






|
|
|
|
|
|
|
|







| The purpose of Business Process Modeling (BPM) is the construction of a concise, unambiguous representation of a business or business area in terms of: |
|
|
|
|
|
|
|
|
|
|
|







|
|
|
|
|
|
|
|
|














|
|







|
|
|
|
|
|
|
|
| More Inclusive Usage |
| More inclusively, the term "Business Process" may be used to denote any set of activities that together achieve some defined goal, regardless of whether that process begins or ends outside the enterprise boundary. |














| This technique uses an activity-based modeling approach called IDEF0. It also references the IDEF1/IDEF1X and the IDEF3 methods. |
| “IDEF” stands for “ICAM DEFinition”, where “ICAM” in turn stands for “Integrated Computer-Aided Manufacturing”. |
| Background to the IDEF Methods |
| IDEF0 Process Modeling |
| IDEF3 Workflow Modeling |
| Use of Data Flow Diagramming in an IDEF0 Context |
| Relationship to Static (Data and/or Object) Modeling |







| During the 1970s, the U.S. Air Force Program for Integrated Computer Aided Manufacturing (ICAM) sought to increase manufacturing productivity through systematic application of computer technology. The ICAM program identified the need for better analysis and communication techniques for people involved in improving manufacturing productivity. |
| As a result, the ICAM program developed a series of techniques known as the IDEF (ICAM Definition) techniques which included the following: |
|
|
|
| Note: IDEF2 is not referenced further in this technique. |
| In 1983, the U.S. Air Force Integrated Information Support System program enhanced the IDEF1 information modeling technique to form IDEF1X (IDEF1 Extended), a semantic data modeling technique. |
| In parallel, the IDEF3 method was also developed as part of the US Air Force’s Information Integration for Concurrent Engineering (IICE) program, sponsored by the Armstrong Laboratory’s Logistic Research Division. |







| The BPM technique makes use of the IDEF0 Process Modeling technique as defined in the Federal Information Processing Standards (FIPS) Publication 183 (see [IDEF0]). IDEF0 has a number of advantages in this context. |
|
|
|
|







|
|
|
|







|
|







|
|
|







|
|
|







|
|
|
|
|
|














|
|







|
|







|
|
|
|
|
|
|
|







|







|







|
|
|
|
| Note, however, that a technique or approach that is used to execute an activity is classified as a control, not as a mechanism. |







| For fast-track application of this technique, certain aspects of the notation and technique may be abridged or omitted entirely, as suggested in the following notes. Further notes, in addition to those provided here, have been embedded in various sections of this technique. The names of the sections containing “BPM Lite” notes have been flagged with an “(L)”. |







| Scoping and information gathering may be carried out and documented informally. However, it is still essential to know or to determine the objectives of the modeling assignment, as described in Step1. |







| Provided that an IDEF0 model is not very large, it need not be documented using the whole repertoire of cross-reference and identification fields that the standard prescribes for the Working and Identification fields on the standard form. In fact, provided that the relationship between different model components remains clear, the form itself may be dispensed with. Instead, the model may be initially drafted on white board, flip chart, or blank paper, and thereafter composed into a document whose structure reflects the hierarchical relationships between the different parts. |







| The IDEF0 standard recommends a carefully sequential procedure for working on different parts of the model. However, this approach can impose too great an overhead in the form of the model-reconcile-rework cycle. Instead of working sequentially, attention may be cycled rapidly between levels and processes quickly re-assigned between diagrams, either between siblings or between parents and children, rather than attending to levels separately as described in Steps 3 and 4. However, it is still advisable to pay particularly careful attention to definition of the context diagram, since the single most critical step in constructing the model is the definition of its scope and context. |







| Depending on the purpose of the modeling assignment, it is not always important to construct a static model in parallel with the activity-based model. If the aim is just to get a clearer view of the activities taking place, but not necessarily to prepare for the commissioning of new business activities and information systems with significant data content, such a model may be superfluous. However, BPM is commonly carried out early on in a series of analysis activities that will lead to the adoption of new technology and/or the commissioning of new systems. In such cases, it will almost always be worth constructing a static model in conjunction with the activity-based model. Note also that, if a static model is constructed, it should be reconciled with the activity-based one as described in Step 5. |







| This technique prescribes a submit-review-finalize cycle within its last two steps. However, in less formal settings, it may be possible to elide these steps, thereby shortening the amount of time required to get from the initial draft of the model to the accepted version. Note that it is still important to review the model and to process all reviewer comments in a systematic manner so that none of the benefit of submitting the model to review is lost. |














| IDEF0 is a predominantly diagrammatic standard, but there are also well-defined rules for the textual components that support the diagrams. In particular, each diagram must be drawn on a standard form, detailing certain fields, which are explained next. The standard layout for such a form may be seen in Annex C.4 of [IDEF0]. |
| In the BPM Lite variant of this technique, some of the formal documentation elements prescribed by the IDEF0 standard may be simplified or omitted. [See the "BPM Lite" section, above.] |
| The standard diagram form consists of three sections: |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







