The Balanced Scorecard: Translating Strategy Into Action (18 page)
Read The Balanced Scorecard: Translating Strategy Into Action Online
Authors: Robert S. Kaplan,David P. Norton
Tags: #Non-Fiction, #Business
While just-in-time production processes and the manufacturing cycle effectiveness (MCE) ratio were developed for manufacturing operations, they are just as applicable to service companies. If anything, eliminating waste time in a service delivery process is even more important than in manufacturing companies, since consumers are increasingly intolerant of being forced to wait in line for service delivery.
Take an example from the banking industry. Many of us are familiar with the process of gaining approval for a mortgage on a house that we wish to purchase. The process starts by showing up at a local bank branch, and filling out an extensive application form that includes employment history, salary, assets and liabilities, as well as a description of the house. After completing the application, the employee thanks us for choosing her bank, and then says that we can expect to hear whether or not the mortgage application has been approved in three to four weeks.
One bank vice president, very familiar with the normal cycle time of 26 days to process such requests, asked employees to keep track of how much time was spent actually processing the application during these 26 days. The answer turned out to be about 15 minutes of work, spread across 26 days, an MCE ratio of 0.0004 (0.25 hours/[26 days × 24 hours per day]). The vice president set a target to reengineer the approval process so that it would only take 15 minutes from completion of the application to a yes/no decision. This target corresponded to an MCE of 1.0. Bank
personnel would continue to do all the value-added processing work but would eliminate all the non-value-added waiting times. At first, all the employees involved in the mortgage approval process claimed this was an impossible target. Among other tasks, credit references had to be requested and confirmed, a process that took at least a week or two. Further study revealed that credit references could be accessed on-line for almost all possible customers. Much of the analytic work and approval routines could also be automated. A reengineered mortgage-approval process, supported by enhanced information technology, was designed that yielded a yes/no decision within 15 minutes. Thus, after customers filled out the mortgage application, they were directed to a cafeteria for a cup of coffee and by the time they returned, a decision was available.
13
A 15-minute one-stop mortgage approval process turned out to be highly attractive to a broad market segment of customers.
Similar studies in other service industries yielded similar conclusions; long cycle times for customer service during which actual processing time was remarkably low. Automobile rental companies and a few hotel chains have now automated, for targeted customer segments, all aspects of checkin and check-out, enabling valued customers to skip waiting in line when initially accessing the service and upon completion of the service delivery process. Thus, companies attempting to deliver products and services on demand to targeted customers can set objectives to have MCE ratios approach 1, thereby producing dramatically shortened lead times for customer orders.
Almost all organizations today have quality initiatives and quality programs in place. Measurement is a central part of any such program, so organizations are already familiar with a variety of process quality measurements:
- Process part-per-million defect rates
- Yields (ratio of good items produced to good items entering the process)
- Waste
- Scrap
- Rework
- Returns
- Percentage of processes under statistical process control
Service organizations, especially, should identify the defects in their internal processes that could adversely affect costs, responsiveness, or customer satisfaction. They can then develop customized measures of quality shortfalls. Metro Bank, as one of its measures of service quality, developed an index called “Trailway to Trolls” (trolls are unhappy customers) to indicate the defects in its internal processes that lead to customer dissatisfaction. The index included such items as:
- Long waiting times
- Inaccurate information
- Access denied or delayed
- Request or transaction not fulfilled
- Financial loss for customer
- Customer not treated as valued
- Ineffective communication
A particularly demanding quality measure, analogous to the MCE ratio described earlier is first-pass yields. Two real stories illustrate the importance of this metric.
Several years ago, one of the authors visited a major automobile company, which we call National Motors (to protect the guilty). The plant superintendent was conducting a tour for the visitor, emphasizing the transformation of the plant to a total quality and just-in-time environment. To illustrate the success of the total quality initiative, a banner at the end of the production line declared that the plant had achieved a perfect score of 155 at the final inspection point of finished products. The superintendent then showed the already impressed visitor to the incoming materials receipt area, where the tracks formerly used for freight car deliveries of raw materials and purchased parts had been ripped out. They had been replaced by loading docks where truckload deliveries were made several times per day. On the way through the plant, however, the visitor noticed many tall racks containing what
appeared to be large quantities of inventory. He asked, naively, why was there a need to store inventory? If materials and parts deliveries were being made just-in-time and moved immediately into linked production processes that could pass intermediate goods from one process to the next without delays, where did all the inventory he saw come from? He was told, quickly and somewhat condescendingly, that he was not looking at inventory. That was the rework area! The plant had achieved its perfect quality score by inspecting items after every production process, and putting to one side any items that failed the quality test. This plant was still operating the expensive way: by inspecting quality in, not designing it in.
