IDENTIFICATION OF EFFECTIVE PROBLEM-SOLVING TOOLS TO SUPPORT CONTINUOUS PROCESS IMPROVEMENT TEAMS Abstract Continuous process improvement (CPI) teams function differently than other project teams. Therefore, engineers who are called on to support CPI teams may need to utilize different tool sets if they are to be effective in improving manufacturing operations. This article explores the differences between CPI teams and traditional project teams. It also reports on the initial identification of problem-solving tools that meet the needs of CPI teams, and reports preliminary findings on their usefulness and applicability. Data from our 2-year study of 16 CPI teams working in light manufacturing, logistics, and automotive assembly show that tool selection is influenced by team life span, data availability, and team member skill level. This study also found that the primary factors influencing tool effectiveness are team member training and the time required for data analysis. Introduction Goals of this Study. A set of 31 continuous process improvement (CPI) teams was formed to improve the effectiveness of manufacturing operations. While each of the teams operated for only 1 week, the study took place over a 2-year period, from May 1996 through July 1998. The CPI teams were formed to support implementation of lean production techniques. The effectiveness of a CPI team was measured with three metrics: reduction in total travel distance for the product, percent reduction in the number of total processing steps, and percent reduction in the number of non-value-added steps. Travel distance was measured in feet. The number of steps in the process was determined by identifying the number of operations, inspections, delays, movements, and storages the product goes through. An operation was defined as a value-added activity, while all other activities were defined as non-value added. This article reports on the effectiveness of a subset of 16 teams selected from the original 31 CPI teams. While all the teams were led by engineers, these 16 teams were led by engineers who were not experienced in direct line operations but, rather, worked in staff operations such as product development and product technical support. The 16 teams included 11 in light manufacturing, 3 in logistics, and 2 in automotive assembly. This subset of 16 teams was of interest because these teams showed a higher degree of variability in their effectiveness than did the larger set of 31 teams. CPI teams led by staff engineers were found to perform either significantly better or significantly poorer than CPI teams led by line engineers. Reviews pointed to two factors that influenced the high variability in team effectiveness: differences between CPI teams and the traditional project teams with which the engineers had been involved in the past, and the use of different problem-solving tools by the CPI teams for task completion. The goals of this study were (1) to gather information on differences between CPI and traditional project teams, and (2) to identify problem-solving tools used by the CPI teams. Information was gathered from CPI team leaders. The primary data collection methods used were interviews and questionnaires. CPI Team Operations. CPI teams were expected to complete their tasks within 5 days (Exhibit 1). This time frame was established to limit the scope and cost of the CPI team activity. In the majority of cases (30 of 31), the CPI team was a pilot team that was the first attempt by the organization to implement lean manufacturing principles in the operation. The team schedule includes 4 to 8 hours of training, 24 to 28 hours of direct work on the task, and a 4-hour closeout. There was a high degree of variability in team operations. For smaller tasks, a 3- or 4-day schedule may be used. It is more common for CPI teams to put in extra hours, usually on days 2 through 4, to complete their work. The hard requirement is that the assigned task be completed by noon on day 5. Lean Production. The concept of lean production was originally proposed as a way to describe Toyota's production system (Womack et al., 1990). The term "lean" was used to illustrate the emphasis on reducing resources used to meet production requirements. For example, lean production systems were designed to use less labor, space, and time than traditional batch manufacturing systems. Toyota's lean production system also emphasizes training and problem-solving tools (Arunoday, 1992). CPI teams are typically used to support the transition of manufacturing operations to a lean production environment (Hancock and Zayko, 1998). Studies have pointed to a need for close management of the manufacturing process as a prerequisite for world-class manufacturing. Successful implementation of lean production is characterized by (1) training programs on the issue of waste elimination and (2) the use of cross-functional teams, such as CPI teams, to implement changes (Hancock and Zayko, 1998). Metrics, such as reductions in travel distance, processing steps, and non-value-added activities, must also be used to gauge the success of efforts to transition to a lean production environment. Problem-Solving Tools. A common first step when building a problem-solving tool set is to look at the seven basic tools of quality (Ishikawa, 1986). This beginning set includes Pareto charts, trend analysis graphs, check