Process Mining

What is process mining?

Apps on our mobile devices have become so ubiquitous that very often we don’t even think twice before we press on the nice little “Install” button in the corner. Let’s say you get a discount on a popular ride sharing application that everyone has been raving about recently. After you install the app, you find out that you need to confirm your ID before you can start using it for ride sharing. There is a conveniently placed button that says, “Confirm your ID”, so you decide to click it. You follow the few simple steps that the app takes you through, and voila! You get a confirmation on your screen that says that all your steps have been completed and that the confirmation of your ID is now pending with the app’s review team – something that they claim will be completed in 24 hours or less. “Not totally unexpected”, you think to yourself as you close the app and go on with your day. A couple of days later you decide to fire up your app to get a ride to take you home after a night out with your friends. Frustratingly, the app still says, “ID confirmation pending”. You close the app and decide to use another app that you have been using for the last couple of years. You also uninstall the new app because you are displeased with their service. What you don’t know is that the promotional campaign that gave you the discount was such a hit that it choked the process over at the app’s office. But you don’t care about that, and frankly, in a world where an alternative app is always available, you shouldn’t have to!

What we saw now was a classic case of a process that is loaded beyond its capacity. John, the manager at the app’s ID confirmation team had set up a robust and reliable team of individuals to enforcethe ID confirmation process. They performed well for months together, the senior management is happy with how well they are performing, and he trusts them to handle slightly-higher-than-usual traffic that he expects from the promo offer that has just been rolled out. However, in the increasingly fast paced environments that businesses find themselves in today, it is often not quite enough to plan in the indefinite. You need to be precise with your plans and you need to be prepared for all types of eventualities. You should be able to quantify how robust your process is to external shocks. So, what could John have done to be better prepared? He could have used historical pricing data to estimate the price elasticity of demand. That would have given him a much better estimate of the increase in traffic than “slightly-higher-than-usual”


However, knowing how much of an increase in traffic to expect would not solve his problem. In fact, it would only tell him in definite terms, what the extent of the problem is. Now that he knows what the extent of his problem is, he is in a much better position to go about preparing for it. The next step could be map the process itself to look for redundancies or bottlenecks and to identify potential bottlenecks in the likelihood of a surge in demand. However, you can do a lot better than simply look at a diagram you have created, marvel at the beauty of it and to make sweeping generalizations based on your ‘intuition’ or ‘gut feeling’. What you really need is to make decisions that are backed up by data. At the very least, you must identify the time taken by each activity in the process, variability in those times, the number of people working on each of those activities and the costs involved, in tandem with constraints such as shift timings, application systems and processing capabilities. There are a host of other parameters you could choose to investigate, if you feel the need to have tighter control over the process. However, one must keep in mind that more is not always better, and in this case, there may be a level of detail beyond which the returns do not justify the costs. What John has on his hands here is an archetypal process mining problem – combining model-based process analysis and data-oriented analysis techniques can help him ensure the process is always under control. Let us take a closer look at process mining and its capabilities.

Capabilities

Process Mining is an approach to analyzing business processes where the objective is tounderstand, track and improve real processes by developing inferences from event logs and problem-specific data. It isn’t about simply presenting key data related to processes; it goes one step further byidentifying the context or active relationships of the processes andpresenting them graphically in order to detect problems and proposeimprovements. The key philosophyis to detect or diagnose problems based on facts and not on speculation, hypothesis or intuition. By combining event data and process models, it will be possible to ensure compliance, detect aberrations, forecast delays, support decision making and propose process redesigns. It is not uncommon to find out from event data that the process in reality is quite different from the

We need to understand that embarking on the process mining journey will not be a trivial exercise. There will be cross-functional teams, multiple applications (sometimes even from different vendors), databases and software systems working involved, and as all businesses know, that is an environment that makes consensus difficult, as is the case with all digital transformation efforts. A typical process mining project makes use of event log data from different information systems from around the business information landscape that is processed using process mining techniques to develop a model of the actual process, which one may find operates differently from how it was intended to run. You may then make use of time, cost and resource parameters to simulate or optimize the process by identifying bottlenecks and potential opportunities for improvement.

