In 1970, Citibank’s services rated poorly when compared with other banks’. In response, management ordered an overhaul of the services area and brought in managers with experience in manufacturing environments to make the services more efficient. That program succeeded in ensuring new processing efficiency and management control; yet it did not go far enough in terms of personalized customer service. Rather, the program tended to treat a service as if it were a manufacturing product, thus deemphasizing its great strength, its uniqueness. So Citibank moved in a new direction, this time redesigning for those who would appreciate the strength of services: the individual market and customer areas themselves. One change led to another: with this decentralizing program came decentralized technology as well as new jobs and a new work environment. The author, who was instrumental in the second design program, describes the history of the change. Using the letter of credit operation as a case study, he also describes how the change was made as well as the services management philosophy that guided it.
It is June 1975. A request to issue a letter of credit arrives in the mailroom at Citibank’s operations headquarters on Wall Street in New York. The request is from a Citibank correspondent bank in Lyons, France on behalf of a local company that needs to finance a shipment of machine tools purchased from a company in Michigan.
Sorting clerks in the mailroom check the contents of the envelope and send the item on to the bank’s letter of credit department on the building’s 24th floor. The item lands on the desk of a preprocessing clerk who determines the source of the item, whether it is from a Citibank branch, a correspondent bank, or a government agency. The clerk shunts the item to the correspondent bank section. There, another clerk has to determine whether the item requests that a letter of credit be issued, amended, or paid, or whether it is a customer inquiry. The clerk sends the Lyons item to the issuance unit.
Three days later, the credit for the Lyons company’s shipment is issued. At least 14 people—a typist, a login clerk, a preparer, a signature control clerk, two checkers, the department manager, a central liabilities clerk at the bank’s uptown corporate headquarters, a marketing officer who approves the credit (also from uptown), the accounting department, the files unit, and the customer service clerk—have acted on it.
The original source document from the Lyons correspondent has been read, reread, checked, and rechecked. It has been crumpled, clipped, stapled, unstapled, rubber-banded in bundles of cards and tickets, stuffed into envelopes, copied, copied from, annotated, and preserved in a cardboard file folder.
The processing of the credit issuance has generated a stack of papers maintained in some six files. An offering ticket has gone to central liabilities and to the marketing officer, with a copy for the file; a five-copy fanfold has been typed, split, and its folds dispersed to a variety of destinations; accounting tapes, proof tapes, and MIS tapes have been punched, rolled, and delivered around the bank; special instructions have been duly noted and recorded in duplicate, one set concerning the Lyons customer, and the other the beneficiary out in Michigan.
The Diebold files are bulging, and the department requisitions additional forms and paper clips. Yet a week after the credit is issued, the customer requests that it be amended, starting the same process all over again.
When the time comes for payment, the appropriate documents are presented at the letter of credit department, correspondent bank section, payment unit. Tickler files are checked, credit documents are pulled, examiners examine, checkers check, typists type (a nine-part fanfold this time), verifiers verify the accounting entries, a proof clerk stashes the tickets into a pigeonhole on a rolltop desk for action at the end of the day, and a check is made out and mailed—five days after the payment documents were presented.
It is now October 1976, a year and a half later; the same company requests a similar credit through the same Lyons correspondent. In the Citibank mailroom, a clerk routes the item to: Letter of Credit, 21st floor, European Division. At the division’s preprocessing desk, a clerk notes the item’s source and puts it in a slot marked “France” on an automatic delivery cart, a mail “robot.” As the robot moves along the aisle, the item is plucked from its slot by an individual sitting in a neat, cockpitlike work station, where she is typing onto a CRT terminal keyboard. As she punches the last key, a printer terminal nearby prints out a formal letter of credit. It is mailed that day. She immediately puts the customer’s original request and instructions on microfiche and stores it in a small file case on the flat top of the work station at her right hand.
Where it once took days, 30-odd separate processing steps, 14 people, and a variety of forms, tickets, and file folders to process a single letter of credit, it now requires one individual less than a day to receive, issue, and mail out a letter of credit—all via a terminal that is fully online to a minicomputer-based system.
Part of the change is dramatically evident. It was a technological leap—from clerks, papers, forms, paper clips, rubber bands, and file folders to workstation professionals, CRT terminals hooked to a minicomputer, and micrographics for record keeping. But part of the change is rooted in the department’s basic function of providing financial services to international corporate customers and is less obvious. This is a change in the very approach to services management, a change in the organizational structure to reflect that approach, a change in the work processes, and a change in the jobs themselves. Some history may put the change in perspective.
