Click for next page ( 36


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 35
4 STRUCTURING THE ORGANIZATION TECHNOLOGY ORGANIZATION AND MANAGEMENT ISSUES Of equal importance to formulation of the Social Security Adm~nistration's (SSA's) strategic directions is implementation of its strategies in terms of the technical organization and management environment. This chapter addresses how overall organizational structure relates to goals and functions and affects the agengy's technology organization and management, and it discusses the specifics of management systems and issues that the committee addressed during both phases of its review. RELATING STRUCTURE TO GOALS AND FUNCTIONS Commissioner King has set forth the SSA's primary goals, which will be used to guide the revision of the Agency Strategic Plan that is under way. These new goals were enunciated (Social Security Administration, 1989) as follows: . . to serve the public with compassion, courtesy, consideration, efficiency, and accuracy; to protect and maintain the American people's investment in the Social Security trust funds and to instill public confidence in Social Security programs; and to create an environment that ensures a highly skilled and motivated work force dedicated to meeting the challenges of the agency's public service mission. 35

OCR for page 35

OCR for page 35
37 (DCU) are organized under Operations, while telephones, the 800 number, and office automation (OA) are organized under Management. Stemming from this choice of overall structure, several symptoms of inadequate organization in the technology management area are evident. As noted in the committee's Phase ~ report (National Research Council, 1990), procurement of telecommun~cations--the largest part of the Information Technology Systems (ITS) budget--is fragmented, with each splinter specifying and procuring its own circuits, either from the General Services Adm~nistration's (GSA) schedules or by separate contract. The committee again suggests that unification of telecommunications management into a single organizational unit would be an effective way of containing this large and growing expense by fostering consolidation of requirements. Even in the committee's initial review of the SSA's plans for OA, it was apparent that the management sector of the SSA needed access to "programmatic" information from the operations sector, and this access was to be provided by the later phases of the OA plan. Further review has shown that the operations sector equally needs access to OA functions, but these functions are being accomplished in an awkward technical manner using mainframe computer applications. This need was not foreseen in the original planning of automation support for the OASI programmatic functions performed by the Operations sector. The effective management of change is handicapped by the divided technical organization. The two technical organizations primarily serve the needs of their respective organizations and share almost no common technical view of SSA's future information systems. The system development organization under Operations holds on to a tradition of centralized mainframes and old computer languages and is reluctant to adopt newer-generation technology. While risk aversion is desirable, excessive conservatism in defense of current operations represents a barrier to beneficial innovation and change. In contrast, the technical organization under Management that zealously favors personal computers (PCs) and distributed processing has seemed to the committee to be overly adventurous in the later phases of the OA plan. ~- 1 ~ The committee concludes that the division of SSA's technical resources between the deputy commissioner for operations and the deputy commissioner for management reduces

OCR for page 35
38 overall effectiveness and cohesion in the delivery of technical systems and services.] Each field office, teleservice center, and program service center at SSA is a geographically decentralized but integral unit of a single enterprise. It seems clear that integrating the technical organizations would be more effective than adhering to the present organizational structure and would offer an opportunity to blend the good ideas from each constituent point of view. In restructuring, care must be exercised to achieve balance rather than obliterating one point of view in favor of the other. As part of its review of SSA's technical organization and management, the committee examined five large commercial service organizations and found that each of . ~ ~ ~ ~ them nau a single ~no~vloual at the executive or senior vice president level, reporting to the chief executive officer, with responsibility for technology, telecommunications, and information resources management. For the five large commercial service organizations, which chose not to be identified, the functions reporting to this individual are shown in Table I. TABLE I. Functions Assigned to a Single Executive in Five Companies Examined Function A B C D E 1. Data-processing operations x x x x x 2. Facilities x x x 3. Systems development/support x x x x** 4. Telecommunications x x x x x 5. Human resources x x x 6. Advanced development x Company* * Company A is a European-based bank subsidiary, company C is a major U.S. financial services firm, and companies B. D, and E are major U.S. insurance firms. ** This organization divided its systems development and support function into five groups, each focused on a specific product line. Whatever the actual title, the chief information officer in the five companies examined 1The reader should refer to the Phase ~ report (National Research Council, 1990), p. 32, section entitled 'decision Making" and subsection entitled 'divided Resources," for further background and information on this conclusion.

