currently, Finance and Administration
Department) representative at the meeting explained that
such a condition was due to a temporary peak work load
which would level off, but stated that he recognized that
larger equipment would be mandatory in the next 12 to 18
months. This official explained also that a decision to
expand the system would be made only when the current sys-
tem had proven itself; was fully warranted; and had demon-
strated that it would "pay off."

     Also in January 1965, a postal inspection report stated
that the source data collection system in Milwaukee and
Minneapolis could not be justified if used for timekeeping
purposes only and that the results obtained for payroll
purposes definitely did not meet acceptable standards. The
inspectors pointed out that, although experiments had been
conducted in recording payroll, mail volume, and motor ve-
hicle information, the recording of information in the two
latter areas had been suspended and the system was being
used for timekeeping purposes only.

     The inspectors stated also that no attempt had been
made by the Department to evaluate the cost of the system,
and they recommended that (1) an analysis be made comparing
manual system costs with automated system costs and (2) un-
til the system tests satisfactorily demonstrated that at
least one part of the mission of the system could be accom-
plished, experimentation and development should be confined
to Minneapolis and Milwaukee.

     A June 1965 Department study described this automated
system as one which provided a more comprehensive data-
gathering method than the then-existing manual system but
which fell short of requirements for a management informa-
tion system.

     We did not find any evidence to show that the compara-
tive cost analysis recommended by the postal inspectors was
made or that the source data collection system was satisfac-
torily tested at Minneapolis and Milwaukee.




                             12
INADEQUATE FEASIBILITY STUDY

     Although a Department study team and postal inspectors
had recommended that the source data collection system be
confined to the two post offices where it was being tested,
the Finance and Administration Department approved a June
1965 study report proposing that a source data collection
system be installed in the Department's 75 largest post of-
fices. This proposal was based on data developed during
the operation of the source data collection system being
tested at Minneapolis and Milwaukee. The study report con-
cluded that:

     1. The system would provide a less expensive way of
        gathering management information.

     2. The information and data gathered would be more ac-
        curate.

     3. Data would be made available to all levels of man-
        agement on request.

     4. Supervisors would be relieved of approximately
        1 hour of clerical work a day.

     5. The system would reduce manual work load in the areas
        of keypunch/key-verify, timekeeping, work measure-
        ment, payroll, and cost ascertainment.

     6. Data would be gathered, processed, and disseminated
        at greater speeds than with the existing manual sys-
        tem.

     The Comptroller General and the Director of the Bureau
of the Budget (now Office of Management and Budget) issued
guidelines in January 1959 and March 1960, respectively, to
be followed by all executive branch departments and estab-
lishments in making preacquisition (feasibility) studies of
automatic data processing (ADP) systems. These guidelines,
which are still in effect, describe the essential charac-
teristics of an adequate feasibility study, including, but
not limited to:

     -- A realistic examination of the cost of an ADP system
        when related to the results that can be achieved.
                               13
     -- A discussion of alternative systems considered.

     -- Specific recommendations directed toward both the
        specific problems involved in getting started and
        the proposed regular operation under the new system.

     We believe that the June 1965 study report fell short
of meeting these guidelines and of demonstrating the feasi-
bility of the proposed source data collection system and
did not justify the decision to acquire and install the
costly equipment in 75 post offices. Some of the shortcom-
ings of the study report were:

     1. The feasibility and costs of the proposed system
        were based on data for a significantly different
        system in Minneapolis and Milwaukee, a system which
        had neither been tested fully nor proved successful.
        We were informed that supporting documentation for
        the cost data was not available. (See pp. 21 to 23
        for a discussion of cost overruns.)

     2. Specific problems--such as adequacy of electric
        power, air conditioning, space, and transmission
        capability of local telephone facilities--which
        might be involved in making the systemsoperational,
        were not identified. (See pp. 23 and 24 for a dis-
        cussion of problems that developed because this was
        not done.)

     3. Only one data collection system was discussed and
        no alternative proposals were presented or consid-
        ered.

     The study report did not contain, although required by
the guidelines, any recommendations for developing the pro-
posed system. Furthermore, the conclusions reached in the
report (see p. 13) were not supported by the results of the
limited tests performed prior to the date of the report.

     On August 6, 1965, at a Source Data Automation Confer-
ence, Department officials made the following comments con-
cerning the Minneapolis-Milwaukee system.

     1. The present system would be changed. Punched paper
        tapes and compilers would be eliminated and data
                            14
       would be transmitted directly to computers for re-
       cording on magnetic tapes. Transactor input devices
       used would be replaced by more sophisticated hard-
       ware at less cost.

     2. The electronic scales used for measuring mail volume
        were experimental and would continue to be experi-
        mental until Department engineers and ADP managers
        were satisfied that the Department had the most ac-
        curate type available.

     Although changes would be made in the prototype system,
officials at the conference proposed a plan for implementing
a source data collection system.  Seventy-five post offices
would be tied to approximately eight new source data collec-
tion computer complexes. Five to 10 post offices would be
tied to each complex with approximately 50 input stations
in each post office. Justification for the plan was based
on the Minneapolis-Milwaukee system limited test results
which had been questioned by Department representatives.

     Representatives at the conference from the Department's
ADP Management Division contended that the first permanent
source data collection complex in the nationwide system
should be leased with an option to purchase and that, after
equipment needs had been fully explored and determined, the
remaining seven complexes could be purchased. For reasons
which we were not able to determine, however, this lease-
purchase option approach was not adopted and the Department
awarded a $22.7 million negotiated contract in June 1966 for
the purchase and installation of the entire system.

     The PSDS contracted for in June 1966 represented a sig-
nificant change from the system discussed at the Source Data
Automation Conference, described in the Department's feasi-
bility study, and presented to the Congress in February 1966.
Rather than processing data in batches at specific intervals,
such as hours or blocks of hours, the system being installed
processes each data transaction as soon as it enters the
system through remote input stations. Instead of eight com-
puter complexes, each serving five to 10 post offices, the
system being installed has

     -- two ADPCs, each with dual large-scale computer sys-
        tems,
                     !        15
     -- four (now five) teleconcentrator sites, each with
        dual medium-sized computer systems, and

     -- 75 (now 110) designated post office data collection
        and concentration center systems.

The installed electronic equipment in each post office is
connected to one of the five teleconcentrator sites by leased
dual communication lines. Each teleconcentrator site is con-
nected to an ADPC, also by leased dual communication lines.

     This system was not operationally tested prior to the
June 1966 contract award which committed the Department to
procure equipment that initally cost $22.7 million. Such
a test was feasible through a lease-purchase option offered
by the prime contractor and, in our view, the nature and
proposed cost of the system fully justified such an opera-
tional test before the Department awarded a contract for
purchase and installation of the entire system.




                            16
SYSTEM SPECIFICATIONS NOT PREPARED PRIOR TO
SELECTION AND PURCHASE OF PSDS EQUIPMENT

     The preparation of system specifications is a necessary
prerequisite in acquiring ADP equipment. Circular A-54,
issued by the Bureau of the Budget in October 1961, specif-
ically states that:

    "The selection of ADP equipment will not be made
    until system specifications are available to
    serve as a basis for selection. For purposes of
    this Circular, the term 'system specifications'
    means (1) the delineation of the objectives which
    the system is intended to accomplish; (2) the
    data processing requirements underlying that ac-
    complishment, i.e., a description of the data
    output and its intended uses, the data input,
    data files, volumes of data, processing frequen-
    cies and timing; and (3) such ADP equipment ca-
    pabilities as may need to be identified. System
    specifications will be designed to insure free
    competition among equipment manufacturers."
    (Underscoring supplied.)

     Department records indicated that detailed system spec-
ifications for PSDS were not completed until 6 months after
the contract for the procurement and installation of PSDS
was awarded in June 1966 and that at least six major changes
were made to these specifications between January and June
1967. Such changes were necessary primarily because of re-
vised user requirements of the customer bureaus within the
Department.