First Management Program: Production Model, 1970 to 1975
In 1970, an Opinion Research Corporation survey of major banks rated Citibank quite low in terms of service to the customer. This dismal finding was made intolerable by the fact that, for the 11-year period of 1960 through 1970, the bank’s internal operations costs had risen at some 15% a year.
When volume soared during the economic boom of the 1950s and 1960s, management hired more clerical staff in an attempt to maintain the one-to-one relationship with the customer that had characterized a quieter era of banking. At the same time, the large, centralized computers of the early 1960s increased the bank’s transaction capacity, which in turn increased the need for more clerical staff to handle the paperwork.
At our peak of inefficiency, Citibank operations starred a cast of thousands who shuffled the paper that was the basis of our transaction processing “system.” These people were handling only a part of the whole, a single task over and over again. The repetitiveness of their work, and the fact that they had no sense of its completion, militated against consistently error-free performance. In fact, out of monotony or lack of attention, our clerks tended to create errors in the processing flow. At one point during 1970, a backlog of some 36,000 customer inquiry items had accumulated. Proof problems and discrepancies in our records were not uncommon.
The institution’s top management decided to effect some fundamental changes in the back office, the seat of the bank’s financial transaction processing. Management made the direction for that change clear by bringing in a new management team recruited from manufacturing industries and from other distinctly nonbanking environments. The new managers, schooled in quantitative analysis and production methods, found the situation very nearly out of control.
The Back Office as a Factory
From the industrial vantage point of the new managers, the dominant underlying cause of Citibank’s problem was organizational. At the time, the operations organization was functional. All of our transactions—no matter what their source or originating customer—came through the same single pipeline to be processed together. It was a matter of first come, first served.
In this organization, a manager was responsible only for a given function which, typically, cut across the whole pipeline. For example, someone managed the function of data preparation; someone else would be in charge of data input; a third person managed the function of item sorting. It was classic horizontal management, a structure in which each manager’s ability to perform his job was dependent on the performance of the function that preceded his in the sequence. And if the pipeline broke at any point along the flow, there was the risk that everything would stop.
In response to this problem, the management team instituted a production management tool, the assembly line. It began by breaking out the items by source—from the Federal Reserve, from our correspondents, from our branches. Independent processing organizations were established for each of these major distinctions.
Within these, refinements were made according to transaction type, that is, by product line. A separate “channel” was established for each product, and all the functions needed to process that product were included in the channel. We had, then, complete vertical management: a person who managed a channel now had control over the total transaction from the time it entered the bank until the customer was advised of its completion.
Control was the focal point of this organization change—the kind of control ensured by the production management disciplines applied in manufacturing concerns. Indeed, the restructuring itself was based on a fundamental conceptual difference: that the back office was really a factory, not a clerical operation. From this starting point, the techniques of production management were adapted to the special realities of our products and to the kind of work involved in producing them.
To implement a services version of the quality control systems that operate in manufacturing, we designed indicators that measured both the timeliness of the processing and the extent to which a channel’s output was error free. To realize financial control, we established analysis and forecasting systems and instituted top-down budgeting and target setting. The target was to hold processing costs flat for two years and then to allow an annual increase of 6% if—and only if—volume warranted.
Within this framework, we addressed the key issue in the processing of financial transactions—and in related fields as well: the labor-intensity of the business itself. In the past, we had simply hired people to push paper. During the 1960s, hardware and software represented only some 10% of our total processing costs; all other expenses made up 90%; labor alone accounted for a full 70% of the total. By its very nature, this ratio constituted an issue of productivity.
Expressed simply, the issue was to flip the ratio entirely, to stand it on its head until the labor and other operations costs constituted only 30%, while hardware and software made up the rest. The management program of controls, forecasting, and accountability that we borrowed from production management was part of the response. The essential addition was technology.
In a sense, the recognition that a new technology was needed was also borrowed from the manufacturing industries. The production management disciplines that ensure the smooth running of mass production plants had been born alongside the introduction of machine tools that could produce more at less cost. In the Industrial Revolution, machinery and the development of modern management techniques were two strands of the same thread. In banking, however, the equivalent of the machine tool is the computer.