OCR for page 35
39 reported directly to top management and had responsibility for was at the same ~ - ~ processing operations. Commissioner King has forwarded to HHS for approval a proposed reorganization of the SSA that would reunify the currently separate technical organizations for systems, telecommunications, and office automation function.c into ~ single ors'~niY.ntir~n 1lnr1~r ~ ~ . . . ~ . ~ telecommunications, which level (e.g., A, C, }a), or no more than one level below (e.g., B. D), data- deputy commissioner tor systems. The proposed division into offices of operations, systems design and development, systems requirements, systems planning and integration, information management, and communications (telecommunications) appears to reflect the experience of commercial organizations as indicated above. We committee commends Cornrtus~oner King 'sieve u' Sing this new arid endorses the proposed orga~afiona changes TECHNOLOGY MANAGEMENT SYSTEMS Organizational structure deals with the division of work to effectively use human specialization to accomplish the enterprise's strategic purpose. To be effective, however, an organization must also have coordination mechanisms that bind together divided work groups for the effective accomplishment of the whole enterprise's goals and strategic purpose. The SSA's management systems must act in three widely different areas: (1) management policy formulation, (2) resource allocation management, and (3) quality assurance of technical design. Issues that have important implications for management policy, resource allocation, and technical management are discussed below. Management Policy Issues While the commissioner, as the chief executive of the SSA, bears the ultimate responsibility for formulating management policy, the committee agrees with the practice of many organizations to establish a formal mechanism for formulating management policy with the next subordinate layer of executives, in a board, council, or committee. The committee did not find such a formally identified group at the SSA but understands that the commissioner's executive staff at their periodic meetings function in this area. However, the Systems Review Board (SRB), which is chaired by the agency's chief financial officer, also seems to act in setting management policy by selecting and prioritizing system projects. So that the SSA can consolidate management policy setting and focus on it at the level it requires, the committee recommends that

OCR for page 35
\ 40 the commissioner establish and chair a formal management policy board with exclusive responsibility for setting management policy, including setting serv~ce-leve' objectives, project- and budget-ranking criteria, and such other matters of management policy that may be assigned to it. As a matter of management policy setting, the following priorities are currently used at the SSA to allocate budget resources among projects: For mandatory projects ongoing telephone expenses, and ongoing automated data-processing (including data communications) expenses. For discretionary projects . programmatic systems modernization, administrative systems modernization, office automation, system replacements and upgrades, and modernization management and support. The criteria for ranking discretional projects, and their order, are: impact on current senice levels; provisions of law, regulation, or higher monitoring authorities (HMAs); commissioner's mandates; Agency Strategic Plan and supporting plans . . _ ~ _ return on investment; and SSA's organizational component priorities. Several aspects of this management policy merit comment. First, this project- ranking policy deals only with the ITS budget and does not include technical labor costs, which are assumed to be fixed. Furthermore, the SSA's two technical organizations appear to set their respective priorities for human resource allocation (a separate budget line item) on a different basis. The SSA's internal guidelines do call for including work-year costs as part of the cost-benefit analysis, but this is required for ITS procurements in excess of i