|
|














|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







|
|
|
|
|







|
|
|







|
|
|
|
|







|
|
|
|
|
|







|
|
|
|







|
|
|
|
|
|
|







|
|
|
|
|
|
|














|
|
|
|
|
|
| In the BPM Lite variant of this technique, it will often be possible to simplify the modeling assignment by limiting the model to one or two viewpoints. |
| Documenting the purpose of the model clarifies the aim of carrying out the modeling assignment, and the goal of communication that the model is intended to serve. IDEF0 models can be used for a wide variety of purposes, including functional specification, implementation design, customer operations, and so on. In the context of BPM, however, it is particularly important to discover and document the following aspects of the model's purpose. |







| Is the model being constructed to enable Business Process Re-engineering (BPR), Business Process Improvement (BPI), or is there some specific business motivation behind the study, such as discovering opportunities to exploit new technology? |
|
|
|







| How extensive is the area of business operations that must be documented? Modelers can err both by including too much and too little in the scope of their assignment. |
|
|







| Is the scope, purpose, criticality and likely content of the BPM study sufficient to merit the construction of a static (data or object) model alongside the process model? If so, the modelers must choose which modeling technique and notation to use, and ensure that arrow data are reconciled with static model content as the process model is built up. (If IDEF0 is being used for the process model, then IDEF1 is usually the notation of choice for the static model, however.) |
| In this technique, we discuss static models particularly in the Concepts section (see sub-section "Relationship to Static (Data and/or Object) Modeling"), and in Step 5 ("Check Model against, and possibly Integrate it with, Static Models"). |
| In the BPM Lite variant of this technique, one way of scaling down the effort and duration required for its application is to omit the construction of a static model and its reconciliation with the process model. However, this can only be done with safety if the data content on which the process model depends is truly insignificant, given the scope and purpose of the model. |







| Every study of organizational processes provides an invaluable opportunity to determine which are the significant measures of process performance, and to take initial samples of those measures. Indeed, it may be said that a purely qualitative understanding of business processes is of little value. |
| Common measures that are always relevant are as follows. |
|
|
|
| Sometimes the effort required to carry out an activity will also be relevant, especially when contrasted with the elapsed time required. For example, one study of a process to generate an initial estimate for a mainframe installation discovered that a few hours’ work was spread over several days because of the number and inefficiency of the hand-offs that were built into the process. |
| Definition and sampling of measures is an area in which a modeling assignment can easily over- or under-shoot the expectations of the sponsors. Modelers must be particularly careful to determine how much effort should appropriately be devoted to this aspect of the work, in light of the terms of reference that they are given, the constraints set, and the overall purpose of the study. |







| In a sense, having a second step concerned with information gathering is somewhat artificial because, in reality: |
|
|
|
|
|
|
|
|
|
|
|
|
|







|
|
|
|
|
|
|
|
|
|
|







|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







|
|
|
|
|
|
|
|
|







|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







| Finalize the model by processing all review comments. File the model according to approved project filing, versioning and configuration management procedures. |







| This tutorial describes a small modeling assignment that is being carried out as a pilot by an analysis team at the Widget Warehouse Company. The Senior Management of the company believes that there are opportunities for significant process improvement, and they wish to start an improvement program. |
| Step 1. Determine the Scope and Purpose of the Model |
| Step 2. Gather and Organize Business Information |
| Step 3. Define the Context Diagram and the Top-Most Diagram of the Model |
| Step 4. Define Lower Levels of the Model and Supporting Material |
| Step 5. Check Model against, and possibly Integrate it with, Static Models |
| Steps 6 and 7. Review the Model and Finalize the Content |
| Step 7. Finalize Model Content |







| In this step, the analysis team is concerned to determine and document the scope and purpose of the model assignment. |
| Scope: Context |
| The analysts are told that the aim of the study will be to gain experience in using the BPM technique, hence only a small area of the company’s operations is being targeted initially. The scope of the model will be to cover a part of the inventory control activities within the company. They may choose exactly which set of inventory control activities to target as the assignment proceeds; the only proviso being that there should be enough in the chosen area to be worth studying and to allow the team to gain meaningful experience in constructing a process model. |
| Scope: Viewpoint |
| The modeling viewpoint will be an operational one; i.e., the view of the activities being carried out as seen by those performing them, rather than, for example, an accounting view of those same activities, which would look quite different. |
| Purpose |
| As stated previously, this is a pilot study to gain experience with the modeling technique, and to “debug” the process. Based on these experiences, a full plan and estimates for the initial phase of the company’s BPI program will be drawn up. |
| BPR or BPI? |
| The bias of the study is towards BPI rather than BPR. Furthermore, the BPI program may eventually lead to the adoption of new technology, but that is not the focus of the pilot, which will attempt only to construct an ‘AS-IS’ model. |
| Depth and Breadth |
| The pilot study must be time-boxed to three weeks, so it will be important not to allow the scope to become too wide or deep. The analysts will aim to find an interesting but well-circumscribed aspect of the company’s operations to study. |
| Necessity for Static Modeling |
| For the purposes of the current study, static modeling may be omitted unless it turns out to be essential to an understanding of the area being modeled. |
| Importance of Measuring Process Performance |
| The sponsors expect that simple measures of process performance will be identified, and baseline readings taken in the area studied. The relevant measures will be activity completion time and, if possible, cost. Documentation of measures and measurements may be limited to the most detailed level of the model in order to preserve time-box constraints. |







