Procurement analytics is designed to provide data-driven solutions to supply management problems. Analytics leverages the organizational capabilities in data management to provide intelligent support for procurement decision-making. Analytics facilitates data collection from several source systems, classifying data to specific taxonomies and visualizing KPIs on dashboards or within business-intelligence tools.
In recent years, such tools and dashboards have become the engines of procurement analytics capabilities. Today, purchasing groups at many organizations are using descriptive analytics (analyzing and summarizing past data) successfully for spend analysis, category management and contract management. Some supply managers have looked to build on those successes by exploring advanced predictive and prescriptive analytics models for decision-making support.
Many research articles have posited that purchasing groups will more extensively use predictive analytics (using past data to predict future behavior) to reap bigger benefits. However, there were few examples of successful implementations of predictive procurement analytics that could be featured and discussed at a November 2019 pipeline meeting of CAPS Research, the Tempe, Arizona-based program in strategic partnership with Institute for Supply Management® and Arizona State University. As a result, attendees identified the current state of advanced analytics in the industry as one of the research topics to study. The research objective was to document successful use cases of advanced-analytics deployment and integration into mission-critical decision-making processes.
Applying Advanced Analytics: Beyond the Ordinary, a report by CAPS Research, examined issues related to successful adoption and implementation of procurement analytics. The report is based on literature review, survey results and interviews with procurement managers from a range of industries.
How Organizations Use Advanced Analytics
In our initial survey of procurement managers, we asked them to self-assess
their analytics use and capabilities. Thirty-four percent of the respondents indicated they use advanced analytics (27 percent using predictive analytics, 5 percent using prescriptive analytics and 2 percent using automation). Sixty percent of respondents use descriptive analytics for data summaries, reports and data visualization; 6 percent said they are developing robust procedures for data collection, cleaning, integration, governance and retrieval.
Respondents were appreciative of the benefits of descriptive analytics through dashboards and pivot tables, but they were also aware of the need for model-driven predictions and recommendations for future courses of action. Further research with a subset of the survey respondents revealed that only about 20 percent of those interviewees had developed and deployed advanced-analytics solutions for procurement processes.
Interestingly, advanced-analytics tools such as robotic process automation (RPA), natural language processing (NLP) and artificial intelligence (AI) are already used by a majority of the survey respondents. However, the use of these tools is limited to the cleansing, aggregation and management of data. Given the explosive amount and variety of data generated and stored, organizations have adopted increasingly sophisticated and complex methods to preprocess the data to suit data-analytics requirements.
While data preprocessing is a foundational requirement for analytics, it rarely translates to direct decisional support. Model-based advanced-analytics
applications will still need to be developed using the preprocessed data sources. Nevertheless, the use of RPA, NLP and AI is a positive trend in the profession. Familiarity with these technologies in the beginning stages of the procurement-analytics journey can help create organizational support for future use of advanced-analytics models in decision-making processes.
Our research indicates organizations are frequently employing advanced technologies like RPA, NLP and AI to facilitate data consolidation and digitization of unstructured data (for example, contracts). Our interviews further revealed that they are putting advanced analytics to better use by consolidating data from different sources and eliminating data discrepancies.
In-House Resources Contribute to Success
Advanced-analytics models are in use for a variety of procurement applications, including (1) predicting product prices based on past data, (2) predicting supplier risk based on financial and production data, as well as external factors such as weather and social media mentions, and (3) optimizing operational and logistics processes tied to procurement decisions.
Organizations that successfully implement advanced analytics have often benefitted from in-house data-analytics teams; 89 percent of respondents indicated that their supply management division had at least one data-analytics specialist. This result is consistent with our interview findings. Our research shows that improving operational efficiency is one of the primary reasons for utilizing advanced analytics; many use cases have also resulted in cost savings and competitive advantage.
One of the main factors affecting the success of analytics applications is the organization’s capability to handle data from multiple sources. Procurement data often resides in centralized ERP systems. Many organizations have invested in third-party systems that extract and transform data from ERP systems to store them in forms that are most suitable for analytics models. From our interviews, we conclude that organizations need to take a strategic approach that lays out a roadmap through progressively maturing stages of procurement analytics data architecture. Organizations that are further along in their data-architecture roadmap have a greater potential to implement advanced analytics.
Successful implementation of advanced analytics in procurement decisions depends on several factors, including availability of data, analytics talent, organizational buy-in and executive sponsorship.