Postal Service: Automation Is Taking Longer and Producing Less Than
Expected (Briefing Report, 02/22/95, GAO/GGD-95-89BR).
The U.S. Postal Service must overcome difficult, if not insurmountable,
obstacles if it is to successfuly introduce optical scanning into its
operations by 1998. Barcoding of letter mail and automatic sorting of
letters to homes and businesses, known as "delivery point sequencing,"
has proven more difficult than the Postal Service expected and lags
behind schedule. The savings from automation continue to be small
compared to overall labor costs and more difficult to achieve than the
Postal Service anticipated.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: GGD-95-89BR
TITLE: Postal Service: Automation Is Taking Longer and Producing
Less Than Expected
DATE: 02/22/95
SUBJECT: Mail delivery problems
Optical equipment
Postal service
Postal rates
Postal service employees
Federal agency reorganization
Operations analysis
Systems conversions
Reductions in force
Cost control
IDENTIFIER: USPS 1995 Corporate Automation Plan
ZIP Code System
ZIP-Plus-4
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* No attempt has been made to display graphic images, although figure *
* captions are reproduced. Tables are included, but may not resemble *
* those in the printed version. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Briefing Report to Congressional Committees
February 1995
POSTAL SERVICE - AUTOMATION IS
TAKING LONGER AND PRODUCING LESS
THAN EXPECTED
GAO/GGD-95-89BR
Postal Service
Abbreviations
=============================================================== ABBREV
AFCS - Advanced facer canceler system
CAP - Corporate Automation Plan
CSBCS - Carrier sequence barcode sorter
DBCS - Delivery barcode sorter
ID - Identification
IPSS - Image processing subsystem
ISS - Input subsystem
MLOCR - Multiline optical character reader
MPBCS - Mail processing barcode sorter
OCR - Optical character reader
OSS - Output subsystem
RBCS - Remote barcoding system
RCR - Remote computer reader
SLOCR - Single line optical character reader
WABCR - Wide area bar code reader
ZIP - Zoning Improvement Plan
Letter
=============================================================== LETTER
B-258949
February 22, 1995
The Honorable Ted Stevens
Chairman
The Honorable David Pryor
Ranking Minority Member
Subcommittee on Post Office and Civil Service
Committee on Governmental Affairs
United States Senate
The Honorable John M. McHugh
Chairman
The Honorable Barbara Rose-Collins
Ranking Minority Member
Subcommittee on the Postal Service
Committee on Government Reform and Oversight
House of Representatives
This briefing report provides information on the U.S. Postal
Service's progress in using optical scanning technology to achieve
its goals of (1) barcoding virtually all letter mail, (2)
automatically sorting mail to individual home and business addresses,
and (3) adjusting work methods and employment to achieve workforce
reductions.
On February 8 and 9, 1995, we briefed your Subcommittee on setbacks
in the Service's automation program and obstacles it must overcome to
achieve current automation goals. This briefing report summarizes
the information presented in that briefing.
BACKGROUND
------------------------------------------------------------ Letter :1
The Postal Service began automating letter mail processing in 1982,
acquiring optical character readers, which could read ZIP Codes and
spray barcodes on letters, and barcode sorters for detailed sorting.
In 1988, the Service began offering lower postage rates to larger
mailers who did their own barcoding. By 1997, the Service plans to
deploy up to about 14,000 pieces of automation equipment costing
about $5 billion. Historically, the Service has viewed automation as
a principal means of controlling labor costs, which for 1994 amounted
to about $39.6 billion or 81.7 percent of annual operating expenses.
In 1992, we pointed out that automation, while producing some
savings, was unlikely to be a panacea that will reverse the tendency
for postal costs to outpace inflation. Through January 1995, we had
issued eight products on Postal's automation program. A listing
appears at the end of this report.
Following a major reorganization in 1992, the Service discontinued
much of automation's central management oversight and decentralized
decision-making to the field. Because field managers complained
about the lack of central direction and emphasis in 1993, the Service
reestablished a central office to oversee automation and prepare a
new Corporate Automation Plan, which is expected to be completed by
March 1995.
RESULTS
------------------------------------------------------------ Letter :2
The Postal Service must overcome difficult, if not insurmountable,
obstacles to successfully complete the program by their projected
date of 1998. Barcoding of letter mail and automatic sorting of
letters to homes and businesses, referred to as "delivery point
sequencing," has proven more difficult than the Service expected and
is behind schedule. The savings from automation continue to be small
compared to overall labor costs and more difficult to achieve than
the Service anticipated.