sheets, histograms, scatter plots, cause- and-effect diagrams, and control charts. A second set of seven management and planning tools, which works on language data and relationships rather than numeric data, has been developed (Mizuno, 1988). These tools are commonly used for teams working outside of manufacturing areas, such as project teams in product design and development (NEC IC Micon, 1987). This tool set includes affinity diagrams, relations diagrams, matrix data analysis charts, hierarchy diagrams, matrices and tables to display value and priority, precedence diagrams, and quality function deployment. Other problem-solving tools have been developed to support product development activities and process improvement. Some of these tools include failure modes and effect analysis (FMEA), fault tree analysis (FTA), and potential problem analysis. One method for organizing these problem-solving tools is to group them by function. For example, problem-solving tools can be grouped as follows: (1) tools for the manufacturing process, (2) tools for cycle time, (3) tools for variability, and (4) tools for problem solving (Michalski, 1998). It is evident that there is no shortage of tools to aid teams in problem solving. The limitation is the teams' ability to correctly select and use the appropriate tool to support their problem-solving efforts. Study Plan. The data from the 16 CPI teams led by a staff engineer were compared with the data from the other CPI teams led by line engineers. All of the team leaders had at least a bachelor of science degree in an engineering field. All staff engineers were currently working in product development or technical support. All line engineers were currently working in production or logistics. The analysis of the 16 CPI teams led by staff engineers made up the bulk of this research effort. This work was conducted in two phases. In the first phase, each team leader was interviewed. During the interview, the differences between the team leader's past work history on a project team and their recent work on a CPI team were explored. The following areas were discussed during the interview: Team goals Team membership . Team roles Team processes Team performance measures Support environment Continuous process improvement Results The interview used an open format rather than a structured form. It was administered after the CPI team had completed its work, in order to address the differences in team leader work Exhibit 1. CPI team schedule of activities Day 1 Day 2 Analysis of Training on problem solving current work process Team tools Lean formulates production process improvements Team skills Day 3 Hands-on shop floor improvement Additional process improvement Day 4 Refine improvements Establish work standards Full production under improvements Day 5 Presentation of results Celebration experience, formal education, and experience with teams. The information collected in this format was qualitative, not quantitative. In the second phase, each team leader completed a questionnaire. The questionnaire asked the leaders to identify how familiar they were with each quality tool, whether they used the quality tool during CPI operations, and how successful the tool was in supporting the CPI team. The questionnaire was completed immediately after the interview. The interviews lasted from 30 to 95 minutes, with an average time of 44 minutes. Team leaders completed the questionnaire in 15 to 24 minutes, with an average time of 18 minutes. Interviews and questionnaires were successfully conducted with all 16 CPI team leaders. Research Results Differences between CPI and Project Teams. On the surface, CPI teams differ from project teams primarily in the duration of team activities and the size and scope of the tasks undertaken. A CPI team is typically used to support improvements in manufacturing operations. The CPI team is given a specific task to accomplish in a set time, measured in days. A project team is used to support research and development of an artifact, so it performs a greater number and variety of tasks than does a CPI team. The project duration is also longer, measured in months or years. These differences in types of tasks assigned and duration of team effort have a much more pronounced impact on team operations than is often recognized. This impact can be better explained by looking at specific characteristics of teams. CPI team leaders reported the following differences between their CPI team activities and their past experience with project teams. Team goals. CPI teams typically have a single, well-defined goal, and the team can determine how to measure its attainment. For example, a common CPI team task is to reduce processing time in a department. The CPI team may measure its performance with several different metrics, including reduction in cycle time to complete a unit of production, decrease in unit travel distance, and the amount of non-value-added time a unit spends in a department. The goal is established outside the team, and team members generally are not asked to accept, modify, or reject the goal. A project team's goal is typically defined both outside and inside the team; metrics for measuring team performance are also developed in this way. Team members review and accept the goal. Throughout the project, the goal is often reviewed and modified by the project team. Team membership. CPI