Process Mining 101 – Misconceptions about Process Mining (Part 1)
in Learning

Process Mining 101 – Misconceptions about Process Mining (Part 1)

Have you heard of Process Mining before but are not sure how it works or if it has any value to you? Or are you completely new to the area of Process Mining and just arrived here by pure chance? Don’t worry. We’ve got you covered. In this five-part series we’ll give you a comprehensive guide to Process Mining. We’ll start with the basics and go into more details regarding application, use cases and possibilities along the way. The series ends with common misconceptions about Process Mining that cause confusion and hold people back from getting started.


Part 1: What is Process Mining

Part 2: Goals and Steps

Part 3: Use Cases

Part 4: Process Mining in Power BI

Part 5: Misconceptions about Process Mining (Part 1)

Part 6: Misconceptions about Process Mining (Part 2)


Misconceptions about Process Mining

There is lots of information floating around on the internet about Process Mining. And we get it. It is an interesting and exciting topic that deserves to be talked about. But there’s also a lot of information that is misleading and causes confusion among our customers and partners and people who are only starting to learn about Process Mining.

We collected the most common misconceptions and misunderstandings surrounding Process Mining to bring some clarity to the topic and set the record straight. This post is concerned with the most frequent technology-related misconceptions.

Content

Is Process Mining only for big companies and big data?

Is Process Mining only useful for some processes or IT systems?

That whole event log thing

Does Process Mining equal Process Science?

Is Process Mining the same as…?


Is Process Mining only for big companies and big data?


Assumption 1

“Process Mining only pays off with huge amounts of data”

What’s true

Process Mining is based on event data that is automatically visualized in a process flow. This automated feature makes it easy to deal with large amounts of data. But that doesn’t mean that you can’t use Process Mining when you’re processes are not fully automated, or you only have small amounts of data from simpler processes. Process Mining does not care about how complex your process is or how much data you can feed into the tool.

In Process Mining the aim is to uncover the ongoing as-is process from end-to-end. Just because that process is simpler doesn’t mean it’s clear how it goes. Actually, when you’re still having doubts, it is a good idea to start with a small Process Mining project and go bigger in follow-up projects. That’s because Process Mining is only one part of an optimization project. Strategy, planning, and implementation of optimization measures are other parts. And it’s easier to start with a complete optimization project on a smaller scale. Plus, you can always add data if you need to. That’s the beauty of Process Mining, instead of a one-time deal, you can use it to constantly monitor and improve your processes.

Assumption 2

“Only big companies can profit from Process Mining”

What’s true

Process Mining looks more impressive when you look at complex processes in huge enterprises. But smaller companies can and should also benefit from Process Mining. What it comes down to is an in-depth analysis regarding process performance, compliance and conformance issues and optimization potential. And no matter the size of a company, everyone can benefit from optimization and improvement strategies.

What’s also true, is that Process Mining was hidden behind expensive paywalls and was designed in an overly complicated way. So the entry barriers were relatively high for smaller companies. But this is changing, mostly due to our own democratization strategy. Process Mining now is affordable. There even is a free version of our solution available in Microsoft AppSource . It’s easy, because the report is pre-built for you and the application is based on an easy to navigate user interface. So there really is nothing holding smaller companies back any longer.


Is Process Mining only useful for some processes or IT systems?


Assumption

“Process Mining is only useful for BPM or ERP systems and only applicable to very specific processes”

What’s true

People often think that Process Mining can only be applied to fully automated processes, meaning processes that are completely controlled by an IT system. But that is not true. In the modern business world, almost every process leaves behind traces in the system. And Process Mining can use those traces. So as long as the data can be accessed somehow, Process Mining is possible.

The main difference here is that one between controlling a process and supporting a process. Think of it that way: technology has changed the way we communicate but it is only supporting us with communication tools. Instead of writing a letter on paper and sending it by post, you can write an email, or in a less formal situation use some kind of chat to talk to people. It is easy to see the traces that sort of communication leaves behind. And it’s the same with business processes. In processes that are supported by IT structures, data is a by-product. But you can still use it in Process Mining.

In fact, the value of Process Mining for those cases can be much higher than that for fully automated processes. In less controlled processes the possibilities to work flexibly are much higher and as such there usually is a lack of awareness of how people actually work. Process Mining uncovers precisely those hidden processes.

This is not to say that automated processes are unsuited for Process Mining. On the contrary. For automated processes based on rigid process models Process Mining can reveal the weak points of the model.