BARCODING PROBLEMS NUMEROUS
AND VARIED
---------------------------------------------------------- Letter :2.1
In 1988, the Service set a goal for mailers and the Service to
jointly barcode virtually all letter mail by 1995--or about 90
percent of the more than 115 billion letters processed annually. To
meet that schedule, about 80 billion letters were to have been
barcoded in 1993 by mailers and the Postal Service. Mailers barcoded
30 billion letters that year, exceeding the 29-billion benchmark set
by the Service. However, the Service fell short of its 51-billion
benchmark, barcoding 30 billion letters. Optical character readers
cannot read about half of the letters they process because of the
wide variation in the physical characteristics of envelopes and
addresses. While the Service encourages standardization, only
mailers who receive volume discounts for barcoding their mail are
required to comply with the Service's automation standards.
Another factor contributing to the barcoding shortfall was delayed
deployment of "remote barcoding," whereby clerks key in addresses
that cannot be read by the equipment. The delay occurred when plans
to use contract employees for remote barcoding were thwarted by a
union-management contract dispute. In 1993, remote barcoding was
expected to generate about 13 billion barcoded letters from an
83-site system, but the actual number was 25 sites and 3 billion
barcoded letters.
Postage rates are out of sync with the Service's automation plans
because incentives offered to mailers give preference to presorting
over barcoding. Before delivery point sequencing, it made sense for
the Service to give a discount to mailers who put their mail in
delivery sequence because it saved time for the carrier. However,
now that the Service must sequence the mail, there is less value in
having individual mailers do it. The Service is also limited in its
ability to verify that letters barcoded by mailers, and receiving
discounted postage rates, have the right barcode and can be sorted on
Service equipment.
In recognition of the barcoding difficulties, the Postmaster General
announced in April 1994 that the barcoding goal had slipped from 1995
to the end of 1997.
NOT ENOUGH MAIL IS SORTED IN
DELIVERY SEQUENCE
AUTOMATICALLY
---------------------------------------------------------- Letter :2.2
Barcoding allows letters to be sorted in delivery sequence, which
should significantly reduce the time that carriers spend manually
sorting letter mail. For many routes, there has not been enough
delivery-sequenced letters given to the carriers. Consequently, they
often continued to manually sort mail already sorted by automation.
The Service has not reduced carrier in-office workhours or adjusted
routes as intended, and no longer assumes carriers will be able to
reduce by 2 hours a day the time they spend preparing mail for
delivery.
About 19 percent of deliveries are to multiple-occupant buildings
where the Service has not yet implemented delivery point sequencing.
Until the Service extends delivery sequencing to the addresses within
these buildings, which has been deferred indefinitely, carriers will
need to continue to manually sort this mail.
AUTOMATION NOT REDUCING
WORKFORCE
---------------------------------------------------------- Letter :2.3
Expecting automation to reduce the postal workforce, the Service in
1992 reaffirmed a 1989 goal of an outright reduction of over 40,000
workyears from the program. This has not happened because of
barcoding and sorting problems highlighted above and the Service's
inability to change work methods and reduce employee workhours after
equipment was deployed. Moreover, a steady drop in career
employment, which began in 1989, was disrupted by a downsizing in
1992 unrelated to automation. The downsizing proved temporary
inasmuch as in 1994 career employment surpassed its pre-downsizing
level and is still increasing, offsetting much of the earlier
reduction. Postal officials said that employment reductions did not
occur because the Service's workload increased. However, we noted no
major difference between actual workload growth and the growth
projected in the Service's automation plans.
The Service also has departed from its automation strategy of
replacing career employees with less costly, short-term transitional
employees who were to fill in until automation reduced employment.
Since the downsizing, the transitional workforce has decreased while
career employment has increased.
Business-related mail that best lends itself to automated processing
is also the most vulnerable to being diverted to electronic
communication, such as e-mail and electronic data interchange. As
this diversion occurs, the benefits of automation are likely to
diminish.
OBJECTIVES, SCOPE, AND
METHODOLOGY
------------------------------------------------------------ Letter :3
Our objective was to determine the status of the Postal Service's
letter mail automation efforts, including its progress in achieving
the goals of (1) barcoding virtually all letter mail, (2)
automatically sorting mail to individual addresses, and (3) adjusting
work methods to reduce the postal workforce.
To determine the automation program's status, we analyzed prior
studies, reviewed the Service's written corporate automation plans
and decision analysis reports supporting the acquisition of
automation equipment. We also reviewed related Postal Inspection
Service audit reports, results of postal employee opinion surveys,
and briefings provided by Service officials to the Board of
Governors. In addition, we interviewed Postal Service headquarters
officials responsible for overall automation planning and management
and conducted on-site work at several facilities in the Service's
Western Area. We did not determine the extent to which automation
has improved postal productivity because Postal Service data did not
permit us to isolate the effects of automation versus other factors.
Our review followed generally accepted government auditing standards
and was done from October 1993 through September 1994.
---------------------------------------------------------- Letter :3.1
A draft of this report was provided to senior Postal Service
officials in December 1994 and we incorporated their comments and
clarifications where appropriate.
We are sending copies of this report to the Postmaster General and to
Members of Congress and congressional committees that have
responsibilities for Postal Service issues. Copies will also be made
available to others upon request.