team membership is defined outside the team. Team member roles are defined by functional area. For example, a typical 8-person CPI team consisted of 4 operators from the production area the team was working on. The remaining team members typically included a project engineer as team lead, a second administrative team member from accounting or purchasing, and personnel from quality and maintenance. Project team membership is determined inside and outside the team. The team may elect to add or remove members as the work tasks become better defined and some tasks are completed. It is more common to add than to remove team members. Individual team members' roles also shift from representing functional areas to contributing to team goals rather than functional goals. Team roles. At team launch, the only role that is typically defined on a CPI team is that of leader. This position was selected outside the team and was not subject to change during team operations. In project teams, more roles are defined and team members may shift roles as the team develops. The longer a project team operates, the greater the likelihood that team roles will change. Project teams with extended operations may face significant problems due to team member rotations on and off the team. There is also a high probability that team members will take on informal roles that were not originally included in the team charter. For example, project team members may become liaisons with external vendors, provide support to other team members on software issues, or be resident experts on specific technologies. These informal roles were not as common with CPI team operations. Team processes. CPI processes are typically defined outside the team and communicated to members by the leader and through orientation training at the beginning of team operations. This is always the case for personnel who have not been on a team in the past. Due to time constraints, the CPI team does not develop new processes to support their operations. CPI teams also tend to focus on the assigned task. Project teams can begin with externally defined processes but are able to develop internal processes to meet team needs. Project teams focus on both the task and the process used to complete the task. Team performance measures. A CPI team's performance is typically based on accomplishment of the assigned task. Typical measures for CPI teams focus on manufacturing process improvements like reductions in cycle time, travel distance, setup time, and the number of process steps. A typical CPI team does not have process-oriented measures, while project teams typically have both process and product measures. Process measures include families of metrics focusing on schedule and budget. Product measures include metrics such as reducing technical risk and achieving performance design targets. Support environment. CPI team members use existing organizational resources and rarely have additional resources allocated to support completion of their task. Project teams typically operate with an independent budget that allows for expenses in manpower, equipment (hardware and software) and, in some cases, capital expenditures such as buildings. Continuous process improvement. CPI teams are focused on task completion. Process improvement, if it occurs, usually takes the form of gathering lessons learned for future CPI teams. Project teams will include some process, formal or informal, to learn from past experiences both internal and external to the team. Results. CPI team results had a high degree of variability. The biggest predictors of team success reported by CPI team leaders were the support provided by the parent organization to the team and the skills of team members; the ability to function as a team, while important, was not as critical a predictor of team success. Project team results are also affected by the parent organization's support, team members' skills, and the team's ability to function effectively. Problem-Solving Tool Use and Effectiveness. The following problem-solving tools were provided to the CPI teams for use during the training session: Facility layout diagrams. A graphical representation of the facility layout. Typical use is to show the special relationships between departments, sections, processes, or equipment. . Work flow analysis. A table that shows steps in a process along with essential information such as processing times and distance traveled. Information flow analysis. Similar to a work flow analysis, but the item tracked is information. Process analysis charts. Similar to a work flow analysis chart, but used to trace sources of process variability. Often used to support root-cause analysis. Fishbone diagrams. Also know as Isakawa diagrams, this tool is used to identify, classify, and prioritize causes of operational problems. Difficulty-impact grids. Graphic that shows the difficulty and potential impact of different problem solutions. Used to rank solutions to operational problems. Activity analysis. Tool to account for how an employee spends time during a workday. Similar to a work log. Time studies. Procedure to determine the correct or standard time to complete a job. Useful for developing production standards to support cost analysis and work scheduling. Statistical process control (SPC). Family of graphical tools that report performance of key process metrics over