     In a PSDS status report, dated February 12, 1968, De-
partment officials stated that, because the proposed system
was unique, the request for proposal released to prospec-
tive bidders in January 1966 allowed the manufacturers con-
siderable flexibility in developing the configuration of
equipment for the system. In this regard, the report
stated:

    "Because of the flexibility provided the manu-
    facturers in their proposals, the system spec-
    ifications were not developed in fine detail.

                            17
     The proposals allowed vendors to propose from one
     to eight computer systems, various methods of
     communications, and various input devices.  It
     was necessary to know the exact configuration be-
     fore the details of the specifications could be
     completed. It was also planned that this task
     would be completed in conjunction with the tech-
     niques provided by the successful vendor."



     "While the RFP [Request for Proposall did de-
     scribe the Postal Source Data System in consid-
     erable detail, it did not constitute Systems
     Specifications. Therefore, it was impossible
     to utilize the *** [prime contractor's] program-
     ming support staff as originally contemplated.
     Instead, under the direction and guidance of the
     three-man POD [Post Office Departmentl PSDS
     staff, the *** [prime contractor'slanalysts and
     programmers were utilized to develop a systems
     specification."  (Underscoring supplied.)

     During the period from May 23 to November 30, 1966, ac-
cording to the report, four system specifications were de-
veloped and "junked," primarily because they did not satisfy
the requirements of the customer bureaus.

      In November 1966 about 5 months after the procurement
contract for PSDS had been awarded, the Department decided
to develop entirely new system specifications using more
comprehensive information available on such system require-
ments as the ratios of types of messages; the number, types,
and frequency of reports; and the number of lines of output
printing. In a review of the revised system specifications,
officials of the prime contractor noted that the system
would require additional computer memory capacity and, if
approved, would result in the Department's ordering addi-
tional memory units.

     It became evident, as indicated by the records of the
Department and the prime contractor, that the significant
changes incorporated in the new system specifications were
necessary if PSDS were to serve its purpose of providing all

                             18
levels of the Department's management with timely, accurate,
and useful information. The revisions covered such items
as:

     1. The extension of work measurement reporting to in-
        clude data from the branches as well as from the
        main post offices.

     2. A 100-percent increase in the number of daily re-
        ports.

     3. A 312 -percent increase in daily print line volume.

     4. The increased complexity of processing the data.

These revisions constituted significant changes in the
scope of the PSDS project and in the system requirements
set out in the Department's request for proposal. Depart-
ment records indicated that these changes in scope resulted
in the procurement of additional equipment totaling at least
$855,200.

     Prime contractor officials stated that virtually all
computer programming work was necessarily suspended pending
completion of the development of the revised system specifi-
cation in January 1967, at which time a careful review was
to be made to (1) determine how much of the computer pro-
gramming work accomplished to date could be salvaged and
(2) develop a realistic schedule for completion of the pro-
gramming.

      In a May 1967 status report, the prime contractor con-
cluded that the development of an ultimate system of the
magnitude of PSDS, including all the features desired by
the Department, would require a very long and concentrated
effort by a full-time team representing all interested or-
ganizational units. The report stated that such an effort
had not been made and was not then being made because, if
the system was to be operational by the target date of
November 1968, time would not permit such effort. The al-
ternative, according to the report, was to implement the
most comprehensive system possible within the allowable pe-
riod and to improve the system at a later date. Subse-
quently, the prime contractor determined that only the time
and attendance function could be initially implemented.
                             19
     In our view, if sufficiently detailed system specifica-
tions for PSDS had been prepared by the Department prior to
the award of the procurement contract, the Department:

    1. Would have been able to make more realistic deter-
       minations of equipment requirements and the cost
       thereof.

    2. Could have established a more realistic target date
       for a fully operational system at the initial 75
       post offices.

    3. Might have selected a different system proposed by
       other bidders.

In addition, if the requirements of the customer bureaus
had been clearly identified before development of the ini-
tial system specifications which were later "junked," the
cost of this unproductive effort might have been avoided.




                            20
SITE SURVEYS NOT MADE PRIOR TO
CONTRACT AWARD

     Prior to the award of the contract for the procurement
of PSDS, site surveys were not made at all of the 75 ini-
tially selected post offices to determine the equipment re-
quirements, the space needs, the necessary site modifica-
tions and improvements, and the implementation problems to
be overcome. The Department sent questionnaires to the 75
offices but visited only eight offices before awarding
the contract. Equipment requirements included in the pro-
curement contract were estimated primarily on the basis of
employee complements for the 75 post offices.

     As late as February 1968--more than 19 months after
awarding the procurement contract for PSDS--only 49 of the
75 offices had been surveyed and their actual equipment
needs determined. As a result of these surveys, the orig-
inal equipment quantities and the amount of the contract
were increased substantially through numerous contract
amendments.

     For example: at the Atlanta, Georgia, Post Office, we
were informed that the Department neither conducted site
surveys nor sought the advice of officials at either the
Atlanta Postal Region or the Atlanta Post Office concerning
the types, quantities, or proposed locations of PSDS equip-
ment until more than 4 months after awarding the PSDS pro-
curement contract.

     Equipment valued at about $216,800 was included in the
original procurement contract for the Atlanta Post Office.
As a result of site surveys completed on June 30, 1967--a
full year after the award of the contract--the Department
determined that equipment, costing about $458,100, was
needed to implement PSDS at the Atlanta Post Office and its
associated branches and stations. This amount represented
an increase of about $241,300, or 111 percent, more than the
amount that had been estimated and provided for in the orig-
inal contract.

     A site survey to determine the quantity and location
of PSDS equipment at the Brooklyn, New York, Post Office
was not completed until May 5, 1967. As a result of the

                            21
survey, the cost of PSDS equipment required by the Brooklyn
Post Office was increased from the original estimate of
$445,100 to $543,000, an increase of 22 percent.

     The following table shows the adjustments in the quan-
tities of equipment that have been necessary at 10 other
PSDS post offices and the effect the adjustments have had
on the cost of PSDS. Similar equipment adjustments were
made at other PSDS post offices not included in this table.

                                               Deliveries to      Percent
                     Original estimates        April 1. 1969       of in-
   Post office       Quantity               Quantity               crease
    location         in units    Cost       in units    Cost      in cost

Albany, N.Y.            41    $   102,500       77- $   203,100    98.1
Birmingham, Ala.        61        148,700       94      255,200    71.7
Charlotte, N.C.         47        120,500       98      217,300    80.3
Greensboro, N.C.        39        100,600       79      180,200    79.2
Miami, Fla.            123        286,900      210      468,700    63.4
New Haven, Conn.        50        121,700       82      197,000    61.9
Rochester, N.Y.         77        193,700      133      301,500    55.5
Springfield, Mass.      53        124,100       85      208,600    68.1
Tampa, Fla.             49        127,200       86      200,000    57.0
Richmond, Va.           60        150,300       95      252.700    68.1

   Total               600   $1,476,200      1.039   $2.484.300    68.3


     From the award of the contract for the acquisition and
installation of PSDS to December 31, 1970, 31 amendments
had increased the amount of the contract from $22.7 million
to $49.7 million, as shown in the following table.




