The subject facility is one of the largest manufacturing or processing facilities under one roof in the world. Its staff of nearly two thousand people operate some of the most highly-automated and computer-integrated equipment in the world on two million square feet (roughly forty-six acres) of floor space under one twenty-seven acre roof. Its annual output represents one-fifth of the total annual domestic cigarette output. It processes over 800,000 pounds of tobacco daily.
Virtually every step in the processing is computerized for maximum precision in process control and product quality assurance. Except for quality sampling, no human hands touch the tobacco. The degree of automation presented numerous challenges in training the floor supervisors and in the design of the managing systems. The consulting project began as construction was ending and continued through the completion of the start-up phase to the beginning of full, steady-state operation.
Within the scope of the project were the following major objectives:
- To develop and install the methods, procedures, and managing techniques necessary for achieving high efficiency, quality, and output from this “state-of-the-art” facility; and
- To refine the traditional organizational duties and responsibilities as well as the staffing levels of all plant personnel in order to improve utilization and productivity of these resources.
The two-year involvement was separated into two distinct projects: Production/Maintenance and Quality Assurance. Each of the objectives was specifically addressed in each project area. The resulting “product” acted as a prototype for the management techniques and organizational structure for the remainder of the client’s manufacturing organization.
Efficiency and Quality
In both of the project areas, the following efficiency and quality elements were addressed:
- operating procedures
- skills training
- problem solving techniques
- management control
- performance measurement
Procedures for how to keep each piece of equipment clean, and for when to clean it were developed and implemented. On-the-floor training in the evaluation of cleaning effectiveness for all levels of plant supervision and management was conducted. Random audits of cleaning effectiveness were installed. Cleaning was stressed early because of its perceived impact on equipment longevity and performance in addition to reasons of basic sanitation.
Because almost every piece of equipment in the facility was new to every operator and mechanic working there, on-the-floor skills training programs were developed. These programs dealt with real-time evaluation of the extensive on-line diagnostics available to the production staff. The skills training ultimately led to the development of a prioritized system of problem solving.
The standard collection of efficiency and production data was enlarged to include various reject and downtime data. This data was summarized for various levels of management in both the Production and Maintenance arenas for reasons which included not only operator and mechanic performance evaluations but also preventive and predictive maintenance. This data collection became essentially fully-automated by the conclusion of the start-up phase.
The combination of prioritized problem solving and automated data collection led to the concept of an on-line, real-time system of diagnostics for production and maintenance personnel to use in managing the production equipment. This system, which combines statistical process control and artificial intelligence, has resulted in improved efficiency and effectiveness at troubleshooting and in the ability to anticipate mechanical problems rather than waiting for them to occur. This, in turn, has resulted in an increase in net production of over ten percent over the projected plant production levels.
Within the Quality Assurance area, while the optimization of all QA resources was the charter, the transition of the organization’s orientation from a “quality control” role to one of actual quality assurance was the goal. Early in the involvement with Quality Assurance, it became obvious that the Quality Assurance function was “trying to inspect quality into the product” which is clearly impossible. Rather than random sampling, the department was using systematic sampling. Sample sizes were so small that statistical confidence in the reported results was negligible. Inspectors were looking for over one hundred defects within each pack of cigarettes, most of which had nothing to do with the consumer’s perceptions of quality as indicated by a combination of consumer panels and customer complaint data.
The specific objectives of the Quality Assurance part of the project were as follows:
- To develop a cost effective Quality Assurance program with both short and long term goals.
- To define the functional roles within the plant organization and create the teamwork essential for a good Quality Assurance program.
- To improve the product sampling validity to achieve acceptable and reasonable confidence levels.
- To implement control procedures that impact and reduce internal costs associated with defects.
- To reduce the cost of in-plant laboratory testing.
- To redirect the efforts of some Quality Assurance resources into areas where cost/quality ratios will be more productive.
As part of this resource optimization effort, the team of client and consulting personnel set out to define the role of Quality Assurance at the plant level and the role of each functional area with respect to quality. A task force was assembled to involve the following functional areas:
- Quality Assurance;
- Product/Process Control Engineering; and
- Plant Production Engineering (Maintenance).
The task force process prompted each member to reevaluate the duties and inter-relationships of each of the functional areas in general. Once these duties and inter-relationships were defined, the task force examined and redefined the required activities of the quality control/assurance process at the plant level, including the identification of the functional area which would be responsible for each activity. All recommendations were approved by plant and company management.
As part of the effort to improve the product sampling validity to achieve acceptable and reasonable confidence levels, an appropriate scenario was developed by which Production could control the visual defects and Quality Assurance could provide a statistically-based estimate of the quality level of the out-going product. The objective was to make the program more consistent, more relevant to customer complaints, and more statistically based. This required fewer defects, a randomized sequence for sampling, and larger volumes of product than had been used in previous inspection programs.