The Market
ResearchandMarkets in the study “Process Analytics Market by Process Mining Type (Process Discovery, Process Conformance & Process Enhancement) — Global Forecast to 2023”, explains that the global process analytics market is expected to grow rapidly from $185.3 million in 2018 to $1,421.7 million by 2023 – that is a Compound Annual Growth Rate (CAGR) of 50.3% over the upcoming five year period. According to their analysis, businesses pursuing Digital Transformation is a major driverof users analyzing and understanding business processes. With data analysis making inroads into all aspects of business, it was only a matter of time before businesses started approaching their problems at a system level or process level. Looking at process mining and its proliferation across geographical regions, there are two things that stand out:

Europe is currently the largest market size, partly due to the fact that most major Process Mining software vendors operate out of this region.
North America has shown the highest rate of growth in recent years, representing the most reliable opportunity in the process mining market

The major vendors of process mining software are given in the image below.

The market leader in process mining in Celonis and since 2015, the software giant SAP has had Celonis Process Mining on its price list, which means that they sell Celonis products around the world, making SAP a global reseller for Celonis. Disco by Fluxicon is the most popular of the independent tools in the market. There are also tools that made its way to the market through the university environment because of students’ research projects. One such example is the free solution ProM which was developed in the Eindhoven University of Technology.

SAP Process Mining by Celonis
SAP process mining by Celonis is the process mining capability provided by SAP within its own landscape. It allows you to take advantage of SAP HANA in tandem with the process mining capabilities of Celonis to help you ensure that key business processes are running at maximum efficiency and in conformance with their intended designs. It also provides you with detailed visualizations with drill-down capabilities, giving you a transparent view into the nitty-gritties of your processes and what you could do to potentially improve them. You could also run simulations to find out what impact changes to the process or parameters of the process can have on your process. The major advantages are:

On-premise and cloud deployment
Complete transparency into processes execution
Process efficiency improvement
Non-compliant processes detection

With SAP process mining, you get automated process discovery (APD) capabilities whereby, you can use event long data from your SAP solutions and third-party applications to understand how your process actually works and what the different process variants are. Here, you are not telling the application what the process is, you are asking it to find out how the process is operating as-is. This will tell you where the bottlenecks are, what inefficiencies are present in your process and can help you identify compliance problems . With further analysis, you will be in a position to pinpoint opportunities for process improvement

With SAP process mining, you also get the flexibility of two different deployment options –cloud and on premise . Depending on the level of security and ease of integration you desire, you can choose between these options. Integration with the existing SAP landscape comes with its own set of advantages such as ease of use and near-instant process mining availability. Either way, SAP process mining provides big data capabilities, allowing you to analyze large volumes of event logs. Root-cause analysis is also possible thanks to the ability to drill down on specific process steps. But perhaps the coolest feature is the ability to annotate diagrams and charts with key performance indicators and then check performance against those KPIs. The consulting, support and services that come with the software giant’s suite of offerings means that the environment has everything that anorganization could wish for.


Conclusion
With process mining, John can be better prepared for the next time the company decides to enforce promotional offers. He will be in a much better position to plan for the demand that comes in, and the changes he needs to make to his process or team. He will be less likely to be caught off guard, with simulated outcomes from process variants at his disposal to help him make his decisions. He can hire new resources or adjust working hours to ensure that the demand will be met. With data analysis poised to change the world as we know it (if it isn’t doing so already), it is only a matter of time before the data craze spills into systems and processes as well, making process mining as ubiquitous as data mining. Being an early mover can certainly be to the advantage of all organizations, as the power of process mining has become evident, and whether you like it or not, it is here to stay!

Leave a Reply

Your email address will not be published. Required fields are marked *

Prev
Current IT Trend in Data Visualization

Current IT Trend in Data Visualization

CURRENT IT TREND IN DATA VISUALIZATION Data: the single most important asset

Next
SAP HANA and R integration

SAP HANA and R integration

Introduction After promising much without any significant breakthroughs for many

You May Also Like