6 Differences Between Traditional Process Analysis and Process Mining
If you’ve come across any process mining advocates, they’ll be sure to boast about the insights, the money saved with process mining, how it’s radically changed the Business Process Management (BPM) environment—the list goes on.
For those new to process mining, all of this has got to sound pretty intriguing. At least we hope so. However, amidst all the praise, it’s also important to show the actual differences that make process mining stand out as a top approach for process discovery and analysis when compared to other methods. Read on to see 6 comparisons between traditional process analysis and process mining, two methods with a common goal but different approaches to getting there.
A Few Definitions
There are a lot of definitions floating around so it’s best to first define some terminology.
When talking about traditional process analysis, I’m speaking about the phase of BPM where a business process is modeled and then analyzed.
Process Mining is the discovery, monitoring, and optimization of business processes.
Both traditional business process analysis and process mining perform process discovery which is the mapping out of processes in the way they are performed. The goal of each method is to analyze the process so that strategic changes can be made for process improvement.
- How Processes Get Discovered
- The Quality of the Information
- The Time Commitment
- Process Improvement Capabilities
- The Type of Model Created
- The User Interaction Data Integration
- Is Process Mining the Only Solution You Need?
1. How Processes Get Discovered
With traditional business process analysis, the process will typically be discovered by reaching out to all the process stakeholders. Whether by individual interviews, group workshops, employee shadowing or even sending out a questionnaire, the goal is to piece together the information from the stakeholders to create a process model. Another preferred method may be employee shadowing.
Today more and more processes leave a digital trace in the form of event logs. Process Mining gathers data from these event logs taken from a business’s systems or a data warehouse. The minimum data requirements needed to map a process are the activity name, a unique case ID, and a timestamp for each case. Once process mining software has the data requirements it uses sophisticated algorithms to automatically discover the process workflow and create an end-to-end process model that displays all activities, the paths between the activities, and the frequency of those paths. You can then compare your as-is model to an uploaded reference model for an instant comparison.
Another noteworthy fact—not only processes can be discovered and modeled. Business rules and organizational models can also be automatically discovered with advanced process mining software.
2. The Quality of the Information
One of the main challenges with interview and workshop methods is the accuracy of the information collected. Interviews and workshops rely on employees to remember every activity they’re involved in with perfect detail but that’s easier said than done. Human bias, disagreements between employees or even as much as one employee having an “off day” will all affect the accuracy of process discovery. Employee shadowing can cause the worker to feel pressure to “perform”.
In contrast, the process model created with process mining is transparent and accurate because it’s derived from fact-based data rather than subjective and often siloed employee knowledge. The quality of the insights makes for better decision-making and decreases risks when developing new business strategies.
3. The Time Commitment
Because process mining automatically discovers, maps, and analyzes business processes, the time spent is drastically reduced not just for process discovery but throughout the entire process management lifecycle. A standard process with an SAP system for instance can be discovered in just a few hours..
Collecting information through manual means is a more complicated matter. It can be difficult to get the proper time commitment from workers who need to juggle priorities and switch between activities at a moment’s notice. Stakeholders will all need to come to an agreement to verify the process. After that comes the modeling and analysis. These steps alone can take several weeks to complete. Then, a high-level analysis of the expected behavior of the reference model can take from a few weeks to several months. You haven’t even begun to monitor deviations and inefficiencies.
4. Process Improvement Capabilities
The fact that process mining automatically discovers and analyzes processes makes it the perfect tool for continuous process improvement. Any time you make changes to a process you can use the newly generated data to create an updated process map with new sets of insights to measure process improvement by comparing to previous process performances, checking conformance, and monitoring KPIs.
Advanced process mining capabilities go beyond measuring process improvement and can be applied as a solution for business process improvement strategy. Simulation engines offer a way to create and test what-if scenarios and analyze them before selecting the appropriate changes for process improvement. This includes testing the results of potential RPA in a process before adding in any automation. When a new process model is created after RPA has been implemented, you’ll see the performance of the bots and immediately see if they are working as expected.
Measuring process improvement gets a lot more complicated without process mining. If you’ve spent the months to manually discover a process, you’ll have to undergo the same ordeal every time you make changes to the process making it nearly impossible to compare the current process performance with past performances in a timely manner that won’t render the analysis outdated by the time you finish it.
5. The Type of Model Created
We’ve already covered the different ways a process workflow is discovered and then modeled. So, what are the qualities of the process models? Do the qualities of the process model change based on how the process is discovered?
The answer is yes.
Because all it takes for process mining to create a process model is a data upload, you can update your process model with the most current available data as often as necessary. With every new data upload, you can perform a process analysis to find new inefficiencies in the process and changes in process operations. The timing, accuracy, and comprehensiveness of these process models make them dynamic process models. They are interactive models that you can zoom in and out of to get a highly-defined or broader view depending on whatever level of context you want.
In contrast, static process models are the result of process discovery where the process steps are collected and evaluated from employee interviews. The procedures captured during interviews represent a specific timeframe. The static process model gets outdated quickly whether due to seasonality, new or updated policies, new employee onboarding, or changing business strategy.
In order to update a static process model with traditional process analysis techniques, you’d need to conduct a whole new round of interviews.
6. User Interaction Data Integration
Tasks like matching information between documents are essential components of a process activity. Leaving people-completed work out of the process model will leave holes throughout the process. Businesses without process mining or task mining will need to resort to the same traditional process analysis methods of employee shadowing and interviews and it’s up to the employees to explain how the manual tasks fit into the process.
Process mining combines with task mining, making it possible to have a truly end-to-end process model. Task Mining is the discovery, monitoring, and analysis of user interaction data on a desktop.
Task mining enriches the information gathered from system event logs by adding in all the steps users perform in the front-end applications or applications without an event log
Filling in the process model with user interaction data gives you an accurate view of all the necessary steps in a process and lets you see what goes on during the execution time of a business activity, how employees are completing their work, all the people that are involved, and how tasks affect other parts of the process.
Just like how process mining analyzes business data, task mining takes user interaction data and automatically performs an analysis to uncover inefficiencies related to time, reworks, and deviations for a better understanding of how a business activity is completed
And when it comes to automating manual tasks, task mining makes it easy to choose the best tasks to automate by showing which tasks cost the most and require the most resources.
Is Process Mining the Only Solution You Need?
After going through this list, it should be clear why process mining has gotten so many supporters despite it still being a new technique. This doesn’t mean, however, that we think you should immediately toss out your BPM strategies.
Process Mining is a tool for anyone who wants an incredibly fast, reliable and continuous way to discover, monitor, and optimize business processes. As mentioned in another article written previously, process mining is a great complement to other Business Intelligence (BI) tools where process mining gives a micro-level analysis of process behavior that fits nicely into the larger macro-level business operations view obtained with BI.
In fact, the process mining technique is complementary to all activities in your digital transformation. The analysis and deep-level insights make it easier to master operational excellence. From workflow to robotic process automation (RPA) initiatives, process mining accelerates any process improvement project you have in your roadmap.
Once you try process mining for yourself, you’ll be surprised at how much faster and more accurate your process discovery, analysis, and improvement will all become and you too will be advocating for process mining as an essential BPM tool.
Check out our other blog posts:
Stefano Pedrazzi has 15+ years of experience in the ICT and BPM field. He spent a significant part of his work life leading and managing Business Process Management and System Integration Projects for OT Consulting. Since 2017 he has been the VP of Sales & Marketing at myInvenio, helping his customers kick-start their Process Digital Transformation revolution using myInvenio’s process mining solution.