لطفا منتظر باشید ...
12.5 Conclusion
12.5.1 Summary of the Methodology
We can summarize the process design methodology as follows: We (re-)design processes by finding a reasonable starting point in the process specialization tree and then generating/exploring the tree, using a set of key questions, to uncover siblings and more distant analogies for each of the core activities and coordination mechanisms. We can also generate alternative orderings of these activities. These alternatives are then organized as a set of multicolumn tables to make the different potential surface structures explicit as selections from each column. We next use trade-off tables to guide our selection of individual column choices, as well as eventually our choice among the candidate surface structures for re-designed processes.
Although the process design methodology has been described here as a linear sequence where we abstract out the deep structure completely before generating and then selecting alternatives, these different steps can be interleaved arbitrarily. One can, for example, create a candidate process re-design by replacing an activity with a ''sibling''(child of a common parent generalization) or even a ''cousin''(child of a common grandparent generalization). This involves, in effect, applying the steps of deep structure analysis, alternatives generation, and selection to a single activity. These steps can also been done to any level of detail; one can generate ''high-level''processes or very detailed ones, depending on to what extent the activities and dependencies are refined by adding decompositions and coordination mechanisms.
12.5.2 Evaluation and Future Work
The methodology described above has been used to (re-)design several processes, of which the hiring example described above is one. Our experience is that the methodology, when used in conjunction with the Process Handbook knowledge base, has been very effective in generating a wide range of novel and promising process design alternatives. The power of this approach appears to come from several sources:
The methodology's recursive application of key questions (who, what, when, where, etc.) fosters the systematic exploration of the space of alternative surface structures.
The use of ''distant analogies''helps uncover farther-flung and potentially powerful re-formulations of the process.
The process knowledge base allows users to leverage the creativity and expertise of process experts from many different domains. The methodology could be utilized, however, even if we did not have such a knowledge base.
Our process design approach offers, we believe, significant advantages over other process design methodologies. It adds, in effect, some science to what today is mainly an art. As discussed earlier, few of the commonly used process re-design methods address the aspect of generating new alternatives. Those that do suggest simple rules of thumb for streamlining an existing process or recommend locking a design team in a room with a facilitator to brainstorm a new ''blank-paper''process. Or they focus on incremental changes despite the usual rhetoric typically associated with business process redesign or reengineering (Jarvenpaa and Stoddard 1995). For example, Harkness, Kettinger, and Segars (1996) found that many process redesign efforts are based on Total Quality Management techniques, which adopt an inherently incremental approach (Mizuno 1988). Other approaches also typically call for an exhaustive analysis of the as is situation (Nissen 1998), which ours does not (although such analysis may well be needed for different purposes such as cost justification).
The Process Handbook and its process design methodology is a work in progress. The generative strength of the methodology is, in particular, a double-edged sword in the sense that it is often easy to uncover an overwhelming number of process alternatives. The procedures are not fully formalized, so human judgment is often needed, for example, to identify core activities and key dependencies, to generate alternatives not represented in the database, to prune out the nonviable surface structure candidates, and eventually select the appropriate processes for one's needs. While we do not expect to obviate the need for human judgment, we do plan to explore how the system can support human users by further reducing the burden of generating and selecting from large design search spaces. For the meantime the general sense of those of us who have used the Handbook is that potentially ''too many''options is preferable to ''too few.''
Other future efforts will include evaluating and refining the process design methodology in other domains (including logistics and manufacturing). Further information on the Handbook and links to several publicly available on-line versions of it are available at the following Web site: http://ccs.mit.edu/ph.
