I am glad because I can start our new semester discussions.
I would like to present you an article that shows process mining usage in real-world company case. Company had a problem with documents flow. Process mining methods were used to solve this problem. Given that this is existing company we have confirmed the process mining methods can be used in real business processes analysis.
https://fedcsis.org/proceedings/2016/pliks/456.pdf
Questions are:
- Do companies compare process model with its execution in production environment?
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
- Do you know any methods for process tracking / analysis / optimization?
Hi,
ReplyDeleteI think that in most companies IT focuses on monitoring process in terms of its health check rather than it's compliance with the design. A good practice, specially with distributed processes is to use one central persistent log with asynchronous interface to improve further analysis and dispose of logs aggregation and transformation step.
I'm not an expert systems expert but I think that process model (e.g. from Enterprise Architect) can be imported and compared to process logs.
For process tracking I use mechanism described above. In terms of analysis and optimizations, it is done programmatically by redesign and re-implementation, since all process I analyze, design, develop are fully automated with no human steps (which cause most problems and create unexpected results).
Thank you for the answer. I agree at the first point where you mentioned that an IT is focused on system condition. In nowadays companies are facing problems with market expectations (high quality, low prices and so on). IT might be a partner in this "war".
DeleteYou can compare process in production environment using process mining methods. Human errors are source of abuses, system crashes etc. but from process comparison perspective it is most interesting part of its analysis.
Process mining techniques are able to extract knowledge from event logs commonly available in today’s information systems. These techniques provide new means to discover, monitor, and improve processes in a variety of application domains. There are two main drivers for the growing interest in process mining. On the one hand, more and more events are being recorded, thus, providing detailed information about the history of processes. On the other hand, there is a need to improve and support business processes in competitive and rapidly changing environments. Therefore process mining is the most important and every year new companies look for more sophisticated issues to achieve your goals in these topics. In my opinion companies should compare process models with their execution in production environment to avoid future problems with compatibility etc. Moreover I don’t know any additional methods for process tracking / analysis / optimization in this area.
ReplyDeleteHello. I agree with you. I want to write some words about input data quality. You highlighted that the quantity of data raises rapidly. (more events are being recorded, thus, providing detailed information about the history of processes). From my experience a quantity of registered information is huge. Never the less its quality is low and in lot of cases when we are trying to extract useful schemes we can not or we have to transform raw input.
Delete- Do companies compare process model with its execution in production environment?
ReplyDeleteI am not an expert in process mining but I think every company should care if methods they use are effective. They should apply methods that are appropriate for their kind of business and supervise if it brings benefits to the company.
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
It is difficult to transfer human know-how to process flow validator because sometimes human knowledge is rather similar to 'unstructured database' than typical process flow. Cooperation of experts and process analyst is crucial. Experts should explain their knowledge in a sufficient way that could be adapted to process analysis. On the other hand, it should be taken into consideration that they don't have to be experts in IT so software should be created as user-friendly.
- Do you know any methods for process tracking / analysis / optimization?
As I'm mentioned before process mining is rather out of the scope of my interest but I think every method should be fitted to the type of company. For some process analysing and tracking could include analysing log files for other analysing social networks if company has any. It would be great to hear about some useful methods which I may use in a future in case of need.
Katarzyna ProM is academic tool but I hope that we have shown its ability to solve real world business problem.
DeleteWhen we are talking about human experts knowledge - we are in strange area. People can give us data about restrictions, or sequence. We met situation where we collected contradictory information from the same person. Next problem is that an important input might be short / single message from the person, which is not involved in process analysis. This seemingly not important data can be crucial to whole activities.
First time I read about this topic so it will be very hard for me to answer on your questions. Human know-how is something more than just understanding information in 0/1 way, it is our experience, our way of thinking. It is very hard to reproduce.
ReplyDeleteHi David. That is the point! ;) Human knowledge is not 0/1. When we are talking about unknown process, which we want to discover - we do not have coherent knowledge. Otherwise we will talk about process comparison (it is part of process mining, too)
DeleteSame as Dawid it's very hard to for me to answer your question because I don't know anything about this topic and second thing is that I don't care. I don't know any methods for process tracking etc. so I don't want to write any bollocks over here. I may that this topic can be quite important when you are working for example in waterfall methodology (this software development methodology love to produces tons of "needed" document, it's real pain in the ass for developer but big companies love it).
