Performance Analysis of the Procurement Process Using Process Mining

Analisis Kinerja Proses Pengadaan dengan Menggunakan Process Mining

Abstract

Procurement process performance plays an important role in supporting cost efficiency, material availability, and operational continuity. However, in Purchase-to-Pay processes, delays and corrective activities often appear not only in the final completion time, but also in the path followed by each case. This paper examines the BPI Challenge, Purchase to Pay event log using process mining, with a focus on completed cases in the 3 way matching invoice before goods receipt process. The analysis covers activity frequency, process variants, throughput time, bottleneck transitions, and rework. The dominant variant represented only 31.43% of cases, which means that most cases moved through paths outside the main variant. Completed cases had an average completion time of 58.17 days. The longest bottleneck occurred from “Record Invoice Receipt” to “Clear Invoice”, with an average duration of 28.51 days, pointing to invoice settlement as the main waiting point. Rework was most frequent in “Change Quantity” and “Change Price”, at 18.14% and 14.16%, respectively. Cases with repeated “Change Price” also had a higher average completion time, increasing from 63.86 to 87.47 days. These results show how process mining can locate waiting points and corrective activities that require attention in procurement improvement.