Spend analysis is the process of collecting, cleansing and classifying a company's purchasing data to show what it buys, from whom, at what price and through which channel. It is the first stage of source-to-pay: sourcing, supplier consolidation and negotiations all rest on it.
The work follows five steps:
- Extract spend from every source: AP ledger, purchase orders, card statements, expense reports
- Cleanse the supplier list, merging duplicates and name variants
- Classify each line in a category taxonomy
- Enrich with contract, price and payment-term data
- Analyse by category, supplier, entity and channel
Classification takes most of the effort, and the chart of accounts does not do it. A GL account records the accounting nature of an expense, not what was bought: the same laptop can land in an asset or an expense account. Spend analysis needs a purchasing taxonomy, and public ones exist: UNSPSC, a four-level hierarchy (segment, family, class, commodity) with two digits per level, and the EU Common Procurement Vocabulary, a nine-digit code used by public buyers, whose last digit is a check digit.
E-invoices could carry that classification line by line, but rarely do. In Peppol BIS Billing 3.0, the item classification (BT-158) is optional, even though its scheme list includes UNSPSC and CPV. The buyer still classifies.
Spend analysis differs from budget vs actual analysis, which compares each cost centre with its plan; spend analysis cuts across cost centres, by supplier and category. It is also how tail spend and maverick spend are measured.
Phacet can feed it with clean, traceable data. The agent that labels supplier transactions categorises invoice lines, the agent that cleans the supplier database detects duplicate records, and Phacet Analyze builds dashboards on verified tables, with no answer without a source.