time. Commonly used in trend analysis. Failure modes and effect analysis (FMEA). Technique that identifies failures that may occur as a result of process deficiencies. Also provides estimates as to the impact of a particular failure. Fault tree analysis (FTA). A tool that looks for potential causes of defects in production or service activities. The output is a listing of potential causes of system failures. Team Leader Familiarity with Tools. CPI team leaders completed a questionnaire on their familiarity with problem- solving tools (Exhibit 2) prior to the team training session. The questionnaire listed each tool along with a single-sentence description. Team leaders were most familiar with facility layout diagrams (94%), work flow analysis (75%), SPC (75%), and fishbone diagrams (75%), and least familiar with difficulty-impact grids (6%), time studies (13%), and activity analysis (13%). Team leader familiarity with the remaining tools was as follows: information flow diagrams (31%), process analysis chart (38%), FMEA (69%), and FTA (56%). Once the questionnaire was completed, the CPI teams had a 4- to 8-hour training session on these problem-solving tools, lean production, and team skills (Exhibit 1). The training was done at an introductory level, adjusted for individuals' prior team Exhibit 2. CPI team leader knowledge of specific problem-solving tools FTA FMEA SPC Time Studies Activity Analysis Difficulty-Impact Grid Fishbone Diagram Process Analysis Chart Information Flow Diagram Work Flow Analysis Facility Layout Diagram 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Team leader familiarity with tools experience. Each tool was described, and an example of its application was presented. Approximately 10 minutes was spent on each tool, although there was a high degree of variability in the training. At the beginning and end of the training session, the instructor explained that the tools were presented to help the CPI teams complete their assigned tasks. All 16 teams used at least one tool, and the majority of the teams used several tools either as a team or in sub-teams. Problem-Solving Tool Usage by CPI Teams. Most CPI teams used multiple problem-solving tools, although the level and quality of tool use varied significantly (Exhibit 3). The most commonly used problem-solving tools were facility layout diagrams (94%), difficulty-impact grids (88%), fishbone diagrams (81%), and work flow analysis (63%), while the least-used tools were FTA (13%), FMEA (13%), process analysis charts (13%), and SPC (19%). Other problem-solving tool usage was as follows: information flow diagrams (25%), activity analysis (25%), and time studies (38%). Problem-Solving Tool Effectiveness. During the presentation of results (Exhibit 1), the CPI team leader reported on the effectiveness of the problem-solving tools in helping the team complete its assigned task. Only tools used by the CPI team were rated. A scale of 0 to 4 was used to report the tools' effectiveness (Exhibit 4). Tools that received a score of 0 were not included in the average results (Exhibit 5). The most effective tools (Exhibit 5) were the facility layout diagram (3.4), work flow analysis (3.3), information flow diagram (3.3), process analysis chart (3.1), fishbone diagram (3.0), and difficulty-impact grid (2.9), while the least effective tools were activity analysis (1.1), time studies (1.6), SPC (1.6), FMEA (1.7), and FTA (2.2). Discussion of Results CPI teams were given smaller assignments than those given to project teams. The CPI teams also were given a very short time in which to complete the assignment, typically only 1 week. Engineers with prior experience in staff positions reported in interviews that this project team experience often hindered their ability to function as a CPI team leader. Team leaders noted that having a clearly defined goal and understanding the role of team leader were critical to the CPI team's success. Developing a quick understanding of the CPI team process was also important. Team leaders indicated that project teams tended to optimize problem solutions, whereas CPI teams worked toward solutions that could be implemented quickly. CPI and project_teams differed on all eight team characteristics studied. Team leaders and other team members appeared more comfortable with the CPI team process once they began the Exhibit 3. CPI team problem-solving tool usage FTA FMEA SPC Time Studies Activity Analysis Difficulty-Impact Grid Fishbone Diagram Process Analysis Chart Information Flow Diagram Work Flow Analysis Facility Layout Diagram 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Reported tool usage Exhibit 4. Tool effectiveness rating system Score Description Not effective Somewhat effective 0 Tool not used 1 2 3 Generally effective 4 Very effective analysis of the current work processes (Exhibit 1). A common approach for the teams was to use the facility layout diagram to scope the project. Using existing documents or developing one from scratch did not appear to impact this tool's effectiveness. Often, use of an existing layout was disrupted when errors were found. Maintaining accurate layouts is not a high priority in all organizations. Work flow analysis was also effective in the early stage of team operations. It is interesting to note that both the facility layout diagram and the work flow analysis were familiar to many of the