The thing about Process Mining is, that it can work with any process as long as there are a CaseID, Timestamp and Activity Name located somewhere in the IT system. Speaking of which…


That whole event log thing


Assumption 1

“Process Mining collects and transforms all the necessary data automatically."

Or the polar opposite:

“Before you can start with Process Mining, you first have to gather, clear and transform huge amounts of data."

What’s true

Process Mining needs very little datapoints. What’s important is the format. Process Mining works with event log data. That data can come from various sources, including databases, csv files, spreadsheets, logs (e.g. transaction or message) or ERP systems. The minimum datapoints are a caseID, timestamp and activity name.

A caseID is a unique identifier for every individual case in a process. So, for example, an order ID.

A case is one route through your process and is made up of a series of events, hence the name event log.

A timestamp simply tells you when something happened.

An activity name tells you what happened. For example, an order placement.

The Process Mining tool does not search your system for appropriate data, you need to feed it into the tool. But that is not as difficult as it sounds. Also, you definitely have appropriate data stored somewhere. That’s where your or our analysts help you.

Sometimes, you need to transform data into the right format. But that also is not as difficult and time consuming as people often think. And the nice part is, after you feed the data into the Process Mining tool, the visualization of the process flow and other visuals, as well as the creation of report pages happens automatically.

So to make it short: neither is data loading in Process Mining fully automated nor is it overly complicated. It’s just that you need the right data in the right format.

a man confused by the term 'event log'
Many people are confused and intimidated by the term 'Event Log'. But it is only the data that is required for Process Mining: a collection of events, categorized by three data points.

Assumption 2

“You can just use a database or other tools such as Excel to analyze your processes, there is no real need for Process Mining."

What’s true

That is often said by people who are intimidated by the event log. But here’s the thing: theoretically you might be able to use the data that you would feed into a Process Mining tool and do a manual analysis. However, that would take much longer and cause much higher costs than any Process Mining project. Plus, the results are not necessarily complete or well-presented.

The strength of Process Mining lies in its ability to show you how things are currently happening, including all outliers. Gathering that insight through any other method may not be impossible but it would take forever. And then you would also have to bring it in a format that allows you to share and discuss your insights with others. Process Mining does all of these things for you – fast, objectively and reliably.

And what’s more, you don’t need to know what you’re looking for because Process Mining will reveal those areas that should be optimized. You can explore your processes and look at various different perspectives.

So isn’t it better to take the limited time that’s needed to create an event log instead of sifting through your data yourself?


Does Process Mining equal Process Science?


Assumption

“Process Mining is Process Science, a scientific approach to discover and optimize processes."

process mining is not process science
It’s easy to mix up Process Mining and Process Science. Process Mining is a sub-category of Process Science. It is used for analytics and to develop optimization strategies.

What’s true

Process Science is a collection of disciplines that combines IT with management sciences. The major goal of Process Science is to improve processes in terms of time, costs, quality, speed, flexibility, and reliability.

This sounds like Process Mining, but Process Science comes with many more tools and areas. In fact, the area Process Mining belongs to is Business Process Management which in turn is only one subset of disciplines in Process Science.

Process Mining fits in on the analysis side of Process Science. It uses event data to visualize business processes and enables detailed analysis of those processes. In this way, Process Mining works with recorded behavior in contrast to perceived behavior.


Is Process Mining the same as…?


Assumption 1

“Process Mining and Business Intelligence tools do the same things”

What’s true

On the surface BI and Process Mining look similar. Both are for analyzing data in the greater area of business management. Both use visualizations to make it easier to analyze data.

But BI traditionally focuses on defining and monitoring isolated KPIs whereas Process Mining looks at processes from end-to-end and helps to find root causes of rework, waste and bottlenecks.

The main difference is that BI assumes that you know the business process, while Process Mining recognizes the fact that no matter how well a process is planned, things always go differently than planned.

That means, Process Mining helps you to find the greatest optimization potential, to define priorities, and to monitor those KPIs that really matter.

In that sense, Process Mining and BI complement each other and both are stronger together than separate.

With second generation Process Mining the boarders between BI, data mining and process analysis get even more blurred. BI tools now also utilize state-of-the-art Process Mining features, but in general the complimentary nature of both tools remains.

That’s also due to a different set of data that is used in Process Mining and BI. Process Mining uses event log data with three obligatory data points: caseID, timestamp and activity name. Process Mining is incredibly strong when it comes to that data format, but BI is stronger in other formats. However, the specific dataset of Process Mining is what makes end-to-end visibility, conformance and compliance checks and root cause analysis possible.