When computers were first adapted to commercial purposes, they were embraced in much the same way as is the latest rage in fashion—and with a similar lack of concern. They were simply something every big business had to have. We all knew they were going to change fundamentally the way things were done, and although we did not know precisely what their impact would be, everyone wanted to be in on the change.
Early on, the bank’s back office fell under the spell of the dazzling potential of technology, attacking labor-intensive operations with new automated systems. We bought machines to stuff, sort, seal, address, send, and file. Bit by bit, function by function, we cut back on costly and error-prone manual routines. We eventually worked down the cost ratio of hardware/software to labor/other to 23:77, a significant advance over 10:90.
The Program’s Results
A dramatic indicator of our results was the measurable reduction in total operating expenses. From 1970 to 1975, as total staffing levels were managed down (through attrition and placement) in the face of inflationary trends, our processing expenses remained flat. Had costs for the period continued on the upward spiral of the 1960s, the drain on Citicorp’s bottom-line net profit would have been substantial. Therefore, our total program in operations—the new management approach and the introduction of new technologies—had made a major difference to the institution.
In 1975, Citibank was again involved in a survey of how our customers perceived our services. We had improved some in the five years since 1970: in terms of accuracy, where we had been last in 1970, in 1975 we ranked at about the middle of the lower half. In speed and responsiveness to customer needs, we were rated just above the mean and on the edge of the top third of the 30 banks surveyed.
While these results in part were grounds for feeling that the internal improvements we had made had positively affected our customer base, they also pointed out a new direction for management. The 1975 survey showed not only that the efficiency of our production process and the very high quality of our product had had an impact on our customers, but also that the impact had not reached far or deep enough. Why?
In taking our cue from the production management disciplines of manufacturing enterprises—a necessary first step, to be sure—we had tended to blur the difference in what a customer expects from a manufactured product as distinct from a service delivered. In gaining the control needed to achieve production efficiency, we had perforce homogenized the services that we processed. By imposing a kind of product uniformity on our processing, we had sacrificed what is the very essence of a financial transaction service: its uniqueness.
There was a second message in the survey results. We were being told that another premise of service had been devalued: personal ministration to the customer. Obviously, we could not go back to the “preboom” era of lower volume and a slower pace. But, just as obviously, it was necessary to find a way to build in the kind of personal service that our customers were still finding in other banks—particularly, as the survey made clear, in smaller ones.
Second Program: Services Management, 1975 to Today
Programs carried out between 1970 and 1975 had been aimed at achieving production efficiency. By 1975, we could be said to be running a very good processing operation. But what we wanted was to be running a financial services business, one that was geared to providing personal services to diverse customers in a variety of marketplaces.
A customer buying a product off a shelf chooses one of many identical products. A customer buying a service is asking for a specific activity that will meet a specific need. Clearly, we would have to go beyond production management solutions to a post-industrial model of services management. This added dimension required stretching our vision to look outside our processing shop to where the customer sat, and then to look back in from his viewpoint.
We devised a new program that has three basic components: first, a reorganization along market-segment lines; second, the establishment of the appropriate technological base to serve the production needs of this market-driven structure; third, the redesign of jobs and processes.
Decentralization into Market Segments
Essentially the new program took a decentralized approach to organization, to the computing function, and to the managerial and processing tasks. As we perceived it, only by decentralizing could we custom-tailor service. For the customer, after all, the window into his bank should be a single entry point, angled toward his marketplace. The actual processing of his transaction should be invisible to him, but fast and accurate. The service he is provided should come from a responsive human being who, through familiarity with the customer and his business, become more than just a disembodied voice on a telephone or a number on an inquiry form.
The exhibits depict the evolution of the organization into the market-driven decentralized structure. In 1974, our corporate services were carried out in channels that suited the management program begun in 1970 (see Exhibit I). All processing functions for a product were vertically integrated under one accountable manager. The marketing, consumer check processing, and data processing functions, however, were handled by separate divisions. In that year, the initial cut at the new organization realignment was undertaken.
Exhibit I Operations 1974: The First Reorganization
First, consumer processing—the high-volume check processing of retail branch items—was transferred to a newly created institutional group, the Consumer Services Group. Within the corporate customer base, each channel was given its own marketing resources and its own specialists in systems development and applications programming. The only element still centralized after this restructuring was the remaining data-center organization—primarily the accounting and information systems.