team leaders, were used by a large percentage of the teams, and were found to be very effective in supporting the teams' efforts to complete their task. We can hypothesize that team leader knowledge promotes tool use, and that when tools are easy to understand and use, the team will more readily adopt them. Fishbone diagrams were also familiar to team leaders and used by a large percentage of the teams. Again, this tool is easy to understand and use. Fishbone diagrams were commonly used to help identify problems with existing operations after initial use of the facility layout diagram and work flow analysis tools. Some tools that were not well known by team leaders were reported to be effective in supporting the team. Information flow diagrams, while familiar to only 31% of the team leaders and used by only 25% of the teams, received a score of 3.3 (of 4.0) for effectiveness. We hypothesize that the information flow diagram is an effective tool when used appropriately. As the trend toward increased information management technology continues, tools like the information flow diagram may see increasing use. Similarly, the difficulty-impact grid, familiar to only 6% of the team leaders and used by 88% of the teams, received a score of 2.9 (of 4.0) for effectiveness. Again we can hypothesize that ease of understanding and use contributes to a tool's acceptance by the team. Some tools were familiar to team leaders, but were not found to be effective in supporting CPI teams. These include FMEA, SPC, and FTA. If we accept the hypothesis that ease of understanding and use increase the probability that a tool will be Exhibit 5. Reported effectiveness of problem-solving tools by CPI team leaders FTA FMEA SPC Time Studies Activity Analysis Difficulty-Impact Grid Fishbone Diagram Process Analysis Chart Information Flow Diagram Work Flow Analysis Facility Layout Diagram 1.5 2 Tool effectiveness rating 2.5 3 3.5 used by a CPI team, then the limited use of these tools is logical. In general, these tools require more time and experience to use effectively. For example, you can use a draft facility layout diagram as an aid in problem solving, but you will not have the same success with a draft FMEA. One area of concern for us is the lack of understanding and use of time studies. Only 13% of team leaders had an understanding of time studies, although 38% of the teams used them in their analysis. The time studies completed by the teams were not coupled with a methods analysis study or used to develop time standards. We were concerned about time studies because 12 of the 16 teams were working in facilities that had some form of production control system, such as an MRP, MRPII, or ERP, in place. These systems used time standards to support schedule and capacity analysis. For all but 2 of the 12 teams, the time standards used were based on heuristics and were not engineered standards. As expected, significant variances between planned and actual production rates and capacities were common in these facilities. Conclusions The CPI teams used in this study were unique: they focused their process improvement efforts on specific manufacturing problems that could be easily identified, and they were allowed to operate for only 1 week. This approach was used to educate the organization on the use of CPI teams. It was hoped that a successful CPI team activity would provide a model for further CPI activities in the organization. Use of a CPI team was found to be an effective method for making improvements in manufacturing operations. It should be noted that CPI team results may not yield the type of fundamental long-term improvements needed if company management is not committed to CPI (Tippett and Westbrook, 1995). CPI teams were effective in solving production floor problems rapidly, although team performance was highly variable. To improve team performance, team leaders with a traditional engineering background need to be aware of the differences between CPI and project teams. Differences in task scope and task completion time affect many team characteristics, such as team roles and processes. The fact that CPI teams strive to improve, rather than optimize, an operation sets them apart from project teams. Team leader familiarity with problem-solving tools was a good predictor of tool use and effectiveness. Also, problem- solving tools that were easy to understand and apply were used more frequently and reported to be more effective in supporting completion of the team task. Finally, some tools were not commonly used by CPI teams because of the time required to train team members in their use or to correctly apply them. If the organization had used some of the more complicated tools previously, then data from these previous applications were used by some teams, but this was generally not a critical factor in team operations. Of note was the lack of data accuracy in organizations with regard to time standards for production tasks and layout and routing documents. The issue of time standards accuracy is of particular concern, because these standards are used as an essential database for MRP, MRPII, and ERP systems. The accuracy problems regarding layouts and routing documents are of lesser concern to CPI team operations, since the teams tended to re-create these at the beginning of team operations to help familiarize themselves with the process under study.