Mining Gold: 6 Steps to Kickstart Your Process Mining Project

You’re looking to make a positive impact on your business, and you have been charged with improving business processes through the transformative power of Process Mining. With myriads of Fortune 500 businesses claiming millions in value realization, and Gartner just recognizing Process Mining as a distinct software category in 2023, your head might be spinning. Beginning a Process Mining program, like any large transformational effort, can seem like a gargantuan and overwhelming experience. In front of you lies a host of challenges – technical, functional, and even political – that make success uncertain and difficult to reach. However, if executed well, the benefits of a Process Mining program can be far-reaching and long-lasting for a team and a business ready to embrace the challenge.  

1. Align to Your Strategic Objectives

Even before a program kicks off, it is important to understand the landscape into which process mining is being introduced. As with every software introduced into an enterprise’s environment, there are several factors to consider such as the technologies and source systems in play. Perhaps most importantly, though, the enterprise needs to ensure alignment to its strategic priorities and ongoing initiatives. Why is Process Mining being introduced? Are there objectives around sustainability? Working capital? Supply chain resilience?  

Without alignment to the strategic objectives of the enterprise, almost any process intelligence initiative will fail. Even if a Process Mining tool or project team can identify process bottlenecks or provide insights into opportunities for improvement, the business itself is required to act to realize value. In some rare occasions, the business is eager to leverage the power of Process Mining technologies. More often than not, though, the thought of yet another tool to work in, another process to manage, and another team to “collaborate” with is not a compelling offer. If, however, the objectives of the program can align to the business’ objectives, and if some high-level goals can be established, business leadership can and should be convinced by the prospect of using data to reach their goals. 

2. Pick Your Process 

With your strategic objectives established and business leadership onboard, it’s time to get a little more tactical. With traditional, case-centric process mining, a process needs to be mined before value can be realized. The decision should be informed by a few pieces of information.  

Which process best fits my strategic objectives? 

If an organization wants to use process mining to improve customer satisfaction, maybe back-office finance is not the best place to start. Identifying which processes align with larger business objectives is often a good place to start. 

Where do I have the most business leadership support? 

This flows logically from the points above. To make any meaningful progress improving any process, the business needs to buy in and have the appetite to make change for the greater good. Without the necessary support of a process owner, a Process Mining program is likely to drag on underdelivering with poor or middling results. 

Where can I shorten my time to value realization? 

To shorten the time to value realization, there are several factors to consider. Several process mining vendors have ready-made connectors or transformation scripts to accelerate the ETL process. With more custom or company-specific processes and source systems, though, there is often more time that needs to be spent engineering a process model before insights can be drawn.  

Some processes run through multiple source systems. In those cases, the extraction and transformation of source system data may need to happen several times – increasing the amount of time that needs to be spent engineering.  

On-hand system expertise is another factor that might impact time to value. For non-standard processes and systems, this is a little more obvious. Even for standard systems, though, there is often some level of customization that happens that may need to be accounted for. Without quick and easy access to source system expertise, the time to value will lengthen. 

Do I have enough case volume? 

Infamously frustrating, and time-consuming processes at your company may seem like easy targets for Process Mining. That having been said, though, if a process is only executed a few hundred or thousand times a year, the value from mining a process may not be worth the effort. Sometimes traditionally process modeling and mapping might be more applicable.  

3. Build the Data Pipeline 

Now that strategic objectives have been established, the business is onboard, and process owners are eager (or reluctantly agreeable) to improve what they own, the actual implementation work can begin.  

The construction and maintenance of the data pipeline is different across many platforms. Some have pre-built connectors for various applications to extract and transform data automatically. Some Process Mining tools are limited to the ingestion of an event log and a case table where others allow for the creation of more complex data models. Some platforms require the extraction of data into the cloud or an on-premise system whereas others can be embedded into the architecture of a data warehouse. Some allow continuous ingestion of new data while others require intermittent file uploads. 

All these considerations will impact the development and management of a data pipeline. Therefore, each different Process Mining suite will need to have a tailored approach. Nevertheless, the data will need to be extracted and transformed into a format compatible with the Process Mining solution of choice. This means that the Case ID should be defined, Activity Steps need to be mapped, and timestamps identified. With the data pipeline built and other relevant details sorted, we are finally ready to see our process! 

4. Choose a Use Case 

Not to be confused with “Pick your Process,” each process can have many opportunities for improvement. Consider the Procurement process. Within Procurement, there are many opportunities for improvement – Reducing Maverick Buying, Canceling Fewer Orders, Vendor Consolidation, Late Receipt of Goods, Increasing Automation, etc. These can all be tantalizing opportunities. In choosing the right one, there are several factors to consider. 

Where does the business/process owner feel pain? 

Particularly if business support for a process mining initiative is tenuous, there might be an opportunity to provide direct support to a process owner by attacking a pain point they feel in their day-to-day work.  

Can the process mining tool (or accompanying AI) provide a recommendation? 

Some Process Mining technologies have native AI capabilities or pre-built analytics that can help analysts pinpoint particular opportunities for improvement. 

Are there common use cases within the process area? 

Maverick buying is a notable example of a common use case within Procurement that is often ripe for opportunity. If there are other process mining customers that have leveraged a technology to their benefit, it is worthwhile considering what has been done to realize value rather than reinventing the wheel. 

Does the data point to a particular use case? 

Process Mining tools are built to help you leverage your data to identify opportunities for improvement. Look at the data and see if there are any obvious areas for improvement. You might identify a root cause to a process breakdown you may never have considered. You may find that an unnecessary and unwanted activity is a common occurrence just by doing a little bit of data analysis. 

What is the time to value realization? 

When looking to get started, sometimes it is best to get the first win under your belt – even if the value realization potential is a bit smaller. This can help the business develop an appetite for larger projects in the future – rather than starting off with a use case that is costly, time consuming, and overly ambitious. 

5. Stay Focused and Maintain Discipline

Presented with a wealth of data and endless opportunity, it is easy to get distracted or bite off more than you can chew. Particularly early on, it’s important to maintain good project discipline and avoid unnecessary scope creep. Take care of your fundamentals. Validate your data. See your first use case through. You can always tackle something larger and more ambitious later in your Process Mining campaign. 

Getting that first use case across the line can set the foundation for a successful program. Kevin Cunningham, Lead Value Engineer from Celonis points out, “We've frequently found that these process intelligence initiatives can become self-funding. The ROI of the pioneer project becomes the funding for a more robust operating model." Especially in a situation where a Process Mining program needs a reason to justify its existence, a quick win early on that generates some additional funding is a great start. 

6. Tell Your Story and Go Again 

Congrats! You’ve completed your first use case. You realized value for your business, improved processes, and you’ve shown your team what can be accomplished. Now it’s time to tell your story to a wider audience. Tell your executive team. Share your success with a new area of the business. Build up some enthusiasm from your future customers. One day these potential stakeholders and customers will be coming to you for help realizing their business objectives and solving their operational problems. And there’s no need to wait to help them! It’s time to move onto your next use case!