In 1975, we carried the concept further by reorganizing into services management divisions corresponding to the market segments represented by our corporate banking groups (see Exhibit II). These banking groups, homes of the account officers of the institution, included:
Exhibit II Services Management 1975
- A National Banking Group, which serves our diverse and extensive domestic corporate and correspondent bank market.
- A World Corporation Group, serving the major multinational corporations.
- An International Banking Group, whose market consists of our overseas branch network, overseas correspondents, and foreign corporate customers.
The logic of the new operational divisions formed was, primarily, the market segment and, secondarily, the products provided to that segment. As a result of this change, the number of fully accountable units dedicated to the needs of a segmented customer set more than doubled.
In 1976, we further refined this divisionalization. First, the services division corresponding to the National Banking Group was split three ways to reflect major market segments within that group: commercial/industrial clients, financial institutions, and securities services. Both the World Corporation Group and the International Banking Group were organized around geographic areas or market segments, and our operations divisions oriented toward these groups were broken down into a pattern that mirrored these regions (see Exhibit III).
Exhibit III Services Management 1976
We continue to refine the organization today; in a market segment, processing is again segmented down to serve discrete customer sets, or, where volume warrants, even a single large customer.
Distributed Processing Network
The technological base to parallel this new organization also had to be decentralized. In 1975, we made the important decision to turn from “big-box” maxicomputers to decentralized minicomputers and microcomputers in a distributed network. That was a key decision with significant implications. To understand fully the compelling reasons for the decision, it is helpful to review briefly some of the history of computers in banking—a history, for the most part, of management’s failure to understand computers. The fact is that for the first two or more decades that computers have been around in banks they have not really been solving business problems; rather, they have been set against systems tasks or production issues.
In the 1950s, the first computers handled routine work such as sorting, check printing, and statement production where volume was an issue. They replaced electric calculators and electro-mechanical bookkeeping machines to perform payroll computations, securities accounting, installment loan updating, trust accounting, and so forth. They simply did faster and more accurately what had previously been done by hand, and they were “managed” with the same unit-record mentality that had directed those manual operations in the past.
In the 1960s, the picture changed somewhat, but management continued to miss the full benefits of computers—mostly by still failing to understand what computers were all about and therefore abdicating responsibility for their use. This was the era of EDP departments and the introduction into finance and other activities of the systems analyst and the technologist. These specialists, with their continuous-flow mentality and sheer love of the machine, tended to ignore the requirements of the business (which they had not really been asked to understand in the first place).
It was also the era of the large-scale computer. Multiprogramming, multiprocessing computer systems, with their capability for handling a number of unrelated items simultaneously, provided economies of scale and were heralded as cost-effective. We were buying equipment, not solving business problems. And we were buying in a marketplace dominated by the vendors who, for their part, were developing technology for its own sake.
In a very real sense, the large, centralized data-center approach drained our ability to provide the uniquely tailored services our customers demanded. It is precisely that ability that distributed processing and the use of small computers makes possible. The distributed network answers our business need.
Where there had once been one centralized data center for the institution as a whole, there are now many minicomputer centers, each meeting the specific requirements of the market segment to which it is dedicated. This decentralization enables us to be considerably more flexible in the services we provide to customers. It promises to be far less costly over the long term. And it ensures full accountability by management to the customers served.
Redesigning Jobs & Processes
The decentralization in organization and processing mandated a fundamental redesign of the work environment and the job processes. The assembly-line channel, in which each worker was responsible for just a single function or piece of the total flow, was obsolete in an organization geared to markets, customers, and custom-tailored services. The idea was to make the individual employee responsible for the complete processing and customer service for a small group of customers. This would entail making our work force more professional, redefining and redesigning jobs, changing and raising the standards of productivity, and establishing the transaction processing employee as a contact point for the customer.
For us, this concept providing small-bank services with big-bank efficiency, know-how, and resources was exciting and powerful.
Letter of Credit: A Case Study
How we advanced the concept is best understood by looking in detail at one of our processing operations, the letter of credit activity. The case study demonstrates the elements of our approach to the services management model: reorganization along market lines, introduction of minicomputer-based distributed processing, and redesign of jobs and of the work place itself to build a professional environment.
A letter of credit is a document indicating a guarantee by the bank that it will pay an obligation of its customer to a third party, usually when certain stated conditions have been met. The letter of credit gives a buyer of goods the prestige and financial backing of the issuing bank. And the bank’s acceptance of drafts drawn under the letter of credit satisfies the third party and his bank in the handling of the transaction. This is an instrument widely used in the financing of shipments of goods in international trade. Because of the variety of terms and conditions possible, it can be a highly complex financial instrument.
