BPM: The value of process mining
Business process management (BPM) is a mature technology that has become a large market, projected to grow from about $11 billion in 2021 to $26 billion in 2028, according to Fortune Business Insights. Its roots date back to the 1980s, when document workflow was developed to expedite document review. Now, BPM covers a wide range of activities and has evolved to be primarily customer-centric rather than document-centric. An important part of digital transformation, BPM provides the ability to automate some or all steps in a process so that human intervention only takes place when needed.
As usage of BPM grew and multiple processes were added, many enterprises were no longer able to track their diverse set of processes. The complexity of enterprise applications, some new and some legacy, made it difficult to do so, as did the lack of process documentation. Visibility into processes was limited, especially for the ones that cut across different enterprise systems. Process mining emerged as a new class of enterprise software. “Process mining is at the intersection between data mining, data science, and process management,” said Marlon Dumas, co-founder of Apromore, which offers a process mining solution of the same name.
The importance of process mining
Process mining collects data from event logs and then visualizes it so process owners, functional managers, or C-level executives can see where their data is located and where the delays are. “Compliance managers want to see if any of the enterprise processes are violating regulations,” Dumas continued. “In other cases, customer support managers want to know why they are spending so much on resolving complaints.” There are many uses cases across a variety of functional areas.
The three verticals that are primary drivers for process mining are insurance and banking, manufacturing, and government. “BFSI has a very large number of use cases,” Dumas noted. “In manufacturing, one of the primary applications is for finding more efficiencies by eliminating re-work and dealing with defects, and in the government sector, many citizen-facing processes would benefit from improvement.”
In many industries, supply chain processes, both inbound and outbound, are of great concern. Awareness of this issue spiked during COVID, but was also gaining increased attention for competitive reasons, since time to delivery is of high importance to customers. “With process mining, organizations can know in advance what the effect will be if a delay occurs in supplying a certain component, and how it would affect all the other systems that are dependent on the timing of that component,” Dumas said.
Process mining allows organizations to increase operational performance by learning from their past experience. For organizations that have gained insights from process mining, the natural next step is to complement these insights with predictive capabilities. For example, Apromore uses historical data to estimate the time until completion of an ongoing process, which aids in planning at multiple levels. It also can predict customer actions such as complaints, returns, and payments. Anticipating outcomes in advance allows organizations to be agile in the face of changing circumstances. The highest level, when data analytics and AI are used to support continuous process improvement, is referred to as augmented BPM.
Improving performance
Substantial performance improvements can be achieved as a result of process mining. A retailer that had entered the ecommerce market 20 years ago struggled to keep up with customer demand after online purchases increased during Covid. In addition, as is often the case with process challenges, the retailer had not been able to identify the root cause of delays. Given the number of factors involved in the organization’s processes, including human resources, suppliers, delivery, and invoicing, the company was eager to figure out how to optimize their profit margins. Data from each of three information systems was extracted and analyzed.
The analysis produced a visualization of bottlenecks, flow issues, and workload imbalances. The company then made several data-driven operational changes. Service level agreement (SLA) fulfillment increased from 74% to 90%, cancellations (which often occur when the customer finds out that their delivery is postponed) decreased, and customer satisfaction increased.