The major contributors to this briefing report are listed in appendix
II. If you have any questions about this report, please call me on
(202) 512-8387.
J. William Gadsby
Director, Government Business
Operations Issues
Briefing Section I BACKGROUND
============================================================== Letter
Automation Overview
(See figure in printed
edition.)
Automated mail processing involves converting printed and electronic
addresses to barcodes and sorting letters according to the barcode
using high-speed equipment. Barcodes can be applied by the Postal
Service, either by automated mail processing equipment or keyers
reading addresses that the equipment cannot decipher (called remote
barcoding), and by mailers who may
Not All Letter Mail Is Automation Compatible
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
Letters make up about two thirds of all mail, and about 90 percent of
the letters are of acceptable size and shape to be processed on
automated equipment. The Service's goal is to be able to barcode
these letters by the end of 1997. The total mail volume
Automation Equipment Deployment Schedule Spans 16 Years
(See figure in printed
edition.)
Note: Postal estimates the
total cost of the automation
program will be $5 billion.
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
EQUIPMENT AND COST
---------------------------------------------------------- Letter :3.2
The automation program requires basically two types of equipment:
(1) optical character readers that read addresses, translate them
into barcodes, spray the barcodes on envelopes, and do some initial
sorting; and (2) barcode sorters that read the barcodes and do the
detailed sorting. As the program has evolved, various modifications
and other equipment have been developed and added, including a remote
barcoding capability where workers read addresses that the optical
character readers cannot. Over a 16-year period ending in 1997, up
to about 14,000 pieces of letter automation equipment, costing $5
billion, are planned to
Evolution of ZIP Codes and Barcodes
(See figure in printed
edition.)
CURRENT CODING EVOLVED FROM
1963 ZIP CODE
---------------------------------------------------------- Letter :3.3
In 1963, the five-digit Zoning Improvement Plan (ZIP) Code was
devised to provide a numbering system that could guide mail to any
U.S. post office. By 1978, the code was used on approximately 97
percent of all domestic mail. Most letter mail was sorted on large
letter sorting machines that relied on operators to manually key in
the ZIP Code before the mail piece could be sorted. When mail
arrived at the post office specified by the 5-digit code, a machine
operator who had memorized street names and numbers for each
carrier's route sorted the letters to carriers. The Service also
offered business mailers lower postage rates for presorting their
mail, allowing it to by-pass certain steps in the Postal Service.
In 1983, the Service added four digits to the ZIP Code, (ZIP+4),
enabling mail to be sorted to a segment of a carrier route without
any special knowledge, or, if warranted, to certain finer
designations, such as one side of the street in a block, a business,
building, or post office box section.
Barcoding was introduced with the 9-digit ZIP Code, thus permitting
letters to be sorted automatically. In 1993, the Service instituted
the 11-digit barcode by adding the last two numbers of the street
address to the 9-digit ZIP Code; it enables
Congressional/GAO Interest
(See figure in printed
edition.)
CONGRESSIONAL AND GAO
INTEREST IN POSTAL
AUTOMATION
---------------------------------------------------------- Letter :3.4
Conversion to the 9-digit ZIP Code was controversial. Business
mailers believed that adding 4 digits to the 5-digit ZIP Code was an
imposition whose costs and inconvenience would exceed any savings.
Congress postponed the 9-digit code until October 1983 while we
reviewed the code's planned implementation and the Service
re-examined its internal options for the expanded ZIP Code. In 1983,
we reported that the 9-digit ZIP Code could be cost-effective if the
Service (1) demonstrated successful performance of the equipment and
(2) established rate incentives for volume mailers that reasonably
assured sufficient use of the 9-digit code.
The Service initially acquired single-line optical character readers
(OCR) that only read the 9-digit ZIP Code line of the address. We
reported that their success thus depended upon mailers applying these
codes in large numbers. But this did not occur. As the Service
considered converting to multiline readers, we reported on the pros
and cons of single versus multiline optical character readers. The
Service decided in 1985 to replace the single-line readers with
multiline readers which could read up to five lines of a printed
address. At that point, the optical character readers no longer
needed ZIP Codes to apply barcodes; however, they work more
efficiently if a correct ZIP Code is present. We subsequently
reviewed the Service's conversion to multiline readers.
In 1992, we pointed out that automation was producing some savings
but was unlikely to be a panacea that would reverse the
Inconsistent Management Attention
(See figure in printed
edition.)
INCONSISTENT MANAGEMENT
ATTENTION TO THE AUTOMATION
PROGRAM
---------------------------------------------------------- Letter :3.5
The Postal Service began major deployments of automated equipment in
1982, and in 1987 it began formulating a formal program with an
oversight office, a written plan, results tracking, timetables, and
goals. However, much of the central management oversight was
discontinued in 1992 after a major reorganization under Postmaster
General Runyon, and more decision-making was delegated to the field.
The number of staff responsible for planning and implementation of
automation was reduced.