The first task was to reduce the number of defects checked to a number which included only those which were truly related to product quality. This, as expected, turned out to be more of a political issue than anything else. The standard inspection categories and the customer complaint items were evaluated to provide the framework for determining which attributes of the final product were significant in nature and which were inconsequential. While most people involved in the decision could agree that fewer than thirty defects could actually be associated with perceived product quality, many of those same people felt that the other defects should be checked if only for historical reasons.
Ultimately, the deciding factor was practical. It was proven to the upper-management task force charged with this issue that, in order to inspect enough product to provide the desired high degree of confidence in the results, maintenance of the existing number of defects would require a seven-fold increase in inspection personnel. Agreement was ultimately reached on twenty-seven defect categories.
Given the new twenty-seven defect categories, the responsibilities and mechanisms for the control and audit of these attributes needed re-alignment. This, in turn, required a re-thinking of the approach to detecting out-of-specification product. In the past, the onus for identifying undesirable product rested with Quality Assurance; in effect, “if QA didn’t catch any bad product, then Production didn’t make any bad product.” This attitude is both inappropriate and costly in today’s highly competitive environment where market-share and competitive pressures must be addressed day to day. To this end, the emphasis shifted to making the product correctly rather than on the ill-founded assumption that Quality Assurance would or could screen out undesirable product after it had been made.
The next task was to transfer the control function from the Quality Assurance area to the Production personnel. Equipment operators and their immediate supervision were given hands-on training in how to inspect for the drastically-reduced number of defects. Plans were also developed, and later implemented, to use optical sensors to perform many of the actual inspections. This ultimately led to the implementation of a random inspection of proper sensor functioning rather than the more labor-intensive manual inspection of product for most of the defects.
Reporting tools were designed around the management structure so that production managers could see those items that related specifically to their area(s) in micro terms and upper management could view more concentrated forms of information to allow them to take a macro view of operations. The reports allowed for the identification of trends and facilitated the determination of any correlation with customer complaints.
Once the control baton was successfully passed to Production and Maintenance (for the inspection of the sensors), Quality Assurance could concentrate on taking larger random samples of product. A sampling plan was developed to accommodate various staffing and production levels. Once implemented, this procedure yielded higher confidence levels in the reported out-going product quality that were then reported together on a shift, daily, weekly, and monthly basis by production unit and by brand.
As mentioned earlier, the new, highly automated environment required new ways of viewing the operation and management of the facility, and presented opportunities to staff the facility more efficiently and effectively than had previously been done.
The elements which were addressed within the staffing/organizational aspects are as follows:
- Task Lists: activities, frequencies, times required, criticality, and degree of difficulty;
- Degree of Human Resource Utilization; and
- Potential for revisions to staffing/organizational structure.
Within each Production and Maintenance area, every task for every hourly worker was identified. Each task was then broken down into the specific activities within that task, the frequency (hourly, each shift, daily, etc.) of the task, the amount of time required to perform the task, the relative importance of the task, and the relative amount of skill required to perform the task. These task lists were then used to update the various skills training programs and to more clearly define individual responsibilities.
While developing the task lists, preventive and predictive maintenance procedures and responsibilities were updated. These task lists also served as a basis for the development of trouble-shooting guides for Maintenance and Production personnel.
The task database ultimately included over two thousand activities to be performed. These included routine operational and preventive maintenance tasks as well as troubleshooting procedures. The process of compiling this database resulted in the identification of duplication of responsibility, equipment errors, and various omissions. It also provided one standardized format rather than the collection of formats which had been assembled over the years.
Ten different staffing options were developed from the task analysis. Each option represented various combinations of the following:
- combinations of operator and mechanic tasks
- reduction of staffing levels for various positions
- transfer of organizational responsibility between Production and Maintenance and/or Quality Assurance
One option was chosen which resulted in the elimination of seventy Production positions, sixty-six Maintenance positions, and sixty-two Quality Assurance positions from the total payroll of almost two thousand hourly and supervisory personnel. These positions were cut through a combination of attrition and early retirement incentives. Total annualized savings exceeded eight million dollars. Where possible these staffing configurations were implemented in the client’s other facilities for additional labor savings in excess of five million dollars.
Summary of Project Results
- Staffing levels within all areas were determined with the direct involvement of plant management and departmental supervision; changes from existing staffing levels were implemented.
- Potential capital expense items were evaluated with respect to their net benefit and impact on process capabilities. One such evaluation forestalled the purchase of seventy-two $700,000 pieces of equipment.
- Organizational roles were defined or re-defined to provide a logical basis for departmental accountabilities.
- Procedures were established or modified to address identified needs that were either not being met or were not being met adequately.
- Work flows were evaluated to minimize bottlenecks and unnecessary handling of materials.
- Management control and reporting systems were designed or updated to accommodate the specific challenges and capabilities of the “state-of-the-art” facility and equipment.
- Supervisors and managers were trained, in actual situations, in the use of the tools at their disposal for analysis of correctable problems and in follow-up for control and to stimulate horizontal and vertical communication.
- A cultural change was begun whereby a continual examination of procedures, processes, and accountabilities could occur. Barriers began to be viewed, not as roadblocks, but rather as items to be overcome in the improvement process.