ReplyDeleteHello, in my opinion methodology is not such important in this case, but thanks for the reply.
Delete1. Do companies compare process model with its execution in production environment?
ReplyDeleteI have never thought about process mining and this subject is absolutely new to me. I can only imagine that management and administration teams in companies take care of it, because it is part of their’s responsibilities.
2. How can we use experts knowledge in process analysis software?
Well in general I think that people who are not experts should use experts knowledge in any area of interests, especially if it can somehow ease the work.
When it comes to process analysis software I think that experts knowledge is crucial because programmers and other specialists have no idea about process analysis and they need a lot of help from those who work on that for years. Experts can explain how software should work and what are the main functions required. They also should test this software and compare their’s results with software’s results.
3. Do you know any methods for process tracking / analysis / optimization?
I only heard about popular products among software. During software development it is JIRA or simple canvases for agile method programming. For analysis there are dedicated softwares and also for tracing, nowadays almost every company have their’s own CMS and CRM.
Thank you for this post. I agree with you. I described human expert's role in this paper and in my previous replies. Its role is very important because we can not be experts in everything.
DeleteThere is a plenty software to store events and little less to analyze them. To be honest it can be very complex and in my opinion the bigger company is the more complex this type of software should be implemented. Other question is how to retrieve information to the expert system from human experts. The role of knowledge engineer can be an oppressor task.
ReplyDeleteThank you, did you use any software to analyse system logs? In process mining we are trying to extract schema from GB of logs and focus at its deviations.
Delete- Do companies compare process model with its execution in production environment?
ReplyDeleteHonestly, despite the fact that this topic seems to be an interesting issue it is really hard for me to say whether companies compare process model with its execution in production environment.
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
It is always hard to process experts knowledge into adequate algorithms, sets of rules, diagrams, etc. On the other hand, up to a point it is possible.
- Do you know any methods for process tracking / analysis / optimization?
Microsoft Sharepoint is considered as a tool for process tracking. Processing logs analysis, files access analysis - every company has its own tools and solutions for process tracing and optimization.
Good evening. Process extraction from event logs is complicated. Even if we take a look on system log ... we have for example an information about time when action was performed but what with concurrent tasks - how to catch this information? Human knowledge might be this additional input for better process analysis.
DeleteUnfortunately I have a little knowledge about process mapping in organizations.
ReplyDeleteA while ago I saw an interesting article about that: Analysis of Customer Fulfilment with Process Mining: A Case Study in a Telecommunication Company.
This paper presents results of process mining implementation in a characteristically unstructured customer fulfillment process in a real Telecommunication Company. Findings from this study can be used by the company to improve their current business process. It also stressed out the importance of resolving data integration issues in implementation of process mining in real cases.
Thanks a lot, I will take a look on this paper.
DeleteHey, thanks for showcasing to us your article. I remember the presentation you gave at the end of the previous semester which was also on this topic. I am not certain whether these questions were answered then, however I 'll try to provide my own thoughts:
ReplyDelete1. I don't know what are the statistics on this. Most likely successful and wealthy companies that have the resources to spare for keeping an eye on this issue are probably doing that as it saves time and decreases complexity and therefore increases revenue in the end.
2. I am not quiet familiar to this topic and currently can't come up with my own idea, however the possible solutions that our colleagues stated in their answer seem logical to accept.
3. Unfortunately nothing more from what was already pointed out.
Hi. That's true - I was talking about that during summer presentation.
DeleteRegarding to yours first point. Business is business :) and if company can save money and get better forecasts by process optimisation - they probably will focus on that. Preparing process analysis does not have a sense if we can not transfer this effort to the money. You have right "it saves time and decreases complexity and therefore increases revenue in the end".