Although the Service is restoring features of the earlier automation
program, momentum has been lost. Parts of the automation program
abolished due to the change in policy included
Automation Implementation Management Department at Postal
Headquarters,
management support structure in the field, and
tracking system that measured automation equipment use and
performance.
In 1993, field managers commented that the automation program
appeared to be on hold or in hibernation. More than a third of
employees responding to an opinion survey in 1993 did not believe
that the Service was doing a good job of implementing automation.
The percentage of executives who believed that the Service was doing
a good job of implementing the program decreased from 53 percent in
1992 to 39 percent in 1993. A central automation office was
reinstated in October 1993 to update the Corporate Automation Plan
and, in coordination with operating functions, develop and implement
a process to manage automation. The new office expects to complete a
new Corporate Automation Plan by March 1995.
Briefing Section II BARCODING
PROBLEMS NUMEROUS AND VARIED
============================================================== Letter
Letters Barcoded in 1993 Fell Short of Benchmark
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
BARCODING IS BEHIND SCHEDULE
AND HAMPERED BY PROBLEMS
---------------------------------------------------------- Letter :3.6
In 1988, the Postal Service set a goal and began developing a
detailed Corporate Automation Plan (CAP) to barcode virtually all
letter mail by the end of fiscal year 1995. The 1992 revised plan
included benchmarks indicating the volume of mail to be barcoded each
year by the Service and by mailers to achieve the 1995 goal. The
benchmark for 1993 was a total of 80 billion letters. The Service
and mailers together barcoded 60 billion letters in 1993. Mailers
barcoded more letters than expected. Most of the 20 billion
shortfall was due to the Service not barcoding its share of letters.
The remote barcoding system (RBCS) was expected to generate about 13
billion barcoded letters from an 83-site system, but the actual
number was 25 sites and 3 billion barcoded letters. Actions the
Service is taking to increase barcoding are discussed in subsequent
sections of this briefing.
Postal Service Barcoding Problems
(See figure in printed
edition.)
POSTAL SERVICE BARCODING IS
HAMPERED BY INADEQUATE
ADDRESS INFORMATION
---------------------------------------------------------- Letter :3.7
OCRs are expected to derive barcodes from addresses that must be (1)
machine readable and (2) verifiable against the Postal Service's
official address directory. Overall, the Service's readers reject
about half the letters they receive for barcoding. Because the
readers reject letters with incorrect or unrecognized addresses, such
letters must either be barcoded by slower and more costly methods or
processed without a barcode, which is the most expensive sorting
method.
Because barcoding discounts are not available to most mailers (e.g.,
a minimum of 500 First Class letters per mailing to qualify), they do
not have to conform their address databases with the official
directory. As a result, many letters have addresses that cannot be
barcoded by the Service's optical character readers. Many others
cannot be read due to the color or composition of the paper, or other
design characteristics. Moreover, handwritten addresses cannot be
read by optical character readers and will be barcoded by remote
barcoding.
The Service is adding computer and camera enhancements to OCRs that
it believes will improve its ability to read and bar code
(See figure in printed
edition.)
INCENTIVES GIVE PREFERENCE
TO PRESORTING OVER BARCODING
---------------------------------------------------------- Letter :3.8
Postage rates are not entirely compatible with the current Postal
Service automation plans. The Postal Service offers a discount for
barcoding mail. But it offers an even larger discount if mailers
presort their mail to the carrier route or in delivery sequence.
Before the implementation of delivery point sequencing in March 1993,
it made sense to offer the largest discount for mail presorted to
carrier routes. However, with delivery point sequencing, this is not
the case. Maximum workhour savings accrue when a carrier receives a
single batch of letters already sorted in delivery point sequence,
which can only be done after the Postal Service receives and makes
final sorts of mail from mailers. The benefits of delivery point
sequencing can be reduced or lost when carriers receive multiple
trays of presorted mail from different mailers. Such mail must be
merged either by automation using the Service's sorting equipment or
manually by the carrier. If neither is done, the carrier must go on
the street with multiple groups of letters, which can increase
delivery time and defeat the purpose of automated sorting. Yet
presorted mail currently receives the largest discount from the
Postal Service.
The Postal Service wants to reclassify the postage rates so they more
closely reflect the level of service given, the use of automation
technology, and the point at which mail enters the system. The
Service expects to propose mail classification revisions to the
Postal Rate Commission in 1995, with rates that, among other things,
will favor barcoded mail over presorted mail. However, presort
discounts are well established. Because mailers and presort firms
have committed their time and money to the existing rate structure,
it will take time to change the basic
(See figure in printed
edition.)