In a letter of credit operation, the basic steps are the issuance of a letter, the amendment of the letter, and payment. The last step involves the bank’s receipt and examination of the third party’s shipping documents after the goods have been delivered for shipment to ensure that the terms of the letter are met.
The process as a whole involves preparation of the letter of credit itself, checking of credit and terms, examination of documents, and accounting activities. In addition, there is a customer service function: answering written and phone inquiries, advising, and problem handling.
Since the late 1960s, when we began our production management approach, the bank management had been considering how to automate the process. All attempts had been nipped in the bud as being totally unsatisfactory and unworkable. One set of specifications, for instance, took up five five-inch loose-leaf binders and was so complicated that it would have required an engineer rather than a clerk to work through the system. The problem underlying all the failures was that the processing operation itself had grown cumbersome and overly complex, and it would have profited us little to automate chaos.
What follows is a description of what we did, using the three basic components of the new program.
Decentralizing Along Market Lines
In 1975, as part of the overall reorganization of all services, the bulk of letter of credit operations was given over to the division serving the International Banking Group (IBG), whose customer base provides the most letter of credit business. While the IBG Services Management Division’s letter of credit department performed work only for the IBG’s customer base, it did so on a functionalized, assembly line basis, with separate units for payment examination, payment processing, files, issuing/amending, customer service, and accounting. All letters of credit, no matter where they came from or what type of customer they served, came into the shop and passed from functional unit to functional unit.
Our first step was to realign the channel around the kinds of customers that provide our letter of credit business: governments, correspondent banks, and our branches. This moved the letter of credit operation away from the strictly production-geared axis to a more market-driven orientation. In a sense, this change mirrored the larger change being carried out across the operating base itself—the shift from a product organization to a customer-segment organization. Nevertheless, within each newly aligned customer cut, the functional, assembly-line work flow persisted.
Later, the reorganization would be made wholly market-oriented, and, within marketplaces, customer-oriented. First, however, it was necessary to understand the process sufficiently so that we could gear our automation effort to an effective customer design.
Adopting the Processing Network
To build the technological base that would support decentralization, we needed to really understand what happens when a letter of credit is processed. Such understanding was essential if the process was to be shortened—its functions integrated—so that a single employee manning a single minicomputer could perform it.
Processing manuals did exist. In three levels of detail, they described exactly how to perform every task of every step along the way. But the management team suspected, indeed assumed, either that the employees had devised their own incidental methodologies, or that some tasks had become so automatic one could no longer identify them as steps the automated system should include. The managers assumed, in short, that things in the manual were not in fact being done, and that things being done were not in fact in the manual. They decided to find out exactly what was going on. How?
For one thing, the managers just walked around the department scrutinizing the operation, watching people at work, checking what they saw against the manuals. Then, they called employees into a conference room and asked them precisely how they did their jobs, and they asked the supervisors how they thought the employees did their jobs. Finally, every morning at 7:00, the letter of credit management team convened to go over what was learned the day before and to match it against the manual.
The end product of this analysis was a book describing the letter of credit operation in full. In this book, the management team charted and described for every step of the processing (1) the inputs to the step, (2) the files required, and (3) the processing tasks.
This guide was then held up for scrutiny and was declared suspect—not as an analysis of what the workers and supervisors were saying, but rather as a mirror of what was actually happening.
Carlos Palomares, vice president and director of letter of credit operations at the time, expressed it this way: “If we were talking to the supervisor, we weren’t 100% sure that what he said was really how the clerk was doing the processing. If we were talking to the clerk, we weren’t 100% sure that he was including every possible form. So we decided that the only way to get down to bottom-level detail was under controlled conditions in a laboratory-like environment.”
The management team created the “White Room,” which was a separate walled-off area in the letter of credit department, and put in it all the equipment needed to process a letter of credit: typewriter, adding machine, rubber stamps, and forms. The managers selected Betty Matos, a seven-year veteran of various functions of the letter of credit department, and assigned her to the White Room.
The analysis of the process began with the simplest kinds of letter of credit transaction. As Palomares said:
“Different kinds of letter of credit have different levels of complexity. We wanted to be able to dissect the simplest letter of credit and do the processing on that one first so that we understood the most basic processing flow completely before going on to analyze the more complex types.