About 1980, the defense electronics division of National Electric was experiencing significant quality problems in its printed circuit-board production and assembly process. It sent a team of engineers to a similar Japanese company to compare the two firms’ production processes. Early in its visit to the Japanese company, National Electric’s team was asked how many of a batch of 100 printed circuit boards made it all the way through their entire production process. The National team leader responded indignantly, “They all do. These are expensive boards. We don’t lose any.” The Japanese inquirer apologized for the poor translation of his question. He meant to ask, “How many boards make it through the entire production process the first time, without any rework having to be done to them?” The National Electric engineers huddled for several minutes and then were forced to admit that they had no idea. It was not a statistic that they had collected or ever considered collecting. They were too busy attempting to minimize adverse labor and machine efficiency variances to contemplate additional production measures, especially a nonfinancial one. Nevertheless, their interest piqued by the question, they asked their Japanese hosts what their percentage was. The Japanese responded that they were currently at 96% first-pass yields. Twelve months ago, they were only at 90% but had been working hard to increase this percentage, with a goal of eventually reaching 100% first-pass yields.
When the engineers returned to their U.S. plant, they asked their plant manager and plant controller the same question. Neither knew the answer. A special study was performed and several weeks later the answer emerged: 16%! Everyone agreed they would not be in this business for long without
a significant improvement in this percentage. Within six months, TQM efforts had raised the percentage to 60%, and this increase in first-pass yields enabled the operating work force to be reduced by 25%: from 400 to 300 employees. In effect, 100 people had been employed at the plant to produce defective products, then to inspect and detect them, and finally to repair them until they were acceptable finished goods. Once the plant decided that it was not in business to make and repair defective goods, the 100 people formerly employed in this activity were no longer needed.
These stories reveal the power of using first-pass yields as a quality measure. The success of a quality program should not be measured by the quality of outgoing items after they have survived numerous inspections and rework processes. It must be measured by reductions in percentage of items, at each stage of a production process, that do not conform to customer-based specifications.
Amidst all the attention to process time and process quality measurements, one might lose sight of the cost dimension of processes. Traditional cost accounting systems measure the expenses and efficiencies of individual tasks, operations, or departments. But these systems fail to measure costs at the process level. Typically, processes like order fulfillment, purchasing, or production planning and control use resources and activities from several responsibility centers. Not until the advent of activity-based cost systems could managers obtain cost measurement of their business processes.
14
For example, an early ABC application in the late 1980s occurred with a branded manufacturer of personal care products. The study focused on manufacturing costs, but the ABC analysis revealed that one of the principal contributors to manufacturing costs and complexity was the production of small lots of new products. As new flavors and varieties of products were designed in the company’s R&D activity, small lots had to be manufactured for initial testing. This often required stopping a high-volume production run already in process to set up for the R&D production lot and then resetting up for the high-volume run. After the new product variant was launched into test markets, feedback from consumers was used to redesign the product, leading to a demand for even more small-lot runs. In the past, the cost of changeovers for the small runs of R&D production lots, and for testing the reformulations, had been treated as part of manufacturing
overhead and allocated to existing products using traditional (and arbitrary) cost allocation procedures. As part of the ABC study, all the production costs, both volume-related and batch-or lot-related (including the round-trip cost of setups when a large volume run was interrupted by an R&D test lot) plus production runs for test marketing, and for reformulating products, were assigned to a newly defined activity, launching new products. The analysis revealed that the company was spending a very high amount per product launch, far more than it had ever imagined. Previously, the company had managed total spending for the R&D group but had neither traced such spending to the outputs produced (number of new products created and launched), nor had it included the costs incurred outside of the R&D department, such as the high manufacturing costs for the small R&D lots. Once they understood the total costs associated with launching new products, the managers were more receptive to suggestions to reorganize the new product formulation and to initiate procedures to obtain a far more efficient and effective process. They also had a much better analytic understanding of the costs associated with simple line extensions that could be compared with the benefits from these extensions.
In general, ABC analysis will enable organizations to obtain process cost measurements that, along with quality and cycle time measurement, will provide three important parameters to characterize important internal-business processes. As companies use either continuous improvement (such as TQM) or discontinuous improvement (such as reengineering or business process redesign) of important internal-business processes, the three sets of measurements—on cost, quality, and time—will provide data on whether the goals of these improvement programs are being achieved.