                                   22
                       Date              Description                        Cost

Original order         6-30-66   Equipment and installation                     $22,699,991
Amendment number:
     1                 8-24-66   Price adjustment (decrease)    $         -80
     2                11- 4-66   Equipment modifications               42,847
     3                 5-24-67   Additional equipment               4,589,475
     7                12-28-67        do.                           4,223,932
     8                 1- 4-68        do.                           1,835,760
    10                 1-10-68   Equipment decrease                      -104
    13                 6-28-68   Additional equipment               2,224,365
    18                 9-13-68   Equipment modifications               11,589
    20                10-16-68   Additional and replacement
                                   equipment                          129,000
    23                 1-15-69   Equipment exchange                    83,000
    24                 1-24-69   Additional equipment                 227,700
    25                 1-23-69   Equipment modifications                3.000      13.370.484

         Equipment and installation costs related directly to
           the initial 75 post offices                                             36,070,475

    27                 5-13-69   Additional equipment for the
                                   initial 75 post offices
                                   and 35 expansion post
                                   offices                       11,155,717
    29                 8-29-69   Equipment modifications              9,585
    31                12- 3-70       "     decrease                  -4,320        11.160.982

         Total contract equipment and installation costs                           47,231,457

Costs of other contract amendments                                                  2,502.422

         Total contract amount as of December 31, 1970                          $49.733.879



     Of the additional PSDS equipment acquired at a cost
of $24.5 million under the amendments to the contract,
$7.9 million (32 percent) represented the cost of equipment
purchased at lower unit prices and $12.7 million (52 per-
cent) represented the cost of equipment purchased at the
same or higher unit prices than prices for similar equip-
ment under the original contract. The remaining $3.9 mil-
lion (16 percent) represented the cost of equipment which
was different from that provided under the original con-
tract.

      Since         the amount of the contract has more than doubled,
 we believe         that, if more precise equipment requirements had
 been known         prior to award of the contract, more competition
 might have         been generated and lower prices for the larger
 quantities         might have been offered by the bidders.

      Further, if the Department had made adequate site sur-
 veys prior to the decision to acquire PSDS, it would have
 been in a position to recognize and correct any expected

                                             23
problems in installing the equipment at post offices. In
an April 1968 survey, the prime contractor reported that,
in attempting to implement PSDS at the first 47 post of-
fices, it had encountered several major problems, including
lack of adequate electrical power and air conditioning to
properly operate PSDS equipment in the post office data
collection sites.

      Of the 47 post offices, only 15 had adequate electri-
cal power, only 19 had adequate air conditioning, and only
11 had both adequate electrical power and air conditioning.
The prime contractor stated that it could not successfully
implement PSDS at a post office site under these conditions
and recommended that the Department inspect each site prior
to authorizing implementation of the system at a post of-
fice.

     The Department subsequently developed an equipment in-
stallation and testing program which required that site
surveys be made before ordering equipment or preparing
sites for equipment installation.




                            24
CHANGE FROM WORK MEASUREMENT SYSTEM
TO WORK LOAD RECORDING SYSTEM

     Prior to September 1968, work was performed on PSDS
computer programs for a Work Measurement System application
designed to provide mail volume and labor distribution data
for each postal operation and for each 8-hour work tour
within a post office and to measure individual employee pro-
ductivity. The Work Measurement System was not implemented
as part of PSDS, although we were informed that system spe-
cifications had been prepared and computer programs had
been developed and tested prior to the fall of 1968. Rather,
the computer programming for the Work Measurement System ap-
plication was scrapped by the Department in the fall of 1968
when the Bureau of Operations (now Operations Department)
replaced this system with a Work Load Recording System. In-
formation was not readily available for us to determine the
cost of the scrapped computer programs.

     Under the Work Load Recording System, mail volume and
labor-hour distribution data was to be provided on a 24-
hour basis for selected operations only,rather than for each
8-hour work tour for all operations, and productivity was to
be measured for groups of employees rather than for individ-
ual employees.

     The first tests of the automated Work Load Recording
System were made during the last week of February 1970, us-
ing data input from the PSDS post office in Richmond, Vir-
ginia. These tests proved unsuccessful, primarily because
of several incompatibilities between the computer program
and the ADPC equipment. The program was to be corrected and
rescheduled for testing in Richmond in early March 1970, and
further testing or implementation of the Work Load Recording
System in other PSDS offices was to be dependent upon favor-
able results of the Richmond tests.

     In July and August 1970, more than 1-1/2 years beyond
its original target date of November 1968, the Department
began implementing the automated Work Load Recording System
in PSDS post offices where the employee time and attendance
application was operational. As of November 1970, a total
of 86 post offices were using the automated system.


                             25
      The change from the Work Measurement System to the
 Work Load Recording System resulted not only in scrapping
 computer programs but also in overprocurement of mail weigh-
 ing equipment and relocation of equipment installed in the
 post offices. These matters are discussed in the following
 sections.

 Overprocurement of mail weighing equipment

     The Department procured electronic mail-weighing equip-
ment at a total cost of about $10.7 million, including
$3 million for bench scales, which was to be used as part
of PSDS to record mail volumes processed. Bench scales and
associated electronic equipment were installed in PSDS post
offices and, in some instances, remained idle for 2 years
or more because neither a work measuring nor a work load re-
cording system had been implemented under PSDS. Some of
these scales were used to weigh mail under a manual version
of the Work Load Recording System, but the scales were not
tied into the electronic network of PSDS.

      In September 1969, about 1 year after the decision to
·replace the Work Measurement System, Departmental ADP Man-
 agement officials completed an evaluation of the equipment
 requirements for the Work Load Recording System and deter-
 mined that only 600 bench scales would be needed at the 110
 post offices planned for implementation of PSDS. However,
 977 bench scales and related accessory equipment, 377 scales
 in excess of the Department's needs, had been purchased.

     As of September 1969, 792 bench scales had been de-
livered or installed in post offices and the remaining 185
scales had been manufactured and were stored in the vendor's
warehouses, pending delivery instructions from the Depart-
ment. As of November 1970, the vendor was storing scales
for the Department under an agreement which extends to July
1971. To our knowledge no evaluation was made in the fall
of 1968 to determine whether equipment needs under the Work
Load Recording System were different from equipment needs
under the Work Measurement System.

     The Department's contracting officer informed us that,
under the terms of the procurement contract, the Department
was obligated to accept delivery of all 977 bench scales

                             26
and related accessory equipment because production of the
scales had been completed in July 1969. We estimate that
the overprocurement of 377 bench scales and related equip-
ment has unnecessarily increased the Department's cost of
PSDS by about $1.2 million.

     We believe that the Department should have reevaluated
its PSDS equipment requirements at the time the decision
was made in the fall of 1968 to replace the Work Measure-
ment System with the Work Load Recording System. Depart-
ment records indicated that, as of October 31, 1968, a total
of 690 bench scales and related equipment had been shipped
by the prime contractor to various PSDS post offices--more
than enough to satisfy the bench scale needs for all 110
PSDS post offices.

Relocation of electronic scale equipment

     The change to the Work Load Recording System required
major relocation of electronic scale equipment that had been
installed in the post offices to meet the needs of the new
mail volume recording concept. For example, the Brooklyn
Post Office conducted a survey in January 1969 for the pur-
pose of determining equipment needs under the Work Load Re-
cording System. The Brooklyn Post Office had begun manual
operations under this system in November 1968. On the basis
of this survey, Brooklyn officials determined that three
of seven electronic floor scales that had been installed in
the post office required relocation to other areas in the
office because of the changed system.

     Equipment relocation costs   were incurred at other PSDS
post offices as a result of the   change to the Work Load Re-
cording System. The Department    negotiated a $600,000 con-
tract in February 1970 with the   prime contractor for the
following purpose.

     "Implementation of the new work load reporting
     system and changes of post office operating con-
     cepts will necessitate extensive relocation of
     PSDS equipment that has already been installed
     in 85 post offices. *** This reconfiguration
     contract will cover the cost of material and
     labor required to effect these changes."

                            27
The contract amount was increased to $700,000 in July 1970
to accelerate this relocation effort.

     Recent changes in the design of the Work Load Record-
ing System, incorporating much of the Work Measurement Sys-
tem concept, are expected to again change the electronic
scale equipment requirements.

UNDERUTILIZATION OF COSTLY
COMPUTER EQUIPMENT

     Because the Department did not implement PSDS as sched-
uled, costly computer equipment either remained idle or was
not effectively used for long periods. For example, the
first large-scale computer, acquired at a cost of about
$1.1 million, was delivered to the Paramus ADPC in January
1967. Another large-scale computer, acquired also at a
cost of about $1.1 million, was delivered to Paramus late
that year. The two computer systems (A and B) were ready
for use by January 1968; however, operations were not begun
until September 1968--20 months after delivery of the first
computer. Between April 1967 and September 1968, some in-
formation was being transmitted from four post offices to
the ADPC on a test basis. The computer systems, however,
remained idle much of the time. The following schedule
shows the percent of idle time of the two computer systems
from April 1967 to February 1970.