Assumption 2

“Process Mining is just Data Mining”

What’s true

They share the mining aspect, but Process Mining has emerged as a subset of business process management and the research and development efforts surrounding that area.

Most Data Mining techniques extract abstract patterns in the form of, for example, rules or decision trees. In contrast, Process Mining creates complete process models, and then uses them to precisely highlight where the bottlenecks are.

So Process Mining has a different perspective and analysis approach. However, it is similar to Data Mining in the way it uses data to generate visualizations and helps to gain insights.

There are also some Data Mining techniques that look at processes, but they lack the end-to-end perspective of Process Mining

Unlike Data Mining, Process Mining focuses on the process perspective: It looks at a single process execution as a sequence of activities that have been performed.

Process Mining is all about understanding the current ‘as-is’ processes. The IT systems record very detailed information about which activities are performed, when, and by whom. By leveraging these log data, fact-based models can be generated that show the actual process behavior from various angles.

a man, dizzy from all the data science different terms
Process Mining is often used together with other techniques from the BPM landscape. Sometimes those techniques get mixed up or are used interchangably. But they all have their own use cases and application areas.

Assumption 3

“Process Mining is the same as Process Modelling”

What’s true

Both Process Modelling and Process Mining will give you a picture of a process. But they are used to achieve completely different things. In a Process Modeling tool, the perceived process is described and documented based on interviews and workshops. This helps to create a shared understanding of the processes in an organization. These models also help to train and educate new hires on how things are supposed to be done. But the problem with those models is, that processes are constantly changing so a process model can quickly become outdated and redundant.

Process Mining, on the other hand, uses the actual data behind your business processes to create a process flow of your as-is process. So where Process Modelling shows an ideal, Process Mining shows reality. Following from that, Process Mining also has a different purpose: analysis, discovery, improvement.

The picture you get from Process Mining also looks much more complex than any straightforward Process Modelling image. But that is the power of Process Mining. It is objective and captures all the details, so you understand your process.

You can even use your most common process variant as your new ideal model. Or you may find that even though there are some variants, that they are as valid as your process model whereas there are some other variants that pose a real problem and require action.

Assumption 4

“Process Mining is Robotic Process Automation”

What’s true

You can trigger workflows from within Process Mining applications. Workflow automation and Robotic Process Automation (RPA) are slightly different things.

A workflow is made up of individual steps similar to the process flow and its individual activities. A workflow has a start and an end point, also very much like a business process. You can, in fact, think of a process as a set of workflows coming together.

When you do workflow automation, you automate the flow of tasks, or documents, or information in a specific predefined order across several work activities. Automation of a particular workflow process can include a simple request / approval process or several different workflow tasks that are triggered by established business rules.

For example a multi-level budget request review is delegated automatically to the right person on each level, depending on the outcome (approved or rejected) and the amount of money.

In more detail: An employee submits a purchase request. This triggers an automated workflow. The request is forwarded to a manager. If the manager agrees, the request is forwarded to the next approval level, depending on set rules ( For example, a request that is below 20.000 $ is forwarded to a high-level finance manager. Everything above 20.000 $ is forwarded directly to the CFO). If the request passes all approval levels, it is again automatically forwarded to the relevant purchase officer.

Here, the individual work activities are carried out by human workers. But the request itself is not send by humans but moves along the workflow automatically.

Robotic Process Automation

In contrast RPA automates individual work activities. Are there highly repetitive, rule-based tasks in your workflow? Those are excellent for RPA. The robots in RPA are not physical objects, but virtual bots that act as a digital employee who can perform rule-based tasks.

For example, bots can pull relevant data from a document and enter it into a database. This might require the use of OCR (optical character recognition) software to translate documents into data. This can also work when scanning in non-digital documents.

With RPA human workers are freed from repetitive tasks, so that they can focus on the more individual and exciting parts of the work life.

So, what does that mean for Process Mining?

  • Process Mining helps you to find automation potential for RPA
  • Process Mining tools allow you to trigger predefined workflows to act on a problem (e.g. sending alerts to the right person if a KPI is passed, the follow up workflow deals with solving the problem)
  • Process Mining enables strategic automation in your company for both RPA and workflows

In the next (and last) part of our series we will talk more about misconceptions about Process Mining and its effects on businesses, employees, and strategy.

If you have any questions or want to learn more about Process Mining, try out our demo version, contact us, get the free visual, or check our website. We are happy to help you, answer your questions and tell you more about our mission.