Presentation of a research paper at the 35th International Conference on Advanced Information Systems Engineering (CAiSE) in Zaragoza Spain. The paper presents a classification of causes of waiting times in business processes and a method to automatically detect and quantify the presence of each of these causes in a business process recorded in an event log.

1. Why am I Waiting?
Data-Driven Analysis of Waiting Times
in Business Processes
Katsiaryna Lashkevich, Fredrik Milani,
David Chapela-Campa, Ihar Suvorau and Marlon Dumas
35th International Conference on Advanced Information
Systems Engineering (CAiSE ’23)

2. 2
Waiting time visualization using process mining
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Introduction
Process mining
IT
systems
Event
logs
Process mining
techniques
Process
insights
Images: Flaticon.com

3. 3
Research questions
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
How can these waiting time causes be automatically
discovered from an event log?
What are the direct causes of waiting time between activity
instances in a process?
How can we effectively describe the contribution of each
waiting time cause to the temporal efficiency of a process?
1
2
3

4. 4
Why am I waiting?
Batching
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Resource contention
Prioritization
Resource unavailability
(off-duty)
Extraneous factors
Enabled
activity
instance
Images: Flaticon.com

5. 5
Overview of the proposed approach
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

6. 6
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Causal directly-follows relations
between activities
Approach
Concurrent
activity pairs
Heuristic
concurrency
oracle

7. 7
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Causal graph of a case
Activity instance transitions:
Register invoice -> Notify acceptance
Register invoice -> Post invoice
Post invoice -> Pay invoice

8. 8
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Causal graph of a case
Activity instance transitions:
Register invoice -> Notify acceptance
Register invoice -> Post invoice
Post invoice -> Pay invoice

9. 9
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

10. 10
Waiting time due to batching.
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Lashkevich, K., Milani, F., Chapela-Campa, D., Dumas, M.: Data-driven analysis of batch
processing inefficiencies in business processes. In: RCIS. pp. 231–247. Springer (2022)
Batching discovery technique

11. 11
Approach
Waiting time due to resource contention and due to prioritization.
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

12. 12
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Waiting time due to resource unavailability and due to extraneous factors.
López-Pintado, O., Dumas, M.: Business process simulation with differentiated resources:
Does it make a difference? In: BPM. pp. 361–378. Springer (2022)
Working Calendar mining technique

13. 13
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Output: Waiting time causes per transition.
Source activity Target activity Total waiting time Case frequency Total frequency
Post invoice Pay invoice 100 h 100 % 1 000
100 h
8 h 30 h
45 h
5 h 12 h

14. 14
Cycle time efficiency (CTE) = PT / (PT + WT)
Impact of waiting time causes = CTE if a particular waiting time is eliminated
Metrics:
1. impact of each waiting time cause on the process CTE,
2. impact of each transition on the process CTE,
3. impact of each waiting time cause in each transition on the process CTE
Approach
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

15. 15
Evaluation
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.
Waiting time causes in the manufacturing process.
Process map of the real-life manufacturing process.
Total waiting time: 20 yrs 7 mths 27 d
Log includes 225 cases with start and end timestamps, and resources.
11 yrs 2 mths 28 days
5 yrs 2 mths 29 days
1 yr 6 mths 13 days
10 mths 11 days
1 yrs 9 mths 10 days

16. 16
Potential CTE improvement per waiting time cause in the manufacturing process.
Evaluation
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

17. 17
Waiting time causes in activity transitions of the manufacturing process.
Evaluation
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

18. 3
2
1
18
What are the direct causes of waiting time between activity instances in a
process?
• batching, resource contention, prioritization, resource unavailability, and
extraneous factors
How can these waiting time causes be automatically discovered from an
event log?
• with the developed technique that decomposes the waiting time into
direct causes from the activity instance log with the enabled, start, and
end times, and resources
How can we effectively describe the contribution of each waiting time cause
to the temporal efficiency of a process?
• by identifying the impact of waiting time causes on the temporal
performance using CTE
Tool implementation:
http://kronos.cloud.ut.ee/
https://github.com/AutomatedProcessImprovement/waiting-time-analysis/
Summary
"Why Am I Waiting? Data-Driven Analysis of Waiting Times in Business Processes" by Lashkevich et al.

19. 19
Thank you!
Marlon Dumas
marlon.dumas@ut.ee