                          Computer System A

                Period                 Percent of idle time

     Apr.    1967   to   Oct.   1967              52
     Nov.    1967   to   Dec.   1967              80
     Jan.    1968   to   Aug.   1968              30
     Sept.   1968   to   Feb.   1970         Less than 1

                         Computer System B

               Period                  Percent of idle time

     Nov. 1967 to Dec. 1967                    60
     Jan. 1968 to Aug. 1968                    43
     Sept. 1968 to Feb. 1970            9 to less than 2

                                  28
     Most of the recorded production time on the computers,
from April 1967 to August 1968, was for system testing and
program development. Although the table indicates that the
computer systems were almost fully used after August 1968,
the computer capacities were significantly underutilized be-
cause data from only a small number of post offices was be-
ing processed for only one of the three proposed system ap-
plications.

     In our view, the significant amount of idle computer
time and the premature investment in costly computer equip-
ment were a direct result of the Department's failure to
provide realistic implementation and delivery schedules for
PSDS equipment.




                             29
                         CHAPTER 3


           PREDICTED SAVINGS MAY NOT BE REALIZED

     Although in 1966 the Department estimated that install-
ing PSDS in 75 post offices would save about $36 million
during the first 5 years of operation, the unanticipated
high costs incurred to date indicate that savings may not be
realized.

     The Department's predicted cost savings were based on
the assumptions that (1) estimated annual cost to operate
PSDS at the 75 post offices, including amortization of equip-
ment costs over a 5-year period, would be $14.6 million and
(2) because annual employee costs of $21.8 million for the
manual time and attendance and work measurement systems in
operation at the 75 offices would be eliminated, the em-
ployees involved in these manual systems would be reassigned
to other duties.

     These assumptions, which in our view lacked validity at
the time they were made because of the inadequate feasibil-
ity study (see p. 13), were not borne out by actual PSDS op-
erations. Employees who were assigned to operate the manual
system have not been reassigned to other duties. On the
contrary, operating costs to date have far exceeded the De-
partment's estimates, and its anticipated cost reductions
have not materialized.

     The Department estimated, in its 1965 feasibility study,
that 740 employees would be needed to operate PSDS in the
75 largest post offices and that the annual employee costs
would total $5.5 million.

     Using the same basis used by the Department in its
feasibility study, and recognizing the increases in employee
complement and mail volume after 1965, we estimated that,
as of October 1970, about 1,050 PSDS employees, at an an-
nual cost of $10.5 million, would be needed to handle PSDS
transactions at the 75 post offices. Total authorized PSDS
employees, however, had reached 1,857 by October 1970, or
about 800 more than our updated estimate.



                             30
     We estimated also that the annual salary costs for the
1,857 authorized employees totaled $19.6 million, or
$9.1 million (87 percent) more than our updated estimate
of $10.5 million and $14.1 million (256 percent) more than
the Department's 1965 estimate. Data was not readily avail-
able for us to determine the PSDS employee salary costs for
overtime pay, Sunday and holiday premium pay, and night dif-
ferential pay.

     The extent of PSDS-related employee increases after
PSDS was installed varied among post offices. For example,
our review at three post offices for a 4-week period showed
that employee costs for processing time and attendance
transactions alone increased substantially as follows:

                   Man-hours       Percent of   Personnel costs Percent of
Post office      Before After       increase    Before   After   increase

Atlanta, Ga.     2,945     5,310        80      $15;200 $ 25,200     66
Brooklyn, N.Y.   7,194    12,250        70       37,000   64,400     74
Newark, N.J.     2,767     6,779       145       14,600   35,700    144

    Total        12.906   24,339        89      $66,800 $125,300     88


     Some post offices have found it necessary to employ
full-time time and attendance clerks to prepare the neces-
sary payroll forms, whereas, before the installation of
PSDS, such forms were filled out by production employees in
their spare time.

     In addition to the 1,857 authorized PSDS employees, a
substantial number of other Workers were involved in PSDS
operations at the local post office level. For example, at
the Brooklyn Post Office, in addition to the 31 positions
authorized for the PSDS, eight other employees were as-
signed to the PSDS data collection site. At the Newark,
New Jersey, Post Office, in addition to the 18 positions
authorized, 21 other employees were assigned to PSDS activ-
ities. The additional assigned employees held other offi-
cially authorized positions in the two post offices but were
diverted from these positions to perform PSDS duties.

     The Department, in its original estimate of cost savings
for the system, significantly underestimated other annual

                                      31
operational costs for such items as supplies and services,
maintenance, and amortization of equipment costs. The De-
partment's 1965 estimate of the annual costs for these items
totaled $9.1 million for the 75 post offices to be included
in the system. We estimated, however, that the annual
costs for these items for the 75 offices totaled at least
$17 million, or $7.9 million (87 percent) more than the De-
partment's estimate.

     Cost studies made by the Department's internal auditors
at selected PSDS post offices indicated that, rather than
eliminating or reducing clerical and operating costs at
post offices, PSDS had substantially increased costs.



     In its fiscal year 1971 budget submission to the Con-
gress, the Department stated, as follows:

     "Cost Reductions, $5,000.000:

    "The gross employment increase for workload is
    offset by a reduction of 737 positions and man-
    years to be achieved by improved time and atten-
    dance processing resulting from the implementa-
    tion of the Postal Source Data System. This
    substantial reduction in the 1970 level of em-
    ployment represents a savings of $5,000,000."

     In view of our findings which indicate that more em-
ployees are needed to process time and attendance data under
PSDS than under the former manual system, we have reserva-
tions as to whether this $5 million reduction in employee
costs can be achieved under PSDS as it is presently being
operated.




                             32
                         CHAPTER 4

         LIMITED USEFULNESS OF PSDS INFORMATION

     The purposes of PSDS, as stated by the Department, are
to provide more timely, accurate, and meaningful data; to
eliminate paper work and relieve supervisors of time spent
on paper work activities; and to provide improved informa-
tion. At the time of our field review, these purposes had
not been fulfilled.

EMPLOYEE TIME AND ATTENDANCE REPORTS

     The volume of time and attendance reports produced by
PSDS is substantial. A Department task force which studied
the PSDS reporting system stated in its September 1969 re-
port that, during one 2-week pay period, a single post of-
fice received an estimated 12 linear feet of PSDS reports--
enough to fill six filing cabinet drawers. We question
whether such volume of data can be examined timely and ade-
quately for effective use by management.

     Of the 24 reports produced by PSDS at the time of our
field review late in 1969 and early in 1970, 19 were used
as control and information documents needed to operate the
PSDS time and attendance function and only five were de-
signed to replace management reports previously prepared
manually by postal mail operations supervisors. The post
offices were still maintaining manual records, however, be-
cause the reports produced by PSDS were inaccurate or in-
complete, or were received too late to be of use to postal
supervisors.
     The Department has stated that the Time and Attendance
Sequence Error Report, which is produced at 2-hour intervals
and identifies sequence errors (e.g., employee checked in at
8 a.m., and 1 p.m., but did not check out for lunch), should
be the fundamental and most effective error signal to the
post office. The timely receipt and correction of reported
errors has a significant effect on the correctness and reli-
ability of other reports generated by PSDS, because subse-
quent reports will contain erroneous data until corrected.
Although this report was scheduled to be received by postal
management at 2-hour intervals and was designed to cover


                             33
sequence errors occurring within a 2-hour period, the report,
at times, was received by management as much as 14 hours
late. It was therefore useless to local postal management
as a means of promptly identifying and correcting errors
listed in the report.

