Analyzing resource behavior using event data
Many activities in business processes are executed by human resources whose performance can affect process outcomes. Process data often includes information about employees who performed different steps in the process. Organizational mining is an area within process mining that focuses on discovering resource behavior patterns from event data. For example, organizational mining approaches can discover social networks and organizational models (and enrich process maps with discovered resource information), evaluate resource performance, identify suitable employees for various process tasks, and identify risky resource behaviors and opportunities for improvement.
The performance of employees in organizations is typically evaluated by managers who might be biased or may not have all information about tasks performed by employees. Process logs often contain accurate information about the types, frequency, duration, and outcomes of activities performed by employees. This information can be used to extract indicators of resource performance (that represent resource skills, utilization, preferences, productivity, and collaboration patterns) and learn how various resource behaviors affect process outcomes . For example, a manager may learn that some of her employees become the source of process bottlenecks when their workload reaches a certain level, or that fast execution of an activity by an employee often causes rework later in the process. It is also possible to derive aggregate performance scores (e.g., using Data Envelopment Analysis), which can be used to benchmark the performance of different teams or track performance progress over time .
To achieve optimal process outcomes, it is important to allocate to the various tasks the best available resources. A subset of organizational mining approaches focuses on optimal resource allocation. Such methods analyze historical process data and identify appropriate resources for a process step based on given criteria. Various methods aim to optimize a given process outcome (e.g., time, cost, or quality), while other methods consider several dimensions of process performance. When selecting the most adequate resources to perform a task, existing resource allocation methods can consider various factors ; for example, the time that it took different employees to complete the task, quality of execution (e.g., how often the task was repeated when performed by an employee), how well different employees work together, resource availability and workload. Based on the selected criteria, a resource allocation approach recommends a resource or a group of resources that should be assigned to a process task to achieve a desired process outcome. For example, for a task in an urgent case, a resource allocation method may assign an expert who produced timely high-quality outputs in the past (to maximize quality and to minimize time), while for the same task in a non-urgent low-impact case it can assign a junior employee (to minimize cost).
Understanding employee behavior and its effect on process outcomes enables timely process interventions (e.g., balancing resource workload or providing additional training) and pro-active management of process performance (by assigning appropriate resources to process activities).
Anastasiia Pika, PhD mindzie