Improving spend data quality to enhance procurement's decision making - a case study

Abstract

Purchase-to-pay process is one of the organization’s backbone as it enables purchasing different services and products to generate revenue. A lot of beneficial information is conducted in different steps of the process and one of the most important data set created is the spend data that reflects what has been purchased over time from which sup-pliers. This data consists of purchase order and invoice data. However, as the data amount has risen, there is a data quality issue. The challenge with vast data amounts is the poor quality that impacts direct to decisions that have been made upon poor data. Therefore, is critical to constantly assess the available data and see if it fits the purposes and targeted data quality level.

The aim of this thesis was to study how the spend data quality could be improved in the purchase-to-pay process for better decision making looking at data reliability, timeliness, completeness and accuracy. This thesis objective was to research what is the data quality in case organization’s purchase order and invoice data and to compare them in price-quantity data. In addition, the thesis researched how different data was generated in various steps of purchase-to-pay process and how it affects decision making. The research methods used in this study was a case study and it had both qualitative and quantitative empirical parts and theoretical section. The quantitative research included analyzing purchase order and invoice data with document analysis method. There were 457 invoices analyzed and 301 purchase orders. In addition, qualitative data was collected by interviewing 23 different specialists, managers and leaders in the case organization in two separate interview rounds. The theoretical part of this thesis was based on scientific literature, though it was scarce on spend data and therefore this study aimed to build the gap between previous academic research and current literature.

Results of this study pointed out that there is a lot of room to improve invoice and pur-chase order data sets to achieve high quality spend data. This thesis results displayed that purchase orders and invoices commonly lack of quantity, item description and price. Furthermore, the study showed that invoice data sets are more complete and therefore should be utilized to have improved quality in price-quantity data. In addition, there were lot of challenges in the purchase order data, which were divided into three different categories: people related issues, process related issues and challenges that stem from too many alternatives. When spend data quality is good, people would also rely on the data and consequently, utilize the data more for various decision making situations and create a positive loop of data iteration. This would have a direct impact for improvements in many areas, such as creating savings, utilizing compliance vendors, improving due diligence and investing correctly.