| Initial investigations of the inventory control activities in the company reveal that there are several separate areas of activity: |
|
|
|
|
|
|
|
|
| After a series of time-limited interviews with key staff, the relationships between these areas become clearer. It seems to the analysis team that materials replenishment will be a good sub-process to use as the basis for the study. It is reasonably complex, although not excessively so, and measures of performance happen to be reasonably easily available. |







| In this step, the A-0 (context) and A0 modeling levels are attempted. The initial draft of A-0 comes out as follows. |
| So far, so good. The analysts have established from their initial information gathering that the inventory control function deals with receiving stock from suppliers, returning defective or incorrectly delivered stock to suppliers, and issuing stock from the warehouse. |
| The next task is to construct an A0 (top-level child) diagram to go with the A-0 (context) diagram. However, constructing this diagram reveals some deficiencies in the context diagram. |
| The A0 diagram raises the questions about: |
|
|
| On the A0 diagram, these questions are answered by two new flows: |
|
|
| Unless these new flow arrows are represented in tunneled form, which does not seem justifiable in this case, they must be reflected on the context diagram also. The analysts therefore re-draw the context diagram as follows. |
| The A-0 and A0 levels now seem consistent and stable enough to justify going on to model lower level activities. |







| In order to stay within their time-box while covering at least one area in full detail, the team decides to concentrate on the “replenish materials” function. |
| A detailed study of the activities in the area reveals the following picture. The estimated costs for each activity are shown in the lower left corner of each activity box. |
| The cycle cost of each occurrence of the “replenish” activity turns out to be around $20. The following picture shows the same activities, only this time the typical duration of each activity is shown in the lower left corner of each activity box. |
| In this sub-model, the activities follow on from one another, so the total cycle time for the “replenish” activity is the total of the individual durations for each of the activities. This adds up to around 48 minutes. |







| Since static modeling has been placed out of scope for this study, this step is null. |







| This small pilot model is reviewed, reworked and finalized using a less formal procedure than that prescribed in [IDEF0]. Copies of the model are prepared in a document that separates the levels into different sections, and carefully documents the scope and purpose, presents the context level, and then the lower levels via a tree-walk of the child diagrams. |
| Review comments are solicited on a simple standard form. Once gathered in, they are classified and disposed according to severity and impact. All review comments that have been accepted and analyzed are processed and the model updated accordingly. |
| The pilot study, small though it is, has already demonstrated that there are several aspects of the replenishment procedure alone that could be optimized. Like many unexamined business processes, it: |
|
|
|
| The scene is now set for further AS-IS modeling to take place, followed by process re-design, leading to improved performance, which will be documented in a TO-BE model. |







| [IDEF0] Draft Federal Information Standards Publication 183 (1993) Integration Definition for Function Modeling (IDEF0). |
| This document defines IDEF0, and contains a large amount of useful information for anyone considering the application of IDEF0. It is available from the NIST web-site, which is at http://www.nist.gov. |
| [Jaco95] Jacobson, I., Ericsson, M., Jacobson, A. (1995) The Object Advantage: Business Process Re-engineering with Object Technology, Addison Wesley, ISBN 0-201-42289-1. |
| This book contains many interesting observations and references on the subject of Business Process Re-engineering and Modeling. However, it focuses on the use of specific object-oriented techniques, which are relatively untested at this time in comparison with IDEF0. |
| [Dave93] Davenport, T. H. (1993) Process Innovation, Re-engineering Work through Information Technology, Harvard Business School Press. |
| This book provides useful definitions, as well as numerous examples of firms that have succeeded or failed in combining business change and technology initiatives. Appropriate emphasis is also placed on new organization structures and human resource programs in facilitating process innovation. |
| [IDEF3] Mayer, R. J., Painter, M. K., Menzel, C. P., Perakath, B., deWitte, P. S., Blinn, T. (1995) Information Integration for Concurrent Engineering (IICE) IDEF3 Process Description Capture Method Report. |
| This document describes the IDEF3 process capture method. As yet it is still in an interim state. Copies can be obtained from the Knowledge-based Systems Inc web-site, which is at http://www.kbsi.com. |
| [DSou98] D'Souza, D. F, and Wills, A. C., (1998), Objects, Components and Frameworks with UML: The Catalysis Approach, Addison-Wesley Pub Co; ISBN: 0201310120. |
| This book describes the Catalysis component-based development (CBD) method. |