“So we would find the simplest types and bring them into the White Room and have Betty try to process them, given the steps that had been identified in the analysis that we had done, that is, the steps, the file requirements, the forms requirements, the rubber stamp requirements—everything we had learned in our 7:00 a.m. analyses.”
In effect, by setting up the White Room, the managers were testing the veracity of their own analysis. They did so, quite simply, by standing next to a seated Betty Matos as she processed a letter of credit step by step.
Several important lessons were learned during this trial in the White Room. For example, some clerks had been keeping redundant files, while others were using forms never mentioned during the initial analysis. One clerk was even using a rubber stamp that had been discontinued in 1970.
The White Room was a testing ground for more than the difference between what was written and what was real. It was an arena to experiment on what should be. As they stood there, watching Betty Matos (and later, an expanding force of workers), the managers would play around with ideas, asking each other and the employees, “What if we combine these two forms, or these five forms? What if we eliminate this step altogether?”
In the three months of the White Room’s existence, without introducing any new automation, the managers reduced the letter of credit department from 142 to 100 people, and the number of steps required to carry out the total processing by half. Yet the figures only reflected the more significant achievement of a clean, efficient operation that was ready to be automated.
Through the downright drudgery of questioning every task, step, and flow, the need for every piece of paper and rubber stamp, management achieved a wholly streamlined operation in the White Room. What was left was bare-bones letter of credit processing; the jargon for it around the department was “plain vanilla.” This was the foundation on which change could be built. For the managers, gaining this kind of detailed understanding and control was like learning to walk. It made it possible to run, and so to automate the operation meaningfully.
The thrust of the design process was to meet the market-oriented, service-driven concept that we had defined as our management goal. We retained a software company that had been brought in by one of the hardware vendors, who had earlier tried, unsuccessfully, to automate the letter of credit process. We chose this company to design the new automated system, LOCAS (Letter of Credit Automated System), precisely because it had detailed knowledge of the product, though its experience with our chosen equipment was minimal. The management team set the following five basic guidelines for the design:
First, the system would be modular. That is, the team wanted to automate one step at a time, so it could test and actually use each module as it was designed. Also, the team wanted to follow the experience of the White Room in going from the simple to the complex. Plain vanilla was to be automated first; and even the plain vanilla—the simplest transaction—was to be automated piece by piece, first issuance, then amendment, then payment. In this way, the more complex steps of automation could be looked upon as enhancements to a basic, simplified process.
Second, the system must be totally on-line with real-time updating of the letter of credit files. Every datum was to be validated at the time it entered the system.
Third, the system should enrich each data element. That is, the managers wanted a capability whereby whenever an employee punched a customer number into the system, he would retrieve all other data about the customer as well.
Fourth, the system’s forms were to be designed according to the needs of the computer file; all special-purpose forms would be eliminated. The system itself would supply the appropriate standardized paragraphs for a particular transaction.
Finally, and perhaps most important, the system should fit the managers’ concept of the single work station. A single individual would process the entire letter of credit transaction using the automated system. The system would, therefore, need to integrate functions in such a way that this was possible. In addition, the work station would contain everything the individual needed to process the transaction.
Given the hookup to the minicomputer, those needs were now simple—a CRT, a microfiche viewer, a box for microfiche files, slots to hold the (now) few pieces of paper needed, and a telephone to handle customer inquiries.
Designing the Work Stations
An entirely new area on a new floor—the 21st—was designated to be the new letter of credit department. We called in a space-planning firm and instructed it to consult with the employees about the design of the work stations and the environment. The designers queried the employees about their span of reach and asked them to test out various kinds of chairs as well as choose colors in order of preference. We selected open-space planning, with modular furniture and movable partitions, to ensure the so-called “shmooze factor,” the ability of workers to socialize when working together.1 The partitions would guarantee each employee a sense of privacy as well as provide an area where they could concentrate.
While the system and the new floor were being designed, the reorganization was proceeding apace. From the basically functional structure of June 1975, the department moved to an interim organization, in which the three basic customer-oriented sections were maintained. Within each, however, market-oriented “clusters” replaced the functional assembly lines. Thus the correspondent bank section was organized into four divisional units, each serving its own specific regional market.
This reorganization meant that jobs were changing. People who had once been checkers or file clerks were now to be full letter of credit processors for customers in one of the divisions. Two kinds of training were required: the first would be in new, expanded kinds of letter of credit processing, the second would be in how to interface with the computer.