MOST MAILERS NOT ELIGIBLE
FOR BARCODE DISCOUNT
---------------------------------------------------------- Letter :3.9
The Postal Service offers no financial incentive but only
encouragement for most mailers to prepare mail that can be sorted on
the Service's equipment. The minimum quantity of letters to qualify
for a barcode discount is a single mailing of 500 pieces for First
Class and 200 for third class. In 1991, the Postal Rate Commission
proposed a public automation rate discount of 2 cents for anyone
using barcoded reply envelopes often provided by business mailers for
bill remittances. But the Postal Service opposed this proposal
because of the difficulties it anticipated it would face in educating
the public about a new rate and ensuring that correct postage was
used. The Service's current plans for reclassifying the rate
structure do not extend prebarcode discounts to smaller volume
mailers.
The minimum quantities to qualify for discounts were first
established in the 1960s and, according to postal officials, have no
current analytical basis. In the meantime, recent advances in
desktop computing have made it relatively inexpensive and easy for
smaller volume mailers to address letters that meet Postal
Mail Preparation Requirements Discourage Barcoding
(See figure in printed
edition.)
MAIL PREPARATION
REQUIREMENTS MAY DISCOURAGE
MAILER PARTICIPATION
--------------------------------------------------------- Letter :3.10
The Service's requirements for mailers to qualify for automation
discounts do not encourage barcoding or necessarily ensure savings.
The requirements are lengthy, complex, and hard to understand. When
mailers submit mail for a discounted rate, it is not easy for the
Service to ensure that the mail meets its specifications and can be
processed on automation. In 1994, the Service had about 3,000
employees whose job, among other duties, was to determine that large
mailings actually qualify for prebarcoded and/or presorted mail
discounts. In many cases, special equipment is required (but not
always available) because flaws cannot be detected visually. Because
of the sheer volume of this mail, the Service's effectiveness in
screening it for correct postage is unknown. We are looking at this
area in a separate review.
The Service offers assistance to mailers on how to prepare their
mailings properly. It also has an effort under way to improve these
field operations by streamlining mailing requirements functions,
simplifying regulations, and increasing training for customers and
employees. One goal is to reduce the administrative workload of mail
requirements personnel so that more time can be spent on such
activities as customer training
Remote Barcoding More Expensive Than Planned
(See figure in printed
edition.)
REMOTE BARCODING METHOD MORE
EXPENSIVE THAN PLANNED
--------------------------------------------------------- Letter :3.11
One automation problem involves how to put barcodes on letters that
have not been barcoded by mailers and have handwritten or poorly
printed addresses that cannot be read by the Service's OCRs. The
Service's solution is a function called remote barcoding; it entails
making an electronic image of letters that optical character readers
cannot read. The images are electronically transmitted to data entry
clerks who enter enough address information to provide the
appropriate delivery point code. This information is transmitted
back to the processing facility and a modified bar code sorter then
matches the new information to the letter and applies a barcode.
The Service initially planned that remote barcode keying would be
contracted out because (1) these workers would not touch the mail,
(2) the work would be part-time employment, and (3) technological
advances in optical character recognition would enable equipment to
read this mail and eventually phase out the remote barcoding.
Twenty-five remote barcoding sites opened initially in 1992 and 1993,
with plans to ultimately have 268 sites. In 1990, The American
Postal Workers Union and the National Association of Letter Carriers
challenged the Service's remote barcoding plans. In May 1993, an
arbitrator decided that the Service's contract with the Union
required the Service to first offer the keying work to postal
employees before contracting it out. In November 1993, the Service
and the Union agreed that the work would be done by postal employees,
30-percent career employee workhours and 70-percent transitional
employee workhours. Because postal employees are paid significantly
more than contract workers, the agreement will reduce the savings
anticipated from remote barcoding. Early in 1994, the Service
reactivated deployment of remote barcoding
(See figure in printed
edition.)
REMOTE BARCODING METHOD MORE
EXTENSIVE THAN PLANNED
--------------------------------------------------------- Letter :3.12
The Service now believes that no major breakthrough in character
recognition is imminent. In 1992, it revised its approach to
research in this area and focused more on products that could be
added to the Service's existing equipment rather than creating whole
new machines, and on products that would not otherwise be funded by
industry. Because OCRs are not barcoding as many letters as
expected, and because no breakthroughs are expected in this
technology, the Service will be relying on the more expensive remote
barcoding function more than planned and for much longer than
planned.
The Service is adding a new piece of equipment, the remote computer
reader (RCR), to remote barcoding systems that is expected to
increase automatic bar code production and reduce encoding costs.
RCR uses the same technology as the OCR but has more time to decipher
an address image than the OCR does. If successful, it will reduce
the number of images that need to be
Increased Use of Remote Barcoding and Reduced Dependence on Mailers
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
INCREASE IN RBCS USE AND
REDUCED ROLE FOR MAILERS AND
OCR
--------------------------------------------------------- Letter :3.13
In March 1994, the Postal Service changed its barcoding projections
to increase the use of remote barcoding and reduce dependence on
mailers and the Service's OCRs for barcoding. The Service originally
planned that RBCS would barcode 17 percent of the letter mail.
However, the Service now expects this method to barcode 28 percent.