     The Department task force, upon completing its study
of the reporting system of PSDS in September 1969, recom-
mended that the Department improve all the reports gener-
ated by PSDS. Their study recognized the need for timely
receipt of reports by post offices. They observed that:

     "*** In many cases the two-hour reports are de-
     layed 6 to 8 hours and in extreme, but not unique
     cases, the post office may not receive a two-hour
     report for more than 14 hours. The result of the
     irregular and late receipt of this report is that
     the report is often ignored and corrections made
     to the Daily Sequence Error Reports because both
     the employee and employee supervisor have com-
     pleted their tour and the information is often
     not available to make the corrections at the time
     the Two-Hour Report is received. These omissions
     cause other reports to lack integrity because
     transactions in the error files cannot be incor-
     porated into subsequent reports."

     We found similar difficulties involving the late receipt
of PSDS reports. For example, at the Atlanta Post Office,
none of the 2-hour reports were being received on time. At
the Brooklyn Post Office, reports for a particular day's
time and attendance transactions were received by postal
stations generally 2 or 3 days after the day covered by the
reports. Station officials at Brooklyn informed us that
the reports were less effective and afforded them less con-
trol over employees than the unofficial manual records they
were maintaining.

     The Department's task force observed an additional prob-
lem related to the receipt of error listings too late to
make the necessary corrections to payroll records. The
task force stated in its report that:




                             34
    "*** At the end of the pay period the entire error
    file (approximately 1 million messages) is purged
    from the system. ***"



    "The purging of one million uncorrected errors
    each pay period reflects adversely upon the in-
    tegrity of the reporting system. *** In order for
    field personnel to ever obtain confidence in the
    system, we believe that it is necessary that they
    attempt to correct all sequence errors."

     The task force found not only that errors were not be-
ing corrected but also that the number of reports being gen-
erated to list errors and other information was excessive.
The task force said that:

    "A total of 24 reports are now produced by the
    PSDS system. Nineteen of these reports relate to
    Time and Attendance. We conclude that this prob-
    ably is an excessive number of reports but field
    personnel contend that these reports are neces-
    sary because of continued malfunction of the sys-
    tem at the ADPC's. Conversely, we have reason to
    believe that one of the major causes of this mal-
    function is the excessive number of reports pres-
    ently required over and above the planned capac-
    ity of the system."

     At the post offices we visited, the number of reports
generated under PSDS had not been reduced from the number
required by the post offices before PSDS was installed.
Local post office employees found it necessary to maintain
manual records and, as a consequence, time spent by mail
operations supervisors had not been reduced. On the con-
trary, their work had increased in certain cases.

     For example, supervisors and timekeepers in the Brook-
lyn Post Office indicated that PSDS had increased their
paper work and that the time spent on timekeeping duties
had increased since implementation of PSDS. Station super-
visors continued to maintain unofficial time and attendance
records after PSDS was installed, because such records were

                             35
believed to be less time-consuming to work with in manually
preparing local management reports on such subjects as the
use of substitute mail carriers. Such unofficial records
were often necessary as a data source because PSDS reports
were not timely.

     Some work stations used manual time cards and time
clocks for substitute employees, in addition to PSDS elec-
tronic input devices. The time cards were used to aid super-
visors in locating employees and in preparing required work
station reports, since report data generated by PSDS was not
received timely enough for the station supervisors' needs.

     Three of the 22 reports received by the Atlanta Post
Office--Report of Loaned Other Hours, Report of Time
Certification-Higher Level, and Report of Transfer To-From
Payroll Accounts--were not used for any purpose. A fourth
report, the daily Carrier Report, was considered to be a
management report but, as prepared by PSDS, was inadequate
for management purposes. The report listed daily, by route
number, all city delivery service. It showed the total of-
fice time and street time for each route and the name of
only the first carrier going on the street.

     The report did not provide information on individual
carriers by name and hours and, consequently, could not be
used to either supplement or eliminate the time-consuming
preparation of Department Form 3997, Daily Record of Carrier
Absences, Replacements, and Overtime. As a result, this
PSDS report was not being used by local management officials
at the Atlanta Post Office.

      The most significant problems encountered by PSDS post
offices with regard to the reports produced by the system
were (1) the large volume of reports, many of which were of
little or no value, (2) the inability of the post offices
to forecast the time of receipt for any report or group of
reports, and (3) the lack of integrity of reports, which
required post offices to maintain supplementary manual rec-
ords.

     In commenting on the problems with PSDS reporting, the
Department task force concluded that:


                             36
    "*** it is deemed appropriate to point out that
    our present problems are not due exclusively to the
    number or content of reports. The continuous prob-
    lems associated with the system, including loss of
    rings and the dropping of Employee Master Records
    has had'a demoralizing impact upon post offices us-
    ing the system. Unless these deficiencies are
    corrected, whether they be caused by post office
    personnel, hardware, software, or operator prob-
    lems, no amount of modification in the reports will
    satisfactorily alleviate the present difficulties."

     The Department's internal auditors have criticized PSDS
because the payroll data produced has not met acceptable
standards. The 1969 Annual Report of the Postmaster General
stated that the internal auditors deemed it impracticable
to verify an estimated $235.8 million in payrolls processed
under PSDS and, as a result, qualified their opinion regard-
ing the Department's financial statements for fiscal year
1969. We were advised that the primary reason for this
qualification was that the PSDS time and attendance applica-
tion had not "stabilized" and that therefore the internal
auditors had difficulty with the controls over, and reliabil-
ity of, the payroll data.

     In our view the lack of reliable payroll data high-
lights the inadequacy of PSDS reports, because the inaccu-
rate information generated by the system cannot be used by
management to effectively evaluate operations and make sound
decisions.

WORK LOAD RECORDING SYSTEM REPORTS

     The Work Load Recording System was to provide essential
information on mail volume and labor-hour distribution to
enable postal managers to improve their ability to schedule
the work force to meet the expected work load.

     Despite receiving information indicating that the Work
Load Recording System would not meet the above objective,
in July 1970 the Department began incorporating it into the
automated PSDS procedures.



                              37
     An official in one post office told us that information
being received under a manual version of the Work Load Re-
cording System was not as meaningful as that which had been
received under the Work Measurement System. He said that
the Work Measurement System information had been broken
down into 8-hour shifts, whereas the Work Load Recording
System provided information for 24-hour periods without
regard to individual work shifts. Therefore production in-
efficiency could not be identified by work shift.

     A Department report for the fourth quarter of 1969
stated that routine work load recording procedures yielded
only the total daily volume processed through each opera-
tion--figures that, according to the report, were manifestly
unusable in determining optimum staffing levels for differ-
ent times of the day.

     Our visits in September and November of 1970 to several
post offices confirmed that Work Load Recording System re-
ports did not effectively serve postal management's efforts
to improve its scheduling of the work force to meet expected
mail work loads and efforts to improve the management of
mail processing during the work tours. Management officials
stated that Work Load Recording System reports generated by
PSDS were received about 2 days after the day covered by the
report, that the data was not accurate, and that the data
as shown in the reports was for a 24-hour period rather
than by an 8-hour work tour. They indicated, however, that
the reports were useful in establishing historical trends
on hours used and pieces of mail handled.



     We believe that the reports and data produced by PSDS
have not resulted in more timely, more accurate, or more
useful data as predicted by the Department in its initial
justification of PSDS in 1966. The Department, however,
has reported progress in eliminating unnecessary reports and
in making changes to other reports to improve their accuracy
and usefulness to postal management.




                             38
                         CHAPTER 5

          PSDS EXPANSION AND MODIFICATION PROGRAM

     The Department is engaged in a multimillion-dollar pro-
gram to expand the PSDS data collection network to an addi-
tional 35 post offices and to modify the system still being
installed in the first 75 offices.

     The Department's 5-year program and financial plan ini-
tially called for the expansion of the PSDS network to an
additional 25 post offices each year during the period 1971
through 1975, a total of 235 PSDS post offices by the end
of fiscal year 1975. In August 1969 the Department revised
this plan to limit the expansion to 150 post offices. Ac-
cording to a Department official, however, this plan was
further revised to limit the expansion to 110 post offices
because of budgetary considerations. Funds committed under
contracts for the expansion and modification program totaled
more than $14 million, and the Department included in its
fiscal year 1971 budget request more than $3 million to con-
tinue these efforts.

