Process Reengineering
4.4 Tools and Techniques for Gap Analysis
The next step is to determine the “gap” between the current processes and the desired future processes and the reasons for the divergence. The primary objective is to establish the basis for major process initiatives—the necessary steps to “get from here to there.” This analysis can also assist in delineating improvements of current processes that can be undertaken immediately (“quick wins”), identifying barriers to longer-term change, and can provide inputs for analyses of the benefits and costs involved in undertaking the required changes.
Several common tools or techniques may be applied to assist in this gap analysis.
Brainstorming is a group problem-solving method that taps the creative ability of participants to identify problems and their causes by eliciting numerous ideas in a relatively short time. The leader of the brainstorming session should encourage participant’s ideas and involvement, while exercising enough control to keep the group on track by walking a fine line between a free exchange
and control. Brainstorming can be a key to team-building. Individual members are encouraged to contribute to the working of the group and to develop trust for the other members. The following guidelines can facilitate a productive brainstorming session.
Be sure that everyone understands the topic or problem under discussion in the brainstorming session.
Each person contributes one idea in turn; if someone thinks of an idea when it is not their turn, it should be written down to be used during their next turn.
Every idea should be recorded (usually on flip charts), and the contributor should be satisfied with the way the idea is written down.
Wild “off-the-wall” ideas should be encouraged; they may trigger someone else’s thinking.
Criticism should be held until after the session since it may block the free flow of ideas. The goal of brainstorming is creativity and quantity.
Allow a few hours (or days) after the initial session for further thought—
an incubation period that allows the mind to release more creative ideas and insights.
If a brainstorming session appears to be slowing down, the leader may suggest
“piggybacking” by building on the ideas of others. Other techniques are to suggest opposites to ideas already recorded, to try quick associations, and to prod the brainstorming by tossing out ideas that lead the discussion in certain directions or explores possible issues or causes in greater depth.
Once a brainstorming session has run its course, it is likely that the list of ideas will be rather jumbled and in need of organization. An affinity diagram (also know as theKJ method, after its creator Kawakita Jiro) may be used to organize the various ideas generated through brainstorming into natural groupings of related thoughts. A header card identifies the common theme of each grouping of ideas. Stray ideas that do not fit any grouping may be explored further or assigned to a “parking lot” for subsequent discussion.
The cause-and-effect or “fishbone” diagram was developed in order to identify, explore, and display all the possible causes of a specific problem or condition [20]. Cause-and-effect diagrams are designed to focus on the cause of the problem instead of the problem itself. Like other tools, it may take a little while to become familiar with its application, but its ease of use makes this method a popular tool for process reengineering.
Cause-and-effect diagrams sort and segregate possible causes of problems into a logical order, identify areas for further data gathering activity, help to educate participants in problem-solving processes, serve as guides for discussion and help to keep meetings on target, and can be developed into a complete project management tool that displays actions taken and results achieved. In building
a cause-and-effect diagram, it is important to define the characteristic (issue or problem) that is to be controlled or improved. This characteristic should be able to be quantified and measured.
The name “fishbone” comes from what the diagram looks like (see Figure 6.3). It is made up of a horizontal line (the spine) with a triangle or box at one end (the head), in which the problem is stated. Several angled lines come off the horizontal line forming the ribs or bones of the fish. A probable cause of the problem is listed at the end of each bone. Major cause categories that typically are used are shown in Figure 6.3. Contributors to the cause (or subcauses) are usually put on smaller bones off the rib. There is no perfect set or number of categories (causes and subcauses); they should fit the problems or processes under analysis.
A cause-and-effect diagram can be used to organize and sort ideas from brainstorming sessions into basic categories. Relationships between ideas are shown, and gaps from the brainstorming that need filling often can be identified.
The diagram serves as a record of the brainstorm and helps to track the status of the group in the problem-solving process.
Cause-and-effect diagrams have two important limitations, however. Due to size limitations, the depth of search possible in the diagram cannot exceed two or three levels. Organizational problems often are very complex and, consequently, limited analytical success may result with this method. The diagram does not offer a quantitative mechanism with which to rank the critical variables. In other words, the fishbone does not help in identifying those variables that most significantly affect the quality of performance.
FIGURE6.3 Cause-and-effect or fishbone diagram.
The best use of a cause-and-effect or fishbone diagram is when a specific problem area is known to need analysis, but there is some uncertainty as to which factors are creating the problem. For example, customer dissatisfaction with a financial process may be identified as the head of the fishbone, with some of the spines being communication problems, process inefficiencies, information reliability, lack of customer service, and issues of affordability.
Using brainstorming techniques, each of these “spines” could be examined in greater detail (possibly even doing a fishbone diagram on each of them). A determination is made as to which areas were most instrumental in creating the dissatisfaction. The focus can then be on solutions that would have the most impact on the customers or end-users.
Storyboarding is a group process, like brainstorming, that encourages participation, creativity, and trust. A major difference is that, from the outset, storyboarding is more structured than brainstorming. Once the purpose of the storyboard session has been identified and understood by the participants, the main categories are selected that will serve as the focus of the exercise. Ideas are then generated and organized under these main headings, similar to the approach used in developing an affinity diagram. Participants write down their ideas on individual cards, which are then posted (pinned) on the storyboard under the appropriate heading (as selected by the participant). If an idea fits under more than one category, a duplicate card is prepared. If only a few cards are posted under a major category, the group’s attention should be directed to this heading and additional ideas solicited or the heading re-assigned to another category.
