Category management is widely recognised as one of the most effective approaches to strategic procurement. By organising spend into logical groupings and managing each category with a dedicated strategy, procurement teams can optimise sourcing, consolidate volume, negotiate better terms, and manage supplier relationships more effectively.
But category management has a prerequisite that many organisations underestimate: accurate spend classification. Without reliable classification, categories are blurry, analysis is unreliable, and strategies are built on shaky foundations.
Spend classification is not glamorous work. It does not make headlines in procurement conferences. But it is the single most important enabler of effective category management — and getting it wrong undermines everything that follows.
What Spend Classification Means
Spend classification is the process of assigning every procurement transaction to a standardised category within a defined taxonomy. The taxonomy provides a hierarchical structure — typically three to five levels — that organises all goods and services into progressively more specific groupings.
For example:
- Level 1: IT and Telecommunications
- Level 2: Software
- Level 3: Enterprise Applications
- Level 4: ERP Software
- Level 5: Oracle Cloud Licences
The most common standard taxonomy is UNSPSC (United Nations Standard Products and Services Code), but many organisations adapt this or use industry-specific alternatives. What matters most is not which taxonomy you use, but that it is applied consistently across all transactions.
Why Classification Is So Difficult
If classification is so important, why do so many organisations get it wrong? Several factors contribute.
Volume and Variety
A mid-size enterprise might process 200,000 procurement transactions per year, involving thousands of different products and services from thousands of different suppliers. Each transaction has a description, a supplier name, and possibly a commodity code — but these fields are inconsistently populated and wildly variable in format.
One transaction reads "HP LaserJet Pro M404n Printer." Another reads "Printer — laser — HP." A third reads "IT equipment per PO 47821." All three are the same category of purchase, but they look nothing alike in the data.
Manual Classification Limitations
When organisations attempt to classify spend manually, the results are predictably inconsistent. Different analysts classify the same items differently. Obscure or ambiguous items are miscategorised or left unclassified. The process is slow, typically taking months for a full classification exercise, by which time new transactions have added another layer of unclassified spend.
ERP Data Quality
Oracle Fusion Cloud captures commodity codes and category information, but these fields are often:
- Left blank because requisitioners skip non-mandatory fields
- Filled with incorrect values because the pick list is confusing or incomplete
- Inconsistent because different users interpret categories differently
- Outdated because the taxonomy has not been maintained as the organisation has evolved
Supplier Name Inconsistency
Effective classification requires understanding who you are buying from, which means normalising supplier names. "Telstra," "Telstra Corporation," "Telstra Corp Ltd," and "Telstra — Melbourne office" are all the same supplier, but they appear as four different entities in the data.
How AI Transforms Classification
This is where artificial intelligence genuinely shines. AI-powered spend classification addresses every challenge listed above, processing high volumes of messy, inconsistent data with speed and accuracy that manual methods cannot match.
EVA from Sharpe Project Consulting (SPC3) incorporates AI classification models trained specifically on procurement data. These models:
- Interpret natural language descriptions to understand what was purchased, regardless of how the description is formatted
- Normalise supplier names by identifying duplicate and variant records for the same vendor
- Learn from corrections — when a human analyst reclassifies an item, the model incorporates that feedback to improve future accuracy
- Process at scale — millions of transactions can be classified in hours rather than months
- Maintain consistency — the same item is always classified the same way, regardless of which business unit purchased it or how it was described
The result is a classified spend dataset that is comprehensive, consistent, and reliable — the foundation that category management requires.
From Classification to Category Strategy
With accurately classified spend, category management becomes dramatically more effective. Here is how classification powers each element of category strategy.
Total Category Visibility
Classification reveals the true size and scope of each category across the entire organisation. A category manager who thought their IT consulting spend was $3 million discovers it is actually $8 million when purchases from all business units are consolidated under a consistent classification. This changes the sourcing strategy entirely — the volume justifies a more structured approach, competitive sourcing events, and dedicated contract negotiations.
Supplier Landscape Analysis
Within each classified category, you can see exactly which suppliers are being used, how spend is distributed among them, and where fragmentation exists. This analysis typically reveals opportunities to consolidate suppliers, increase volume leverage with preferred vendors, and eliminate redundancy.
Price Benchmarking
Comparing prices within a consistently classified category becomes meaningful. You can identify price variance — situations where different business units are paying different prices for the same goods or services — and address it through contract consolidation or policy changes.
Demand Pattern Analysis
Classification enables analysis of purchasing patterns within each category. When are purchases typically made? How do volumes fluctuate seasonally? Are there recurring demand spikes that could be anticipated and managed more strategically?
Contract Strategy
Understanding the full scope of spend in a category informs contract strategy. Should this be a single-supplier agreement or a multi-supplier framework? What volume commitments can be made? What terms will deliver the best total value?
Implementing Effective Classification
Step 1: Define Your Taxonomy
Start by selecting or designing a taxonomy that fits your organisation. UNSPSC is a good starting point, but you may need to customise it to reflect your specific industry and purchasing patterns. Keep it deep enough to be useful but not so granular that maintenance becomes impractical. Three to four levels is sufficient for most organisations.
Step 2: Clean Your Data
Before classification, address the most critical data quality issues. Normalise supplier names to consolidate duplicate records. Standardise units of measure and currency. Remove or flag test transactions, cancelled orders, and other noise.
Step 3: Apply AI Classification
Deploy an AI-powered classification engine to process your transactional data against your defined taxonomy. EVA handles this automatically for Oracle Fusion data, applying trained models that achieve high accuracy on initial classification while continuously improving through feedback and learning.
Step 4: Review and Refine
AI classification is highly accurate but not perfect. Establish a review process where category experts validate the classification of high-value or ambiguous transactions. Feed corrections back into the AI model to improve future accuracy.
Step 5: Maintain Continuously
Classification is not a one-time project. New transactions flow in continuously, and they need to be classified as they arrive. AI-powered tools automate this ongoing classification, ensuring your category view remains current without manual effort.
The Downstream Impact
The ripple effects of accurate spend classification extend well beyond category management:
- Executive reporting becomes more meaningful when spend is organised into consistent, understandable categories
- Budgeting and forecasting improve when historical spend is accurately categorised and can be projected forward
- Compliance monitoring is possible only when spend can be matched against contracts and policies at the category level
- Supplier performance management gains depth when performance can be evaluated within the context of specific categories
SPC3's consulting services help organisations design classification frameworks, implement AI-powered tools, and build the ongoing governance processes that maintain classification quality over time.
Build on Solid Foundations
Category management is only as good as the data it rests on. If your spend is unclassified, inconsistently classified, or classified using an outdated taxonomy, your category strategies are built on unreliable foundations.
Investing in accurate, AI-powered spend classification is one of the highest-ROI activities in procurement. It unlocks better category strategies, stronger sourcing outcomes, improved compliance, and more credible executive reporting.
Get in touch with SPC3 to learn how EVA can transform your spend classification and power more effective category management across your Oracle Fusion Cloud procurement data.