     As of June 1970 funds had been committed and work was
in progress in the following areas.

     -- Two additional teleconcentrator sites, not contem-
        plated in the orginal PSDS proposal, were being
        planned and/or constructed and equipped at Oklahoma
        City and Binghamton, New York (the Oklahoma City
        teleconcentrator is now in operation).

     -- Additional equipment was being or had been installed
        in the ADPCs at Paramus and St. Louis to establish
        dual data processing systems with multiprogram capa-
        bilities at each site (this equipment installation
        has since been completed).

     -- The ADPC at Paramus was being relocated to Wilkes-
        Barre, primarily because of the lack of adequate
        space for expansion at Paramus (this relocation has
        since been completed).


                             39
     -- The Washington teleconcentrator site was being planned
        for relocation to Richmond, primarily because of the
        lack of adequate space for expansion in Washington.

     -- PSDS data collection equipment was being installed
        in postal facilities which would be replaced by new
        facilities within a relatively short period of time--
        this would result in additional relocation and/or re-
        placement costs for PSDS equipment and cabling when
        these new facilities were completed.

     In September 1970, after we reported our review findings
to the Postmaster General, Department officials informed us
that plans to establish an additional teleconcentrator site
at Binghamton and to relocate the Washington teleconcentra-
tor site to Richmond had been revised. The revised plans
call for consolidating, at the new ADPC in Wilkes-Barre,
all teleconcentrator operations supporting the ADPC which
were to have been located at separate sites in Binghamton,
Paramus, and Richmond. The Department believes that the
planned consolidation will reduce overall system costs and
will increase operating efficiency.

     In our opinion PSDS, as it is being operated in the
post offices, is not resulting in cost savings nor improving
the internal management of postal operations and therefore
PSDS should not be expanded at this time to include addi-
tional post offices.




                             40
                           CHAPTER 6

              CONCLUSIONS, AGENCY COMMENTS AND OUR

                 EVALUATION, AND RECOMMENDATIONS

CONCLUSIONS

     In our opinion the deficiencies and problems of PSDS
and the high costs associated with its installation and op-
eration can be attributed to inadequate study and insuffi-
cient testing of PSDS prior to its nationwide installation.

     In summary, changes in PSDS after the award of the con-
tract for its acquisition and installation have resulted in
overprocurement of costly PSDS equipment. Unrealistic sys-
tem implementation and equipment delivery schedules have
resulted in equipment not being used. Employee costs and
other operating costs have greatly exceeded the Department's
original estimates. The PSDS payroll (employee time and
attendance) operation has been more costly than the previous
manual payroll operation. The reports being generated by
PSDS at the time of our review were less timely, less mean-
ingful, and less accurate than reports available prior to
installation of PSDS, and therefore the reports were less
useful to postal management.

     Despite these deficiencies in the system, the Depart-
ment continues the expansion of PSDS to additional post of-
fices.

AGENCY COMMENTS AND OUR EVALUATION

     On June 26, 1970, we advised the Postmaster General of
the results of our review and proposed that the Department
not expand PSDS to any more post offices before making a
comprehensive evaluation of PSDS, including a cost-benefit
study, to determine its efficiency, economy, operational
reliability, and usefulness to postal management. We sug-
gested that expenditures for PSDS be held to a minimum un-
til its operation evidences that cost savings and demon-
strated improvements in postal management can be achieved.


                                41
     By letter dated September 9, 1970 (see app. IV), the
Postmaster General stated that the results of our review
confirmed many of the findings previously reported by the
Department's Internal Audit Division and that the difficul-
ties, delays, and added costs involved in establishing PSDS
were a matter of concern to the Department.

     He concurred generally with our proposal to suspend ex-
pansion of PSDS until a comprehensive evaluation had been
made, except that he believed the limited further expansion
to 110 post offices must be completed if the Department was
to avoid further adverse effects on the overall objective
of providing information useful to control operations at
large post offices. He advised us that commitments had been
made and obligations had been incurred for most of the equip-
ment involved in the limited expansion to 110 post offices.

     He acknowledged that difficulties had been experienced
which impaired the effectiveness of PSDS to date. He stated,
however, that, although some mistakes had been made, the
Department did not agree that all of them could have been
avoided. In explanation of these difficulties, he cited,
among other things, the adoption of an accelerated, and pos-
sibily unrealistic, schedule for implementing PSDS and the
concurrent development and implementation of PSDS.

      The Postmaster General stated also that, over the pre-
vious 12 months, postal management had introduced consider-
able improvement intended to increase the reliability and
effectiveness of computer processing operations and the
timely submission of report data to post offices. Redesign
of computer programming systems, introduction of quality
control checks to assure continuous integrity of data, and
adjustments in equipment components were cited as being
among the changes made to improve operations. He said he
believed that these efforts had, to a significant degree,
overcome the difficulties we had presented in our draft re-
port.

     The Postmaster General said also that the Assistant
Postmasters General for Finance and Administration and for
Operations had assured him that substantial benefits were
forthcoming from the operation of PSDS.


                             42
     We were subsequently informed by Department officials
that, among improvements introduced in PSDS, changes were
made in the equipment to (1) enable PSDS to process various
kinds of information simultaneously and provide a higher
speed sorting capability and (2) increase the memory (data
storage capacity) of PSDS. We believe that these actions
should aid in achieving a more productive system.

     Although efforts have been made to improve the accuracy
and reliability of the PSDS time and attendance application,
our review has shown, and Internal Audit Division studies
have shown also, that it is more costly to perform time and
attendance functions under PSDS than under the former manual
system.

     Reports produced under the Work Load Recording System
application of PSDS at the time of our field visits were not
effectively serving postal management. We believe that con-
siderable improvement is needed before reports produced by
PSDS can be effectively used to improve management of mail
operations.

     We agree that providing useful information to manage-
ment for controlling operations is a desirable objective.
However, the work load data produced by PSDS is not accurate,
is not timely, and is not effectively serving managment in
either controlling or improving postal operations. We there-
fore believe that it is not desirable to expand PSDS to ad-
ditional post offices without first making improvements so
that it will produce data which is not only reliable and
timely but also helpful to management in doing a better job.

     We believe also that the expansion of PSDS to any ad-
ditional post offices, when its effectiveness and economy
have not been demonstrated at the facilities where it has
been installed, would not be a judicious investment of Depart-
ment funds.

     Although the Postmaster General stated in his September
1970 letter that commitments had been made and obligations
had been incurred for most of the equipment involved in the
expansion of PSDS to 110 post offices, the installation of
this equipment in the post offices requires the expenditure
of substantial additional funds for such items as electrical

                             43
cabling, building modifications, special air conditioning,
and auxiliary electrical power generation, as well as for
operation and maintenance of the equipment once it has been
installed.

     We noted that on October 12, 1970--1 month after com-
menting on our draft report--the Department awarded a
$400,000 contract for furnishing and installing cabling and
related equipment at PSDS post offices. We note also that
the Department has budgeted, for fiscal year 1972, additional
funds for PSDS equipment for the 110 post offices--$l1 mil-
lion for data collection equipment and $600,000 for systems
reconfiguration--and $4 million to maintain and repair in-
stalled PSDS equipment, including the additional equipment
budgeted for fiscal year 1972.

     Further, the Department has projected additional PSDS
equipment requirements and related building alterations to-
taling $1 million annually for fiscal years 1973 through
1976 for continued system expansion and modification within
the 110 post office network. We believe, therefore, that,
notwithstanding the Department's existing commitments and
obligations for equipment to expand PSDS to the 110 post of-
fices, significant future expenditures of funds could still
be avoided if the current expansion were suspended.

