Task mining and process mining are transformative technologies that help leaders understand how work gets done. They are similar, but not the same. An easy way to remember the difference is that task mining focuses on frontend data and process mining focuses on the backend. Task mining is used to understand exactly what people do on their computers. Process mining is more useful for understanding the high level steps that make up business processes.
Task Mining Overview
Task mining works by collecting data from the clicks and keystrokes that users perform. This is typically done by applying computer vision to screenshots, querying the operating system, or some combination of both. Task mining then uses machine learning to discover structure in the data it collects and generate insights around how users interact with their computers. These insights include visualizations and analytics that can be used for automation, optimization, and other types of transformation.
One of the key advantages of task mining is that data does not need to be in a specific format. In fact, existing data is not necessary at all. This means that task mining can be used to analyze work regardless of the application it takes place in. For example, if processing an invoice requires Outlook, SAP, legacy applications, and Excel, task mining can accurately capture all process details.
Any time people are doing work on their computer, task mining can be applied. Use cases include process discovery, process definition, and process analysis. The most common applications of task mining are automation and optimization. Task mining enhances automation efforts by automatically identifying and prioritizing automation opportunities and generating detailed process maps. Similarly, task mining guides optimization efforts by calculating machine learning assisted metrics and identifying steps in a process that are ripe for improvement.
Process Mining Overview
While task mining focuses on the interactions between a user and their computer, process mining focuses on backend data in the form of event logs. These system generated logs must contain a case ID, activity, and timestamp to qualify as an input for process mining. Process mining tools connect directly to the system(s) of interest to provide real time analytics and visualizations regarding process performance.
Process mining is commonly used for following the path of a transaction throughout a process. A common example is the Order-to-Cash process. In this example, the event log would be made up of the following:
- Case ID: order number
- Activity: step that the order is in
- Timestamp: time when an order moves from one step to the next
With process mining, millions of these transactions can be analyzed and compared with the click of a button.
Process mining is great for understanding end-to-end processes within and across large enterprise systems. Common applications include system migrations, identification of optimization opportunities, and conformance checking (i.e., comparing the actual performance of a process to the expected performance).
Which is right for my business?
Because task mining and process mining are similar, there is confusion surrounding which approach is best for a given situation. Both can be powerful and transformative technologies when applied strategically and intentionally. Neither approach is fundamentally better–both have strengths that make them better options for specific applications. In some cases, it even makes sense to use them together with a hybrid approach.
Automation: Task mining and process mining vendors both focus heavily on automation when marketing their products, but the reality is that task mining has the advantage when it comes to identifying opportunities and accelerating the development of automation. If the goal is to automate activities that are currently conducted manually, granular detail on manual work is critical. This is where task mining shines. Process mining does not provide the level of detail needed to identify and document automation opportunities. Additionally, business processes don’t always take place exclusively in systems that generate event logs, which limits the applicability of process mining. In the previously mentioned invoice processing example that spans Outlook, SAP, legacy systems, and Excel, only task mining can capture the entire process at a detailed level.
System Migration: Large scale system migrations (e.g., migrating to SAP S/4HANA) helped launch process mining onto the enterprise technology scene. Historically, these migrations take countless hours and cost exorbitant amounts due to the manual mapping that is required to document as-is processes. Process mining eliminates a lot of the manual work involved in these migrations through automatic analysis of system data (i.e., event logs). Process mining continues to be a great starting point for migrations, but doesn’t tell the whole story–parts of a process that are not executed within an ERP, CRM, or other type of event log producing system will not be captured. When process mining was first introduced for migrations, manual process mapping was used to fill in these gaps. With recent technological advances, task mining can now be used to capture the detail that process mining misses.
Optimization and Analytics: Task mining and process mining are both capable of generating meaningful analytics and identifying optimization opportunities. The best approach here depends on what type of insight is desired. Task mining is best suited for generating insights regarding performance from the employee perspective. For example, task mining can generate metrics on average handle time and variability for a given process across multiple employees or teams. Another common use of task mining is analysis of how much time is spent in various applications, which can drive efforts focused on eliminating process bottlenecks. Process mining is well suited for calculating metrics focused on high-level process performance. For example, what percentage of orders follow path A vs path B, and how often do orders get stuck somewhere in the process and never make it through? Similar to task mining, the metrics that process mining generates can be used to drive process improvements.
Both task mining and process mining have their place in the modern enterprise. The key to remember is that task mining focuses on the process from a user’s perspective, and process mining focuses on the process from the system’s perspective. As with choosing any technology, the best approach is to define your business objectives and work backwards from there.