Business Process Discovery (BPD) in relation to process mining, is the endeavor to portray current
business processes and business variations in an honest light. With this, organizations can asses situations and problems with the intent of improvement. There are several components to BPD, each playing a part in its functionality.
Emergent Paradigm- Where current business methods are organized from top-down manual operations, relying on second-hand interpretations of business data. An automated system of data collection relies on the collection of data over a certain period of time. From there, data can be analyzed to reform the business process.
Automated Process Discovery- By automating the analysis of data, the subjectivity of current manual process analysis techniques is removed. The analysis system has an ingrained methodology that- through trials and periods of reform- has been shown to more accurately depict business data and process variations without the infringement of bias.
Accurate Information- Since the information is being collected directly from the source, it cannot be inaccurate. This is a large step up from the second-hand system that automation replaces.
Complete Information- An automated system collects all the data about a given business process: users, names, dates, and subject, compiling it all into a manageable report. Since the data is collected in real time, there is no possibility for memory loss or partial information due to strained time periods. Included in this completeness is the presence of business process exceptions. Quite often, exceptions are regarded as kind of, "white noise". In reports, however, they can do a lot to shed light on exactly what is going on in the business process.
Standardized Process- The automated collection of data yields process data that can be grouped, classified and organized in almost any way that the user wishes. This organization supplies a base from which new processes can be discovered and developed, and benchmarks can be placed. These benchmarks serve as the foundation for the new business process, as well as to point to the root of the original problem.
An example may better depict the usefulness of BPD. Automated BPD captures the required data, and transform it into what is necessary to make a diagnosis. What is challenging abut this however is organizing habitual actions into meaningful events.
Next, BPD tools propose possible process models. Possible new models are crucial for diagnosing the faults in the current model. The following is a chart where a probabilistic repair process is recovered from user actions. The, "as-is" model shows exactly where this business is lacking. 5% of faulty repairs does not create a reliable name, and the extra repairs required to correct these mistakes are cumbersome.
Figure 1- A model Business Process Discovery
A deeper look into this, "as-is", chart could even reveal the specific parts that are at fault for the extra repairs. This discovery could lead to subgroups f necessary repairs that require executive input for improvement.
Figure 2- A Deeper Look Into the, "As-is" Chart
With some analysis, it would become obvious that the faulty parts are what create the need for repetitive repairs.
History of BPD
Business Process Discovery emerged nearly twenty years ago, with the dawn of automated business processes such as EDI. However, some BI programs and systems are not yet fully equipped to analyze data in the detail necessary. In thee cases, the human dynamic is still preferred.
The main root of BPD came from about twenty years ago, when the Alpha algorithm was introduced. The Alpha algorithm made it possible to extract process models from event logs. The algorithm itself was very primitive, and could not do very much beyond it's stated function.
BPD is an essential piece of Business Process Management (BPM). It allows organizations to assess success rates and factors, and decipher what can be done to improve overall performance inside the organization.
business processes and business variations in an honest light. With this, organizations can asses situations and problems with the intent of improvement. There are several components to BPD, each playing a part in its functionality. Emergent Paradigm- Where current business methods are organized from top-down manual operations, relying on second-hand interpretations of business data. An automated system of data collection relies on the collection of data over a certain period of time. From there, data can be analyzed to reform the business process.
Automated Process Discovery- By automating the analysis of data, the subjectivity of current manual process analysis techniques is removed. The analysis system has an ingrained methodology that- through trials and periods of reform- has been shown to more accurately depict business data and process variations without the infringement of bias.
Accurate Information- Since the information is being collected directly from the source, it cannot be inaccurate. This is a large step up from the second-hand system that automation replaces.
Complete Information- An automated system collects all the data about a given business process: users, names, dates, and subject, compiling it all into a manageable report. Since the data is collected in real time, there is no possibility for memory loss or partial information due to strained time periods. Included in this completeness is the presence of business process exceptions. Quite often, exceptions are regarded as kind of, "white noise". In reports, however, they can do a lot to shed light on exactly what is going on in the business process.
Standardized Process- The automated collection of data yields process data that can be grouped, classified and organized in almost any way that the user wishes. This organization supplies a base from which new processes can be discovered and developed, and benchmarks can be placed. These benchmarks serve as the foundation for the new business process, as well as to point to the root of the original problem.
An example may better depict the usefulness of BPD. Automated BPD captures the required data, and transform it into what is necessary to make a diagnosis. What is challenging abut this however is organizing habitual actions into meaningful events.
Next, BPD tools propose possible process models. Possible new models are crucial for diagnosing the faults in the current model. The following is a chart where a probabilistic repair process is recovered from user actions. The, "as-is" model shows exactly where this business is lacking. 5% of faulty repairs does not create a reliable name, and the extra repairs required to correct these mistakes are cumbersome.
A deeper look into this, "as-is", chart could even reveal the specific parts that are at fault for the extra repairs. This discovery could lead to subgroups f necessary repairs that require executive input for improvement.
With some analysis, it would become obvious that the faulty parts are what create the need for repetitive repairs.
History of BPD
Business Process Discovery emerged nearly twenty years ago, with the dawn of automated business processes such as EDI. However, some BI programs and systems are not yet fully equipped to analyze data in the detail necessary. In thee cases, the human dynamic is still preferred.
The main root of BPD came from about twenty years ago, when the Alpha algorithm was introduced. The Alpha algorithm made it possible to extract process models from event logs. The algorithm itself was very primitive, and could not do very much beyond it's stated function.
BPD is an essential piece of Business Process Management (BPM). It allows organizations to assess success rates and factors, and decipher what can be done to improve overall performance inside the organization.
Comments