According to Postal Service officials, the increased reliance on RBCS
is due in part to two factors: (1) lower than expected OCR barcoding
performance due to a degradation of the mailbase and (2) technology
enhancements in remote barcoding. The degradation of the mailbase
occurred as more of the easier to barcode mail became customer
prebarcoded. The remaining mail is more difficult for the Service's
OCRs to barcode. According to Service officials, a team has been
formed to address the problem by persuading mailers to improve
address accuracy and readability.
Briefing Section III OBSTACLES TO
ACHIEVING DELIVERY POINT
SEQUENCING
============================================================== Letter
(See figure in printed
edition.)
DELIVERY POINT SORTING IS
NOT AVAILABLE FOR MULTIPLE
OCCUPANT BUILDINGS
--------------------------------------------------------- Letter :3.14
The 11-digit barcode identifies the street number of the address so
that letters can be sequenced according to the carrier's route.
However, many carrier routes have large numbers of apartments and
suites, and most mail for these units cannot now be sorted in
delivery sequence. In implementing delivery point sequencing, the
Service gave priority to residential addresses, and apartments and
other units in multiple-occupant buildings were deferred. This was
because of the additional complexity involved in interpreting the
address information and because the Service believed that greater
savings could be achieved by concentrating on residential deliveries.
About 23 million delivery points are affected, about 19 percent of
total deliveries. The Service expects to eventually include
(See figure in printed
edition.)
CAPTURING CARRIER SAVINGS
BEHIND SCHEDULE
--------------------------------------------------------- Letter :3.15
When carriers receive their letters in delivery point sequence, the
time they spend preparing mail for delivery should be reduced. This
would allow the carriers to cover larger routes, resulting in a
savings in workhours. The Service originally expected that 2 hours
of mail preparation time could be saved per carrier.
Achieving the carrier savings expected, however, will require the
cooperation of the carriers and the unions representing them. In
September 1992, the Postal Service and the National Association of
Letter Carriers issued joint agreements that resolved past disputes
on the subject and provided a plan for setting delivery point
sequencing volume targets and rules for realigning carrier routes to
capture savings. As the Service gained experience with
implementation, it became clear that original target volumes to
trigger major realigning of routes were too high. In January 1994,
negotiations on updating the joint agreements broke down, leaving
delivery units in various stages of their plans. As a result, the
Postal Service, using the existing agreements, issued detailed
delivery point sequencing management strategies to the field,
including changes to target percentages for interim route adjustments
and new directions on when carriers should cease casing automation
sequenced mail and carry it directly to the street. The Service also
amended the handbooks to provide for route inspection procedures to
include automation sequenced mail.
The Service no longer anticipates that carriers will save 2 hours a
day preparing mail for delivery, and now uses a savings estimate of
80 minutes per day based on standard sorting rates
(See figure in printed
edition.)
OBSTACLES TO CAPTURING
CARRIER SAVINGS
--------------------------------------------------------- Letter :3.16
At the end of fiscal year 1994, after a year and a half of
implementation, 50,112 routes, about 30 percent of all city routes,
were receiving letter mail sequenced on automation, and carriers for
about 33 percent of the routes continued to manually sort letters
already sequenced by automation. However, daily consistent volumes
were insufficient to carry out major route realignments. Frequently,
target offices received less than 40 percent of the letter mail in
delivery point sequence order. And, overall, time spent in the
office by carriers in 1994 was up 6.2 percent over 1993, while volume
was up 3.4 percent.
The Postal Inspection Service studied delivery point sequencing at
several of the first delivery units to implement the new method, and
thus should have been the furthest along in its use. The Inspection
Service concluded that not only was there no significant savings in
carrier in-office workhours, but also that the amount of delayed mail
also increased. The chief reasons for the lack of savings were
reported to be the low volume of delivery point mail and the
carriers' distrust of automation's accuracy, which caused them to
continue to sort the mail manually. (A 1993 Postal Service employee
opinion survey indicated that fewer than one fifth of the city
carriers who responded believed the Service was doing a good job
implementing
Space Needed for Sorters
(See figure in printed
edition.)
LACK OF SPACE FOR SORTERS
--------------------------------------------------------- Letter :3.17
A long-standing problem that has required the Service to revise its
plans and that can reduce expected cost savings is the lack of space
at mail processing plants or delivery stations for all the barcode
sorters needed. Most of the delivery point sorting was to be done by
a piece of machinery called the delivery bar code sorter. The
Service estimated that it needed about 5,600 of these. However,
after committing to about 2,400 delivery bar code sorters (DBCS), the
Postal Service realized that it did not have room for all of the
remainder. The Service then developed plans for a smaller barcode
sorter. However, it must buy more of these than the original pieces
of equipment because they have less capacity. This still has not
solved the space problem entirely. The Service currently estimates
that it lacks space for about 20 percent of the needed delivery
sequence sorting capacity.