The method for teaching the new processing was cross-training. The first phase was vertical: within a product area, people trained one another in specific functions.
The individual who knew government issuance worked with the individual who knew government amendment. They both did both jobs until each knew the other’s. Only after the employees completed the cross-training did they begin the second phase of the program: training across products. This horizontal cross-training followed the same pattern: an individual who knew the branch area would work with one who knew the correspondent area until both knew both.
The cross-training effort went on during regular working hours, while employees performed their usual day-to-day operations. This was not the case with the training in computer interface skills. To teach employees to read the messages on the various screens, to fill in fields, even to work the keyboard required a classroom situation. These sessions were held after working hours.
From the very beginning management had been careful to communicate to the work force exactly what was at stake. The employees were told that their environment, the tools of their trade, and their jobs would change. Management tried to represent the change as an opportunity for growth and advancement for those who applied themselves—which, happily for all, it turned out to be.
Fifteen months after the original plan was presented, the system was ready, the newly decorated floor was ready, and the cross-trained employees were ready. The final structure for the letter of credit department was established.
Now reporting to a director are four operations heads, one each for divisions 1, 2, 3, and 4. Each division consists of a cluster of work stations. The individual at the work station processes completely the letters of credit for customers in “his” or “her” countries. Managers’ open offices are set around the periphery of the work-station area. The office noise consists mostly of light tapping on a CRT keyboard, talk, and the occasional understated ring of a new phone system. The back office has all the amenities of the front office.
That is very much at the heart of the matter. Financial transaction services are of growing importance to corporate customers. The day may not be far distant in banking institutions when services are as important as credit. The individual who used to be a clerk with a green eyeshade is now, at Citibank, a first-line service professional. His work station, with all the tools of his unique trade at his fingertips, is a microcosm of the institution’s raison d’être, and his job equates to the institution’s mission. The one-time clerks were prepared for this mission by being treated as professionals in a carefully managed program of training and practice. On one Friday night everyone went home reminding one another to go to the new floor on Monday morning. Over the weekend, the new furniture was set up on the new floor, and workers’ nameplates were placed at the work stations. That Monday morning the managers arrived early, so that when the employees arrived, they could “escort” them to their work stations.
This shift in job content and environment has had enormous impact on our employees. They are no longer cogs in an assembly-line wheel; they are providers of services to real customers. Gone are the stultifying effects on human energy and motivation often caused by the specialization of labor. People have full responsibility for an identifiable service that they personally deliver to individual, known customers.
The work-station professionals have responded visibly. They take a proprietary interest in both “their customers and the knowledge they have about those customers, the transactions, and the technology they ‘own.’” One man placed the flags of all the countries he serves atop his CRT terminal. People tend to feel displaced and to get edgy if someone else needs to use “their” equipment for a minute. The employees are talking to customers again, as in the old days. They are answering questions and handling problems about transactions they themselves have processed, getting instant feedback on their own performance from the people they perform it for.
In 1969, following high school and a stint in military service, Jim Mayer, 30, came to Citibank. Since then he has worked in the letter of credit department. When automation came, Mayer was glad. “I could see the automation could help make the job easier. Maybe automation puts some people out of work, but it only puts them out of boring work. It’s given me a better job, more responsible.”
The job is better because its most tedious portions are now handled by the machine. It is more responsible because, as Mayer said, “You feel you have complete control of the item from the time it comes in till it goes out.” Mayer feels too that the redesigned job in the new system recognizes the level of knowledge that the workers always had: “Before, you had to get a decision from the supervisor for any problem that came up. But the supervisor was not actually doing the work. So he wasn’t familiar with it. So you had to discuss the whole thing, and he would say, well, what do you think? So you’d really resolve the thing yourself, but you didn’t have the responsibility. And it took time. Now, you do have the responsibility. You know what has to be done because you’ve processed the whole item.”
Another aspect of this new responsibility is the accountability for performance. It does not frighten Mayer: “The bank is trying to put confidence in you. I feel aware of what I’m doing. I don’t feel nervous. Look, the work can be corrected before it goes out. Each screen on the terminal shows you exactly what you put in. All you have to do is look at it.”
For the customer, the work-station environment means a restoration of person-to-person service. An inquiry is no longer met by the response: “I only know issuance; I cannot help you amend.”