RECOMMENDATIONS TO THE POSTMASTER GENERAL

     We recommend that the Postmaster General

    --suspend the PSDS expansion program pending a compre-
      hensive evaluation and cost-benefit study of PSDS to
      determine its efficiency, economy, operational reli-
      ability, and usefulness to postal management and

    -- curtail the procurement of PSDS equipment and soft-
       ware (programs) and keep operational costs to a min-
       imum,pending the outcome of the comprehensive eval-
       uation and cost-benefit study of PSDS.




                            44
                        CHAPTER 7

                      SCOPE OF REVIEW

     Our review was made at the Department's headquarters
in Washington; at the postal regional offices and postal
data centers in Atlanta and New York; at the ADPC located,
at the time of our review, in Paramus; at the teleconcen-
trator sites in Washington and Paramus; and at the post of-
fice facilities in Atlanta; Baltimore and Riverdale (Prince
George's County), Maryland; Detroit, Michigan; Newark;
Brooklyn; Richmond and Norfolk, Virginia; and Seattle,
Washington.

     We reviewed Department budget justification documents
for PSDS and legislation and records of hearings of the
House and Senate Committees on Appropriations pertaining to
the planning and the justification for PSDS, At Department
headquarters and the various field locations we examined
(1) pertinent records, correspondence files, contracts, and
other documents related to the feasibility, planning, ac-
quisition and operating costs, installation, and use of
PSDS and (2) reports being generated by PSDS under the em-
ployee time and attendance and the Work Load Recording Sys-
tem applications at selected post offices. We also held
discussions with officials at various management levels
within the Department to determine the usefulness of PSDS-
generated reports to postal management.

     In addition, we reviewed pertinent documents furnished
us by the prime contractor for PSDS.




                             45
APPENDIXES




47
                                                                          APPENDIX I



                                   PSDS POST OFFICES
                                 (as of November 1970)


Alabama:                Illinois:             Missouri:              Oregon:
    Birmingham              Chicago               Kansas City            Portland
                            Chicago North         St. Louis
Arizona:                      Suburbana,b                            Pennsylvania:
    Phoenix                 Chicago South     Nebraska:                  Harrisburg
                              Suburbana,b         Omaha                  Lehigh Valleyab
Arkansas:                   Peoriaa                                      Philadelphia
    Little Rocka            Springfielda      New Jersey:                Pittsburgh
                                                  Camdena
California:             Indiana:                  Hackensacka        Rhode Island:
          a                           a
    Fresno                  Fort Wayne            Jersey City            Providence
    Inglewooda              Indianapolis          Newark
    Long Beach                                    Patersona,b        South Carolina:
    Los Angeles         Iowa:                     Trenton                Columbiaa
    Oakland                 Des Moines
    Pasadenaa                                 New Mexico:            Tennessee:
    Sacramento          Kansas:                   Albuquerque            Knoxvillea
    San Bernardinoa         Wichita                                      Memphis
    San Diego                                 New York:                  Nashville
    San Francisco       Kentucky:                 Albany
    San Jose                Lexingtona            Bronx              Texas:
    Santa Anaa              Louisville            Brooklyn               Austina
    Stockton a                                    Buffalo                Dallas
             a
    Van Nuys            Louisiana:                Flushingb              Fort Worth
                                        a
                            Baton Rouge           Hicksvillea,b          Houston
Colorado:                   New Orleans           Jamaica                San Antonio
    Denver                                        Long Island City
                        Maine:                    New York Cityb     Utah:
Connecticut:                Portland              Rochester              Salt Lake City
    Bridgeporta                                   Syracuse
    Hartford            Maryland:                 White Plainsa,b    Virginia:
    New Haven               Baltimore             Yonkersab              Norfolk
                            Prince Georgesa                              Richmond
Delaware:                                     North Carolina:
               a
    Wilmington          Massachusetts:            Charlotte          Washington:
                            Boston                Greensboro             Seattle
District of Columbia:       Springfield                                  Spokane
   Washingtonb                                Ohio:                      Tacomaa
                        Michigan:                 Akron
Florida:                    Detroit               Cincinnati         West Virginia:
    Jacksonville            Flint a               Columbus               Charlestona
    Miami                   Grand Rapids          Cleveland
    Orlandoa                Royal Oaka            Dayton             Wisconsin:
                   a
    St. Petersburg                                Toledo                Madisona
    Tampa               Minnesota:                                      Milwaukeeb
                            Minneapolisb      Oklahoma:
Georgia:                    St. Paul              Oklahoma City
    Atlanta                                       Tulsa

aDenotes the 35 post offices included in the expanded PSDS network.

bDenotes post offices where PSDS had not been installed as of November 30, 1970.




                                            49
APPENDIX II


       BASIC COMPONENTS OF POSTAL SOURCE DATA SYSTEM

INPUT DEVICES

     1. Badge readers are used to record time and attendance
        and work assignment data. (See picture on p. 53.)

     2. Transactor input stations are used to record leave,
        personnel actions, adjustments, and certain mail
        volume information and can be used as badge readers.
        (See picture on p. 54.)

     3. Electronic scales are used to record the weight of
        the mail by type, classification, and operation.
        Three types of scales are used: floor, bench, and in-
        motion (conveyor). (See pictures on pp. 55 and 57.)

    4. Alphanumeric input terminals are used to initiate
       inquiries relative to any data pertinent to a post
       office, such as employee pay records. (See picture
       on p. 56.)

COMMUNICATION DEVICES AND NETWORK

     1. One or more concentrators are located within a data
        collection site at each of the PSDS post offices.
        These concentrators continuously collect all data
        generated by the input devices and automatically
        forward it to one of five teleconcentrator sites.
        Each concentrator contains dual components to assure
        continuous operation.

    2. Each of the five teleconcentrators consists of dual
       medium-sized computer systems (as a safeguard against
       system failure in the event oneunit becomes inoper-
       ative) and various subsidiary equipment. The tele-
       concentrators receive data from the PSDS post offices
       they serve, transmit acknowledgements to the concen-
       trators when data has been received, and forward the
       data to the ADPCs for processing. In the event an
       ADPC is unable to accept data transmission, the tele-
       concentrators can store the data on magnetic tape
       until it can be transmitted to the ADPC. (See pic-
       ture on p. 59.)

                             50
                                                APPENDIX II


     3. Local and long-distance transmission lines leased
        from various telephone companies serve as data trans-
        mission channels between (a) branches, stations, and
        local post office data collection sites (b) post of-
        fice data collection sites and teleconcentrator
        sites, and (c) teleconcentrator sites and ADPCs.
        Alternate data transmission channels are provided
        for use between the post office data collection sites
        and the teleconcentrator sites.

AUTOMATIC DATA PROCESSING CENTERS

     Each of the two ADPCs in the system consists of dual
large-scale computer systems. This redundancy provides
backup capacity in the event that one system fails. At the
ADPC data is received, processed, manipulated, and stored,
and reports are prepared and dispatched to remote locations.
Each data message received is edited and validated; employee
files are updated; payroll information is computed, formatted,
and transmitted to the appropriate postal data center for
the actual preparation of salary checks; labor distribution
data is compiled; and mail volumes are computed.

     All data messages received and stored in the ADPCs are
also recorded on magnetic tape as a safeguard against loss
of data should the disc files used for working storage mal-
function. (See pictures on pp. 60 and 61.)

OUTPUT DEVICES

     Each PSDS post office, teleconcentrator site, and ADPC
is equipped with a high-speed printer terminal which pro-
vides printouts of any messages or reports directed to the
particular site where the printer terminal is located. (See
picture on p. 58.)