It was fresh input to discussion :)
I guess I don't have enough experience to fully answer these questions. Nevertheless, one thing definitely caught my attention. It is the question related to "translating" human know-how to process flow validator. In my opinion it will never be fully possible. As human knowledge and way of thinking is so unique and so illogical in terms of structure and associations that no machine, no programme, no application will ever be able to recreate it. Why google translator is a bad translator? Because it cannot fully recreate to process of translation, it cannot make use of context, cultural knowledge, associations. It translates directly a sentence from one language into another, choosing the first word matching the source word. To illustrate it with a good example I would like to "thank you all from the mountain" for reading my comment. You see what I mean? Associations , metaphores and machines... this will never work.
ReplyDeleteHello, an example: "thank you all from the mountain" made me laugh :). I have to agree with you, but from business perspective we should take care about process transparency. (I know that is not possible in every case but in most of them it is) When our process is not understandable people involved in its execution are trying to find workaround to simplify their work. It can be source of further errors.
DeleteIn my opinion there is a missing gap that could be fulfilled with software which can be able to learn what it should be checked during approval process. If document should be approved by few persons, I know that each of them are responsible for different check areas. Every stage could be candidate to add the most recurrent issues which can be standardized. Practice of adding standardization can be used to automate the process. The simplest example: check versioning, or check if all required fields in document was filled. Then we can use this as helper during approval process to minimize human error.
Delete- Do companies compare process model with its execution in production environment?
ReplyDeleteFrom the presented article I can deduce that the process executed in production environment can differ from the one that was modelled (or even it could be no model created). This fact can cause a lot of difficulties in a company as missing documents, inconsistencies or delays of some acceptances.
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
As it was mentioned in the paper, experts, thanks to process mining method were able to find and analyse process inconsistencies and errors. They looked at the process from the productive perspective, what let them fixing the most important problems occuring in the process.
Some changes in a process were required. They resulted with optimizations in an area of decision support and better acceptance of the process by decision makers.
- Do you know any methods for process tracking / analysis / optimization?
Sorry, but I know very little about process mining, so I am not able to describe any other method.
- Do companies compare process model with its execution in production environment?
ReplyDeleteThey tend to and fail miserably :) Unless, it is a factory. If you look at the software companies they have process models, special people to maintain them and still fail somehow.
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
We can manage every single step and still fail. Especially in so called agile environment. It's hard to win with client which is paying us. By doing that, we are probably breaking some rules and most validators will fail.
- Do you know any methods for process tracking / analysis / optimization?
Using process trackers, managing it. Trying to avoid being agile :) Or be flexible and have flexible validators. However, with flexible validators we can fail in managing process and track it. This is most likely to be a problem for project managers not me. We have a lot of methods for handling processes, making analysis and optimizations and still most companies fail. Even those big ones, like Microsoft or Google.
- Do companies compare process model with its execution in production environment?
ReplyDeleteI really don’t know… but I think they should. Why? Because it has important influence on company’s finances and efficiency. Well-created process model might protect company from negative consequences in business.
- How can we use experts knowledge in process analysis software?
This question regards to papers conclusion. It highlights expert knowledge usage. How to "translate" human know-how to process flow validator?
The whole process mining has to be supported by people because of their understanding of the relationship and situation in real-world processes. Experts can help in process modelling however there should be created algorithms capable of collaborating with them.
- Do you know any methods for process tracking / analysis / optimization?
No, I don’t know such methods. In company that I work for there is no dedicated process management. All processes are managed by managers of each department. Is it correct? Only experts should comment…
Do not know much about this area, it seems to me that there are a lot of problems, but our thinking is significantly different from the thinking machine. Intuitively, it reminds me of multivariate statistical analysis of the data, where it is always necessary initial knowledge of the field so as not to deal with the obvious facts.
ReplyDelete1) Article doesn't state that companies do that, so it's hard for me to tell if companies do it or not. But, basing on the article, it's quite important to do this kind of analysis, because it can improve your processes and it can find drawbacks of your system
ReplyDelete2) Experts knowledge can lead to system patch, experts know-how and their understanding of the events in the process can be helpful in findind and pointing problems
3) Unfortunately, I don't know any methods of process tracking/analysis/optimization. I don't want to try to be an expert in a field that is completely out of the scope of my interests.