Briefing Section IV AUTOMATION NOT
PRODUCING ANTICIPATED REDUCTIONS
============================================================== Letter
Postal Service's Expected Savings
(See figure in printed
edition.)
SAVINGS HAVE NOT MET
EXPECTATIONS
--------------------------------------------------------- Letter :3.18
The Service developed a detailed Corporate Automation Plan in 1989,
which it last updated in 1992. In so doing, the Service predicted
savings of from 84,000 to 100,000 work-years by 1995, including an
outright reduction of over 40,000 work-years. Instead, work-years
have increased, both overall and for work such as letter sorting that
is most directly affected by automation. Work-years in manual letter
sorting, which should decline as the Service turns to automation,
actually increased 5.3 percent in 1993 and 2.9 percent in 1994.
Postal officials said that increased mail volume and deliveries
account for the growth in work-years. However, workload growth is
normal in the Postal Service, varying according to rate increases and
the general health of the economy. The Corporate Automation Plans
took workload increases into account. For example, the 1990 plan
estimated that volume growth would average 1.7 percent a year from
1990 to 1995; through 1994, it has grown
Career Employment Dropped but Has Returned to Pre-reorganization
Level
(See figure in printed
edition.)
CAREER EMPLOYMENT
FLUCTUATIONS
--------------------------------------------------------- Letter :3.19
By far the largest group contributing to total work hours are career
employees. Career employment steadily declined from a high in May
1989 until a major reorganization and downsizing occurred in 1992,
which was unrelated to the automation program. As a result of the
downsizing, career employment dropped suddenly and quickly in 1992
and early 1993. However, career employment now exceeds the level
achieved before the downsizing and is increasing rapidly, apparently
in an effort to remedy some widely publicized service problems that
the Postal Service encountered after the downsizing.
Change in Career and Transitional Employment
(See figure in printed
edition.)
Note: In November 1992, career
employment equaled 674,983 and
transitional employment equaled
20,218.
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
CAREER VERSUS TRANSITIONAL
EMPLOYEES
--------------------------------------------------------- Letter :3.20
The Service also is no longer following an explicit cost-savings
strategy whereby career employees were to be replaced by less costly
transitional (temporary) employees during the period that automation
was being phased in. The theory was that once automation began
displacing employees, the transitional workforce could be more easily
reduced than career employees, who have significant job protection.
Accordingly, from May 1992 until August 1993, the transitional
workforce grew to over 41,000 employees. However, in a 1993
agreement with the American Postal Workers Union, the Service agreed
to curtail the number of transitional employees. Thereafter,
transitional employment leveled off and began dropping while career
employment increased
Budgets Project Some Automation Savings in 1994
(See figure in printed
edition.)
Source: Postal Service data.
(See figure in printed
edition.)
COST-SAVING IMPACT OF
AUTOMATION HAS BEEN
RELATIVELY INSIGNIFICANT
--------------------------------------------------------- Letter :3.21
From inception of the automation program in 1982, the Service
justified automation on the basis of its potential for reducing
costs. If automation were performing entirely as planned and
producing all the savings that were cited to justify the equipment
initially, the savings would still be a small percentage of total
operating expenses. For example, in 1994, the Service estimated the
budget impact from automation to be a savings of $41 million, or less
than one tenth of a percent--a relatively insignificant amount
compared with cost increases due to higher mail volume ($716 million)
and the higher cost of labor ($1.1 billion) in 1994.
The Service estimated that the total amount of budget savings and
cost avoidance from automation, beginning in 1987 when the Service
began investing in advanced OCRs, through 2005, would be $14.4
billion. This is about 1.3 percent of estimated total operating
expenses for the same period--about $1.1 trillion.\1
(In a recent request to the Postal Service Board of Governors for
approval of RBCSs, the Service also included in its justification
delivery data showing that barcoded mail was delivered more
Electronic Diversion Could Adversely Impact Automation Benefits
(See figure in printed
edition.)
--------------------
\1 Assuming operating expenses continue to grow by 6.1 percent
annually, the actual rate of growth from 1987 through 1993.
THE IMPACT OF ELECTRONIC
COMMUNICATION ON LETTER MAIL
--------------------------------------------------------- Letter :3.22
A factor largely outside the Postal Service's control--the electronic
diversion of letter mail--could adversely affect future benefits from
automation. Mail that best lends itself to automated processing is
business-related correspondence and financial transactions. This is
the mail that is also the most susceptible to being diverted by
mailers to electronic communication such as fax, e-mail, and
electronic data interchange between businesses. Such mail accounted
for about 44 percent of the Service's total volume in 1993. The
Service estimated that business-to-business correspondence and
transaction mail would drop from 30.4 billion pieces in 1988 to 20.5
billion in 1994. While the projected decline was offset by increases
in advertising mail and mail between businesses and households, many
postal experts believe that in coming years a major portion of the
mail clearly will be at risk.