OCR for page 35
41 $100,000. The committee concluded that this policy of detaching accounting for technical labor allocation from accounting for allocations for system procurements has led to completion of system designs and software by developers well ahead of the availability of hardware. Such software can become obsolete sitting on the shelf, and the investment is then lost. This should not be a normal practice allowed by management policy or by a proper resource allocation mechanism. Second, and perhaps more important, are several of the specific policy decisions that appear to have been made in developing the ranking criteria. "Ongoing telephone expenses" are given a mandatory project classification, but these uncontrolled and uncontrollable expenses are likely to increase because of more teleservice. Therefore, this expense category will likely consume a larger part of the ITS budget and will consequently curtail discretionary procurements that in turn are needed to complement the systems development investments carried out under the labor budget line item. To mitigate this effect, certain on-going telephone expenses such as those for service to field offices, program service centers, and the SSA's headquarters complex may be retained in a mandatory category--but the 800 number, including teleservice centers expenses, may be moved to the discretionary category, requiring annual approval in competition with other discretionary project requests. Also, the top ranking factor for discretionary projects--"impact on current service levels"--fails to address the management oolicv question of what service levels should be established for the SSA. , . . ~ .. . ~ _ The "provisions of law, regulation, or higher monitoring aumorltles surely ranks above the "impact of management-set service levels," as "commissioner's mandates" might be better titled. As noted in the committee's Phase I report, "return on investment" might be better titled "discounted benefit-to-cost ratio." Also, "commissioner's mandates" and "Agency Strategic Plan and supporting plans" should not escape the discipline of financial analysis as the order of ranking factors suggests. The discretionary project ranking factors should be reviewed, with consideration given to reordering them as follows: provisions of law, regulation, or HMAs; discounted benefit-to-cost ratio; management-set service levels; commissioner's mandates; Agency Strategic Plan and supporting plans; and SSA's organizational component priorities.

OCR for page 35
42 Resource Allocation Issues Once ciari~cy is achieved on the policy issues discussed above, the resource allocation process should proceed more easily than in the past. The principal resource allocation mechanisms at the SSA are the budget process and the operations of the SRB. While the committee recognizes that the federal budgeting process limits an agency's flexibility, this should not prevent the SSA from developing management systems that provide proper information for decision-making purposes. The need to report to outside entities and the need for useful internal management information are not unique to the ~~ . .. LEA or other government agencies. Me private sector also has legal requirements to report certain information (e.g., financial accounting standards set by the Securities and Exchange Commission and the Financial Accounting Standards Board). However, industry has long established a system of internal management accounting, separate from financial accounting, as the primary source of information for orderly decision making. As noted above, for effective management, both programming labor and the hardware to support it need to be tied to a common set of schedules and cost goals. Clearly, development of the latest releases of systems software without adequate computer ., . . . , .. . .. . . .. it. .. capacity on when tO run tnem suggests that netter coordination is needed to prevent such uneconomical developments and procurements in the future. Rectifying this technical management disparity is clearly a task that remains to be done in the budgeting process. The committee recommends that the commissioner, and/or the chief financial officer, review the management aspects of the budgeting process to effectively tie together the approval systems of both the labor and Information Technology Systems budgets. Resource allocation is a critical element of technical management at the SSA. The SSA should consider establishing a resource allocation board to focus on this area. A possible approach that the committee favors is to revise the charter of the SRB and its staff to Reemphasize technical design review and concentrate on resource allocation. The SRB might even be renamed as the resource allocation board. Technical Management Issues As defined by the committee, technical management issues are those that are internal to the technology management structure of the SSA. If a new deputy commissioner for systems is appointed' these issues will become the responsibility of this executive. .