New headings also may be added during the storyboard session. During the follow-up evaluation, the most appropriate items are selected and prioritized by the group, and specific assignments are made.
Root cause analysis helps to describewhat happened during a particular event, to determine how it happened, and to understandwhy it happened. The goal of root cause analysis is to find the real problem as opposed to simply treating symptoms. The most basic method to find a root cause is to ask “why”
(at least) five times (see Figure 6.4). A decision diagram (root cause tree) may aid the investigator in identifying root causes. Once the root cause analysis has been completed, it is important to make sure all team members see the problem in the same manner to help direct the search for alternative recommendations.
Each team member should prepare a brief description of the problem. Individual statements should be collated, and the team participants then should prepare a collective statement of the problem (consensus) for review and concurrence.
Atree diagramsubdivides a process, event, or purpose into its component items (see Figure 6.5). The entries on the diagram, when read from top to bottom, progress from general purpose to specific actions, and answer the question
“how is the purpose accomplished?” If read from bottom to top, the diagram answers the question “why?”—the logic or justification for certain actions being undertaken.
FIGURE6.4 Root cause analysis.
FIGURE6.5 Tree diagram.
A matrix diagram shows the relationships between two or more sets of items. The items are arranged in rows and columns on a chart that shows the presence or absence of linkages among collected pairs. The diagram facilitates an analysis of the relationship of each item in one set to all items in the other set(s). Thinking is often triggered that might not have occurred if this organized approach had not been used. The diagram is also helpful in identifying patterns of relationships: which items are major factors and which items do not relate to anything.
ThePareto diagramis a fundamental tool that can be extremely effective in determining which characteristic is causing problems in a given process. This technique is a deceptively simple, yet powerful, method for looking at the data to help find the root cause of a problem. It derives its name from the work of Vilfredo Pareto (1848–1923), an Italian economist who concluded that a fairly consistent minority (about 20%) of the people controlled the large majority (about 80%) of a society’s wealth. Joseph M. Juran expanded on this notion to suggest that only 20% of the possible causes produce 80% of all effects—that is, a few root problems are responsible for the large majority of consequences.
In practice, the percentages may not always be exactly 20:80, but there usually are “the vital few and the trivial many.”
A Pareto diagram is constructed by categorizing and ranking data by frequency of occurrence or impact of problems. These data are then plotted in bar-chart form in descending order along the x-axis (see Figure 6.6). A cumulative sum line shows the percentage contribution of all preceding bars.
This plot allows the analyst to determine the most important factor in a given process or situation. Sometimes it can be advantageous to plot dollars on the y-axis to emphasize the cost factor. After a Pareto diagram is plotted, the data will indicate what problem is most prevalent, and what kind of leverage can be gained from solving it. The diagram also provides strong clues as to what is the root cause of the difficulty. A Pareto diagram serves as a way to compare problems that existed before efforts were initiated to improve a process with those problems that still existed after the improvements were launched.
In some instances, however, the Pareto diagram will not appear to work—
as when all categories are virtually equivalent and there is only a small gain to be made by solving the “worst” one. Not all Pareto analyses produce results leading to the detection of determinate causes and workable solutions. However, if the analysis does not produce a readily apparent cause at a reasonable leverage, it may be possible to sort the data using another characteristic to determine the true culprit variable.
Forecastingseeks answers to key questions about possible and/or probable futures: (1) What happens to object A in time B if the current course of activities is allowed to continue without change? and (2) What are the likely outcomes if change C is introduced? The most primitive method of forecasting isguessing.
FIGURE6.6 Pareto diagram.
The forecast may be more acceptable if the person making the guess is an expert in the matter. Guessing makes use of tacit knowledge that a specialist cannot express in exact words or numbers.
The best method for eliciting such an expert forecast is the unstructured interview that permits inquiry into the reasons and explanations for the presented forecast. Some things may be learned when interviewing an expert that later can be used to construct expanded forecasts through other methods. A questionnaire may be used instead of an interview when the number of experts to be consulted are numerous and/or at a distance.
In the Delphi method, identical questions are directed to a group of experts, asking for their opinions on the future development of a specific topic.
A summary is made of the replies received, and this summary is sent to the respondents, giving them the opportunity to revise their original responses. This procedure may be repeated several times until the responses approach some sort of consensus.
Extrapolation—the most usual method of forecasting—is based on the assumption that the course of events will continue in the same direction and with unvarying speed (or with steadily growing or diminishing speed, i.e., as a logarithmic extrapolation) until some major intervention occurs. The basis of an extrapolation is knowledge of the recent development of the phenomenon—
sequential observations made at known points in time.
Statistical models, such aslinear correlationandregression analysis, can be applied if the time series of observations is sufficiently large. Forecasting on the basis of statistical models often is feasible and successful even when the reason or explanation of the mathematical association found in the historical data is not known. Generally speaking, an effort should be made to determine the rational explanation behind a statistical association assumed as the basis of a forecast. It is always safer to forecast on the basis of a causal model than to forecast only on the basis of statistical association.
The causal model is the most accurate method of forecasting. This approach becomes possible when, through research, relationships are identified that show how the attributes and variables to be predicted depend on each other.
In the best case, one of the variables in the causal model is time. When the correct time period is inserted, the model immediately becomes the desired forecast.