LETTER MAIL AUTOMATION
=========================================================== Appendix I
(See figure in printed
edition.)
MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix II
GENERAL GOVERNMENT DIVISION,
WASHINGTON, D.C.
James T. Campbell, Assistant Director
Leonard G. Hoglan
Loretta K. Walch
Katherine M. Wheeler
DENVER REGIONAL OFFICE
Arleen L. Alleman
Christopher R. Moos
GLOSSARY OF LETTER MAIL AUTOMATION
EQUIPMENT
=========================================================== Appendix 0
ADVANCED FACER CANCELER
SYSTEM (AFCS)
------------------------------------------------------- Appendix 0:0.1
Faces (arranges mail so all addresses and stamps are facing the same
way), cancels, and sorts letter mail into three mail streams:
pre-barcoded letters, OCR readable (typed/machine imprinted) letters,
and hand-written or script letters.
CARRIER SEQUENCE BAR CODE
SORTER (CSBCS)
------------------------------------------------------- Appendix 0:0.2
Sorts barcoded letters into the carrier's walk sequence, i.e., the
order in which letter carriers deliver their routes. Using a
three-pass operation, it sorts one route at a time.
DELIVERY BAR CODE SORTER
(DBCS)
------------------------------------------------------- Appendix 0:0.3
Primarily used to sort barcoded letter mail to the order in which
letter carriers deliver their routes. Using a two-pass operation, it
sorts many routes at a time.
MAIL PROCESSING BAR CODE
SORTER (MPBCS)
------------------------------------------------------- Appendix 0:0.4
Primarily used to process/sort barcoded letters to the destination
post office or 5-digit ZIP Code level only. Can also sort barcoded
letters to the carrier route for larger delivery offices.
MULTILINE OPTICAL CHARACTER
READER (MLOCR)
------------------------------------------------------- Appendix 0:0.5
Reads the entire address block of a letter, consults the address
directory, determines the proper ZIP Code for that address, sprays a
barcode on the letter, and sorts the letter to one of a number of
stackers.
REMOTE BARCODING SYSTEM
(RBCS)
------------------------------------------------------- Appendix 0:0.6
Provides barcoding for mail that cannot be read by OCR. It processes
images of letters lifted by AFCS and MLOCR and determines the correct
barcode, which allows the letter to be barcoded by a barcode sorter
modified with an OSS.
MAJOR COMPONENTS OF RBCS
------------------------------------------------------- Appendix 0:0.7
Image Processing Subsystem (IPSS): Receives images of letters not
read by RCR and transmits the images to the remote encoding sites,
displays video images on video terminals, determines ZIP Codes from
operator-keyed extraction codes, and transmits ZIP Code and ID back
to the decision storage unit.
Input Subsystem (ISS): A modification to AFCS and MLOCR that allows
AFCS and MLOCR to spray an ID tag on the back of each OCR unreadable
mailpiece and then provide a video image to the RBCS for further
processing.
Output Subsystem (OSS): A modification to a barcode sorter that
enables it to interface with RBCS and print a barcode on the letter,
as well as sort the letter based on the barcode.
Remote Computer Reader (RCR): Uses address recognition techniques
similar to the MLOCR but has more time to determine the ZIP Code
information.
SINGLE LINE OPTICAL
CHARACTER READER (SLOCR)
------------------------------------------------------- Appendix 0:0.8
Reads the city/state/ZIP line of the address block, consults the
address directory, and determines the proper ZIP+4 barcode for that
address; prints the barcode on the letter; and sorts the letter to
one of a number of stackers.
WIDE AREA BAR CODE READER
(WABCR)
------------------------------------------------------- Appendix 0:0.9
A modification to MPBCS that permits the sorter to "read" a barcode
virtually anywhere on the letter.
LIST OF RELATED GAO REPORTS
=========================================================== Appendix 1
Conversion to Automated Mail Processing Should Continue; Nine-Digit
ZIP Code Should Be Adopted If Conditions Are Met (GAO/GGD-83-24, Jan.
6, 1983).
Conversion to Automated Mail Processing and Nine-Digit ZIP Code-- A
Status Report (GAO/GGD-83-84, Sept. 28,1983).
Comparative Review of Single-Line and Multiline Optical Character
Readers Used in Mail Processing (GAO/GGD-84-78, Aug. 7, 1984).
Postal Service: Information on the Change to Multiline Readers for
the ZIP + 4 Program (GAO/GGD-86-62BR, Mar. 28, 1986).
Postal Service: Automation Is Restraining But Not Reducing Costs
(GAO/GGD-92-58, May 12, 1992).
Postal Automation and Pricing in the 1990s (GAO/T-GGD-92-39, May 12,
1992).
Postal Service: Restructuring, Automation, and Ratemaking
(GAO/T-GGD-93-15, Mar. 25, 1993).
Postal Service: Role in a Competitive Communications Environment
(GAO/T-GGD-94-162, May 24, 1994).