OCR for page 35
43 However, in the absence of such an individual, a number of technical management activities appear to fall between organizational gaps. This has had undesirable effects on important activities that (1) are not being done, such as telecommunications integration; (2) are being done improperly, as in the case of technical review by the SRB; or (3) are being done informally in a less than satisfactory way, such as the establishing of new links between the programmatic system and the PC network. It is too much to expect that such systems integration will be accomplished effectively given the present organizational structure, which requires the resolution of closely interacting problems only at an unusually high level in the organization. However, for the discussion that follows, it is assumed that either a deputy commissioner for systems exists or that some other effective integrating mechanism will be devised. Software Quality Assurance Because of the size and complexity of the informs~tinn earl t~l-~^mmiln;~tir~n~ ~`lclPmc .. ^ A ~ . 41~ ~ Cat Cat A 4 ~ 1 ~ _ _ ~ ~ A, ~ . . .~ ~ ~^ ^~^ ~~ ~_~__~_~$v44~ If If uy all; arm, lecnnlcal Design ana software quality assurance must be the responsibility of individuals having the requisite technical expertise. There are well-established procedures for a technical organization to assure nuali~v incllltlin~ rlesisrn r~vi~.w.~ ~nr1 ~~ . ~ . . ~ ~ :~ A ~ ~ ~ ~ ~~ ver1ncat1on and valloatlon testing. In fact, the present verification procedures used in the programmatic development organization are a good example of such a self-policing activity. The practice of reviewing one's own work leaves room for doubts about objectivity. A solution often adopted by industry is to name a director of quality assurance as a separate line organization. This is an adequate anDroach for man~fact~rin~ hilt for ~ _ ~ ~~ in_ ~^ ~^ ^~ ~ ~^ ^~~ vow ^~^ . ^. . . . esoteric software Development, software quality assurance is a haltingly developing art. A reason for the slow evolution is general scarcity of individuals with the high-level of expertise needed to assure software quality, which should equal the level of the developers. For this reason, most software-based organizations employ rigorous software engineering methods to assure that quality is built into the product.2 Another option would be to contract out to qualified software consultants for the equivalent of a quality assurance department. The advantages of this approach are the independent perspective and access to requisite expertise the disadvantages are nrincinallv in the areas of timeliness and consistency with an outside reviewer. On balance, the committee believes that a preferred approach for SSA exists and the committee recommends that ~ , ~ D ~ ~ ZRefer to the Phase I report (National Research Council, 1990), pp. 19-24, section entitled "Software Engineering Methods."

OCR for page 35
\ 44 the Deputy Commissioner for Systems have full responsibility and be accountable for technical design review and software quality assurance and for establishing the appropriate policies, practices, and procedures to uphold this responsibility. -I In this role, the deputy commissioner for systems must also have responsibility for ensuring the quality of the product as well as its development. Technical Design The committee believes that the major technical issue at SSA is not centralization versus decentralization but rather how much decentralization makes sense. SSA currently maintains a large centralized client database, developed over many years to support its programs. The agency's ability to carry out its mission depends on this database, and there are no plans or budgets proposed to change it in any significant way. The committee did not find compelling reasons to suggest that immediate steps be taken to abandon the centralized database in favor of a distributed approach. The committee believes that, because of its huge investment in a centralized system, the SSA should retain a centralized client database and continue to maintain it and back it up for the foreseeable future. Therefore, the issue of decentralization is relevant largely in the area of displays, formats, calculation, and other routine tasks that can be distributed. Notwithstanding the arguments of technical proponents for centralized and distributed systems and their opposing rationales, the SSA needs to consider the economic dimensions of this issue in terms of cost-benefit analysis. The added costs of decentralizing certain tasks may be calculated in additional lines of code that need to be developed and maintained. The added benefits of decentralization will show up in reduced traffic through the communications network and reduced loading of the costly NCC mainframes. Such benefits will result in lowering costs or avoiding costs by deferring expansion for some period of time. Another advantage ascribed to decentralized systems is that they allow greater response to unanticipated requirements, are more accommodating to user differences, and adapt easily to new opportunities. During its review, the committee found a strong divergence of opinion amens S:SA's ~ e ~ ~ . =7 _ _ ~ ~ e ~ ~ ~ ~ -D ~ D~ ~~ Art ~'''~''D ~^ ~ two technical organizations. One side favored centralization; the other favored full distribution. In light of the consolidation of these two organizations into one, it is necessary for the deputy commissioner for systems to address this divergence and blend it into a vision of the core level of decentralization most desirable for the 1990s. ~ . ~ . .

