The Challenge: Hundreds of Millions of Euros, Thousands of Suppliers, and Very Little Visibility
For a large, UK-based private equity organization operating across 22 offices worldwide, with numerous acquired companies folded into its structure, understanding how and where money is being spent is no small task.
The firm manages a vast and highly decentralized global supplier network, channeling hundreds of millions of euros annually across thousands of supplier relationships. But with spend data scattered across a complex portfolio of entities, getting a clear, consolidated picture of that activity had been persistently out of reach.
Without structured, categorized spend data, informed procurement decisions were difficult to make and future-focused analysis was harder still. The firm needed more than a snapshot. It needed a repeatable, scalable foundation for understanding its global spend.
The Approach: Intelligent Categorization at Scale
The engagement began with a substantial data challenge: 30,000 invoices covering an 18-month period from January 2023 to July 2024, each requiring accurate categorization to build a meaningful view of global spending.
Recognizing that a generic taxonomy would not do justice to the firm’s unique and complex spend structure, the team developed custom categories tailored specifically to its business. Seven iterations of spend analysis configurations were created and refined, each cycle sharpening the focus and improving the accuracy of the categorization framework.
To accelerate the process and to test the potential of machine learning in this context 25,000 invoices were manually categorized by the team. This substantial human-led effort served a dual purpose: it produced high-quality, structured data immediately, and it trained a machine learning algorithm to handle the remaining 5,000 invoices autonomously.
The machine learning model performed well, though as with any first deployment of this kind, some refinements were needed. Those early outputs were used to further train the algorithm, improving its accuracy and laying the groundwork for more reliable automated processing in the future.
The work also opened a clear path forward: integrating the spend analysis capability directly with the firm’s B2B systems, where supplier payments and spend data are already housed. That integration would enable real-time data access and significantly streamline ongoing analysis.
The Results: A Structured, Intelligent Foundation for Global Spend Management
The engagement delivered something the firm had not previously had: a comprehensive, structured overview of its global spend, one that reveals patterns, highlights allocation trends, and equips the finance and procurement teams to make better, faster decisions.
Three levels of spend categorization now give the business the flexibility to examine its spending at a high level or drill down into granular detail, depending on the question being asked. Spend trends that had previously been invisible are now visible and actionable.
Equally important is what has been built for the future. The machine learning model, now trained on a robust dataset and refined through real-world use, is ready to process new invoice data with greater speed and accuracy than was possible before. Each new cycle of data will make it sharper.
The firm leaves this engagement not just with answers to today’s questions, but with the infrastructure to keep asking better ones.
How SpendQube can help
If your organisation is struggling to understand where money is being spent across multiple business units, entities or systems, SpendQube’s spend analysis solution can help. Whether you’re dealing with patchy supplier data, inconsistent categorization, limited spend visibility or the challenge of analyzing large volumes of transactions, SpendQube combines AI-powered analytics with procurement expertise to create a clear, reliable view of spend. From data cleansing and classification through to actionable insights and ongoing support, we help organizations build the foundation for better procurement decisions, stronger governance and sustainable cost optimization.