The work-station processing environment has resulted in cost savings, productivity gains, and improved service quality. It has also begun to create a new kind of business manager in the back office. The span of control of this new breed of manager is narrow—some three or four work stations. But the depth of his or her customer involvement is much increased, a development that clearly affects the bank’s marketing posture. The flexibility of the new environment and its customer-driven focus give the bank the ability to enhance products or tailor them as needed to meet or anticipate customer requirements.
The Bottom Line
It is widely acknowledged that early efforts at automation, in both the manufacturing and the service industries, often paid inadequate attention to the human factor. The result was a serious morale problem for many companies, and a worker alienation that persists in many places today as skepticism or fear about management’s automation plans. At Citibank, our first efforts at automating various areas of operations followed this pattern, but we learned from these experiences. By early 1973, in fact, human resources had become a key staff function of the operations area, and “the human factor” was instituted as one of the five performance indicators by which managers were measured.
In undertaking to automate the letter of credit operation in 1975, how people would respond was of paramount importance and figured in the planning from the earliest stages. The management team that implemented the letter of credit automation also laid out precise programs for the following:
- Communicating the coming changes to the workers.
- Carrying out the cross-training.
- Effecting the grade changes that would correspond to the enriched jobs and greater responsibility.
- Designing career paths for the new letter of credit work-station professionals.
- Transferring those personnel who would not be asked to move forward with the letter of credit automation and job enrichment or who themselves asked to be transferred.
In fact, those workers who were transferred out of the operation, either at their request or at management’s decision, were retrained and reassigned to other jobs within the corporation. Nobody lost his job.
The relocation of all the letter of credit personnel into equal or better jobs took just under one year. Some of these people were placed in assignments that they find challenging and rewarding; others no doubt feel that their new jobs are inadequate—but all have secure jobs and have received what support and training they needed at corporate expense.
Those workers who did embark on the cross-training program to become work-station professionals have in fact moved from single-function clerical jobs to skilled, professional assignments with a corresponding upgrading of compensation. People who had been hired as clerks to perform a single task for which they received weekly wages are today regarded by this corporation as professionals in whom the corporation has happily made an important investment.
Many of these former clerks are now officers of the bank; all of the jobs in the letter of credit department today have the potential to lead to officer status, and formalized career paths exist for all personnel in the area. The salary range of these professionals is at least double what the salary wage was when they were clerks, and the potential for growth is, of course, very great.
Our success with the work station has not, however, been unqualified. Through the regular feedback sessions that Citibank arranges for employees, and through supplemental interviews specifically with the letter of credit work force, we have learned that the single-product work station may have swung the pendulum a little too far back. While the workers like the responsibility and the content of their “beginning-to-end” jobs, a certain feeling of isolation is also present. The workers still help one another—in a kind of ongoing training—and they are able to socialize freely and well. What is lacking, however, is a real sense of teamwork, as well as the support and greater confidence that teamwork can instill.
We hope our concept of the next step, the multi-product work center—in a sense, the ultimate customer orientation for processing services—will answer this problem. The multiproduct work center will be geared to a set of customers, or even, if volume warrants, to a single customer, but it will handle all the financial transaction services those customers use or require.
The actual configuration of the work center is still being planned, but an overall profile can be described. There exists today, for example, a services management center (SMC) for Swiss banks. It handles all the transactions those important correspondent relationships require. The people who staff the center are becoming expert in the “style” of banking these customers practice; they are getting to know the personnel, procedures, and processes of the Swiss banks so they can respond to and anticipate customer needs.
Another services management center has been established for customers in the Middle East/North Africa region. We call it the “MENA SMC.” It has been extremely innovative—and highly successful—in designing special procedures and new products to serve the unique needs of customers in this region of the world.
Both the Swiss SMC and the MENA SMC serve as stylistic prototypes for the multiproduct work center to come. What is required now is the technological integration of functional capabilities, the development and implementation of appropriate systems to handle and deliver the full profile of transaction services for the various customer segments, and the development of solid training programs to create the multiproduct work center professional. None of these elements is that far from being realized.
In sum, then, we have made and intend to make more significant changes. Yet more important than the scale of the changes is their direction. We are using technology to create an environment in which services are tailored to customer needs and efficiently delivered by a professional work force—that is the real message of our actions.
Citibank’s letter of credit operation today is, in our opinion, a model of financial transaction services processing. In making our back office a mirror image of our front office, we are delivering the personalized service of the “old days” via the sophisticated technology and management thinking of today.