OCR for page 35
45 Configuration Management and Change Distribution Configuration management in which software changes are controlled and distributed will become more important at the SSA in the future. Currently, with a totally centralized architecture and "dumb" terminals, system changes are made only at the NCC and can be ~ ~ _ ~ ~ _ 11 ~ 1 ~ ~ ~ . ~ . ~ , ~ e ~ ~ ~~ ~~ _ manaueu and con~ro~ea DV release numbers. nut untn tne tntrnn~,ct~on at ~t :s and networks ~ , _ _ , ~ ~1 ~ ~ , ,~ ~ ~ ~ ~ ~ ^. ^~ . .. in the MA system. the need for and burden on software configuration management wall Increase. O ~ ~= At present, the limited use of PCs and their minimal interconnection with each other and with the Programmatic system (lo not require more than managing the nhv.cicn1 D r J e _ ~ ~ 1~ e ~ ~~ ~ ~ ~ ~ ~ ~ . ^~~ . alslrlDuuon or software ana changes. Ants is so because separate ottlces on separate local area networks can still operate effectively with different software releases as changes or upgrades are physically distributed and installed separately at each office. As the programmatic and management systems become increasin~lv interlinker1 in the 1 n n ^ ~ _ ~ ~ ~ ~ _ 1 1 ~ ~ 1 ~ ~ ~ 1~ aim 1melllgence migrates lo Ine terminals ana workstations, simple manual configuration management as currently practiced will become inadequate. The loss of synchronization caused by time delays in installation will not be acceptable in a more distributed and integrated systems architecture. Thus the SSA must plan for distributing _ _ ~& ~ to _ 1 ~ ~ 1 ~ ~ software changes or upgrades throughout Its system in a more synchronized and controlled manner. Electronic down-loading over the telecommunications network offers such a means. Centralized network control and device management used in concert with a sound configuration management program will be necessary and vital technical management tools at the SSA in the future. The committee believes that any incremental cost associated with extra telecommunications capacity needed to support electronic down-loading will be more than offset by the economies of scale of an integrated voice, data, and video telecommunications network. For example, T-1 capacity links (24-voice channels) are now being offered by commercial long-distance carriers at tariffs equivalent to those for 10- to 12-voice channels. Intelligent telecommunications network planning on a unified media basis would also suggest that many management and control operations, including down-loading, can be done during off-peak periods. Systems Integration Management In discussing systems integration management here, systems are addressed in the larger context of information (voice, data, and video) and telecommunications. In this way the issues of management are highlighted. Not surprisingly, the committee found examples of systems integration at the SSA at all levels that were done well and others that could have

OCR for page 35
46 been done better. In supporting its programmatic functions, the original Systems Modernization Plan, with its massive hardware increase in its earlier stages, provided a good example of systems integration management. Where it began to weaken in effectiveness was in its later period, from about 1987, when the technical work force continued at full speed despite the absence of usable capacity at the NCC. This resulted in a number of software releases that cannot be fielded because they were developed in advance of hardware availability. This reflects inadequate hardware and software integration planning. The OA projects were initially planned independently of the programmatic systems. ~ I _ ~ _ ^4 ~ ~ A ~ ~ 1_ . , , ~ . ~ The OA projects moved toward prOVlUtIl~ Ins ADA Elm an 1megralea system for management data, but this foms on supporting "Management" requirements should be expanded by adopting a broader set of goals that will serve the overall technical management effectiveness of the SSA. For training purposes, a very small aperture terminal network was apparently planned without consideration of other telecommunications requirements. The 800 number team similarly procured its telephone lines independently and without regard for potential economies from collocation with other services requiring telecommunications support. For example, the committee found that, at the program service center in Richmond, California, where major telecommunications lines terminate, the local telese~v~ce center for the 800 number was located elsewhere. Effective systems integration management has been difficult to achieve at SSA as a direct consequence of the Operations/Management technical organization split. Between these two halves of the technical organization, systems integration management has been minimal. Each component seems to have gone its own way. Some integration has been achieved by the efforts of (1) the system planning staff to integrate at the Agency Strategic Plan level; (2) the chief financial officer to integrate at the SRB level; and (3) the natural informal linkages that have sprung up to improve the exchange of information between the programmatic and management systems. Systems integration management can only be as good as the top leadership structures the organization to achieve it. The committee's principal concerns derive from the division of technical resources and programs at the deputy commissioner level. The committee recommends that the Deputy Commissioner for Systems be charged with establishing plans, programs, and policies that direct and strengthen technical systems integration management.

OCR for page 35
47 ~~ ~0 Nations Resend C-ndl. 1990. ~ ~~ ~ ~e ~ ~ ~ ~e ~-~^~ W~b1~to~ D.Cx National Academy Press Soda Seedy Ion. 1989. ~~e\~ ^ 2. Baltimore, ~d.: Social Seedy ~~strabon.

OCR for page 35