                             51
APPENDIX III




                 PICTURES OF SELECTED

               PSDS EQUIPMENT COMPONENTS




                          52
                                                                    APPENDIX III




                                                I-
                                                C=
             z           ~~~I-           ICi    p
             Cj~~~            =    I=
                                   =
                  Lai               _j                                           =    CL.~~~~~~La
       C,)                 LU            v)                         P:   L       ~    0,

             CD            2        &                          2C   Ca           A-




LAL

                                   ,c    LJJ6        J0                          =        C


LLML

                                                            c.oo
LLL
       LU                                                 Z I-L Z   ~C       r




       Ck5


                        LuC




                                  53u
APPENDIX III




        L3                    L

        I2l C~                L16I


                                   LUJ
        LLJ                   LLJ


          :23CI
        CCl


                                        2c



   I-
    z~                            CCD-C



                                  Icl




                CDWLa             C,


               I-




              2c             c5



                        54
                                                    APPENDIX III




                    z!    C.3    o=   0         Zu.
                    Co           o    c~~~~C=   Z
                -   0..    _nP




           CD             W.~1
    w      i ic~·
                                                          OctI
                    CZ




o

II.




      I-




                                 55
APPENDIX III




                   C-,
                   C,

LU          LJ
            .rr

            V)

            A...
              ui                                          w C




           CJE
            YI
            LLJL               UI J~~~~~~~~~~Z~~U             LU


 -~~~~
 w~wV                          uj       2m~~~~~~~~~~-I




                                        O                         ~~~~~~~~~~~~~~~~~

 z~                                                                    I-
                                                         JI
     NEI                                                      r    C

                                u~~~~~~~~~~~r~~~0

                                                                       0-


                         A.~~~~~~~~~~~~~~~~~~~~~~~~L
                                      56~~~~eb



                          4~        /
                                           APPENDIX III




'U

     V.i




z




W-
'U




           LL'   0        n       _    '
                           C       -
                     LL   L~~$?
APPENDIX II I




               _ l 101
           | | !11  IS~~~~~~71
                                                           |_0E
i·;ll        I·--           ;, :0                   i^1-
1I
           4i-: I 1
             Iz               H

                               _        -~~~~~~;
 i~:--
               I_     _wc-:·
                       ·:aa
                                       -··
                                         .g:~~~b:
                 ,·i- 1.·
         i·:;i·.i~~*~
               la    _

                                   l _ XFist~5
     APPENDIX III




                U--




                      i


                          4,




59
APPENDIX III




                                             a
                                             U




   i                   \t   i                w

 ! ,   '       "   0




                                 60~~~~~~~




                            60
 APPENDIX III




  ~.   . .   .-... :;       I-




                                 a.



                                 I-

                                 w
                                 z



                                     -3

-61                              <




                                 U

                                 z

                                  cU
                                  0


                                     U

                                 0




             ,f,        X   t
APPENDIX IV



                          h'r 3postmadat Meati

                            September 9,   1970




  Dear Mr.    Neuwirth:

  Thank you for your letter advising the results of your review
  of the Postal Service Data Syste.n and for the opportunity to
  comment on your proposed report.

  Your study, which has been helpful to us, confirms many of the
  findings previously reported by our own internal audit. The
  difficulties, delays, and added costs involved in establishing the
  system, as experienced in the period since the decision was made
  and the program initiated in 1966, are a matter of concern to our
  management.

  Our concern is twofold: First, to assure that a fully productive
  system is completed and, second, that any deficiencies in plans or
  procedures which may have occurred in the past have been completely
  eliminated. I believe that we are taking necessary actions in this
  dir ection.

  The original project envisioned a system which would readily make
  available to management essential data relating to employee time
  and attendance for pay purposes, the volume of mail handled, and
  workload distribution for operational and cost control purposes.
  This was the initial objective at the time the decision was made in
  1966 and was the basis for the planning and procedures that followed.

  As you point out, difficulties have been experienced which impaired
  the effectiveness ofhhe system to date. In recognition that some
  mistakes were made, we do not, however, agree that all could have
  been avoided. There are reasonable explanations which account for
  many of the items covered in your report.

  The urgent need within the Postal Service for improved management
  and control information apparently caused the adoption of an



                                62
                                                           APPENDIX IV


accelerated and possibly unrealistic schedule which was extremely
difficult to meet. The development as well as implementation of
the system were to be accomplished concurrently to a significant
degree. In a developmental environment, especially in a system as
complex as this, it is not uncommon that schedules originally fixed,
especially on an accelerated basis will slip -- systems specifications
will require altering --   and change in equipment requirements and
reconfiguration will be necessary.

A management decision was reached during the development of the
 system to revise the workload and labor distribution factor. This
decision substantially affected the system being developed and equip-
ment requirements. The new workload recording system is designed
to enable postal management to receive, on a timely basis, information
of far greater value in the efficient control of operations. The additional
costs for equipment over the costs originally projected, are also
accounted for in part by the need, which clearly developed as the system
progressed, to embrace a far greater number of employees and to pro-
vide a level of redundancy of equipment necessary to support the
continuous uninterrupted operation of the system.

Over the past 12 months, our management has introduced considerable
improvements to increase the reliability and effectiveness of computer
processing operations and the timely submission of report data to
post offices. A re-design of both the on-line and off-line computer
programming systems, the introduction of extensive quality control
checks to assure continuous integrity of data, and adjustments in
equipment components, are among the changes made in recent months
for improved operations. We believe these efforts have, to a signifi-
cant degree, overcome the difficulties cited in your draft report.

We continue to expect that over the coming months many of the direct
and indirect benefits projected earlier will be realized from the appli-
cation of the mail volume, labor distribution, and time and attendance
recording data. A number of other actions are contemplated, including,
but not limited to, the recovery of manpower released upon completion
and satisfactory operation of the system.




                                     63
APPENDIX IV



Your report recommends that any further expansion of the system
be suspended pending a comprehensive evaluation to determine its
efficiency, economy, and operational reliability. We concur
generally with your recommendation with the exception, however,
that a limited further expansion to 110 offices must be completed if
we are to avoid further adverse effect on our overall objective of
providing information useful to the control of operations in our large
post offices. Moreover, commitments were made and obligations
incurred some time ago for most of the equipment involved in the
limited expansion.

I have personally reviewed this matter with the Assistant Postmaster
General, Bureau of Finance and Administration, and the Assistant
Postmaster General, Bureau of Operations. Both have assured me
of the substantial benefits that are forthcoming from the operation
of this system. Moreover, they look ahead to the system also serving
in a number of other activities which have generated great volumes
of paperwork such as the motor vehicle reporting and inventory control.

Further detailed explanations of each of the specific items mentioned
in your report, because of their complexities, can best be delivered
and understood in mutual discussions between your representatives
and ours. I suggest that this be done.

                                   Sincerely,




                                   Winton M. Blount




Mr. Max A. Neuwirth
Associate Director, Civil Division
U. S. General Accounting Office
Washington, D. C. 20548




                                     64
                                                               APPENDIX V


                        PRINCIPAL MANAGEMENT OFFICIALS OF

                           THE POST OFFICE DEPARTMENT

                        RESPONSIBLE FOR ADMINISTRATION OF

                            POSTAL SOURCE DATA SYSTEM


                                                  Tenure of office
                                                  From             To

POSTMASTER GENERAL:
       Winton M. Blount                       Jan.      1969   Present
       W. Marvin Watson                       Apr.      1968   Jan. 1969
       Lawrence F. O'Brien                    Nov.      1965   Apr. 1968

DEPUTY POSTMASTER GENERAL:
    Vacant                                     Jan.     1971   Present
    Elmer T. Klassen                           Feb.     1969   Jan. 1971
    Frederick C. Belen                         Feb.     1964   Jan. 1969

ASSISTANT POSTMASTER GENERAL
    FINANCE AND ADMINISTRATION
    DEPARTMENT (note a):
      James W. Hargrove                        Feb.     1969   Present
      Ralph W. Nicholson                       Mar.     1961   Feb. 1969

ASSISTANT POSTMASTER GENERAL
  OPERATIONS DEPARTMENT (note a):
    Frank J. Nunlist                           Apr.     1969   Present
    Vacant                                     Dec.     1968   Apr. 1969
    William M. McMillan                        Feb.     1964   Dec. 1968

aEffective December 1, 1970, all bureaus within the postal
 establishment were redesignated as departments.




U.S. GAO, Wash., D.C.

                                        65