The 5 steps of Spend Analysis
Objective: Master the 5-step methodology to run a complete, reliable Spend Analysis, from data collection to actionable recommendations.
Step 1 - Data collection
The quality of the Spend Analysis depends directly on the quality and completeness of the data collected.
- ERP sources: SAP, Oracle, order and invoice data - the main source (80-90% of the data)
- Purchasing cards (P-cards): Often forgotten but represent 5-15% of spend, especially in indirect procurement
- Contracts: Contract data (negotiated prices, committed volumes, payment terms)
- Expense reports: Travel, entertainment - an often untapped reservoir
- Off-system purchases: Orders placed directly by operations without a purchase order
⚠️ Warning: On average, 20 to 30% of a company's spend bypasses the procurement system (maverick spend). Failing to include it significantly distorts the analysis.
Step 2 - Data cleansing
The most time-consuming step, but the most critical. Dirty data produces false analyses.
- Deduplication: The same supplier can appear under 5 different names (IBM, IBM France, IBM Corp, International Business Machines, I.B.M.)
- Standardisation: Harmonise units, currencies, date formats
- Enrichment: Fill in missing data (category code, country, legal entity)
- Validation: Identify and handle outliers (a EUR 1M invoice for office supplies = probable error)
Step 3 - Categorisation
Group spend into homogeneous categories to enable analysis and comparison.
- Standard taxonomy: UNSPSC (United Nations Standard Products and Services Code) - 4 levels, usable internationally
- Custom taxonomy: Tailored to the sector and organisation, more relevant but less comparable
- Practical rule: 3 to 4 levels of depth, 15 to 25 level-1 categories, 80 to 150 sub-categories
Step 4 - Analysis
Apply analytical techniques to extract the insights.
- Pareto analysis: Identify the 20% of categories/suppliers that account for 80% of spend
- Trend analysis: Year-on-year evolution (N-1, N-2, N-3) per category and per supplier
- Anomaly detection: Aberrant unit prices, inconsistent volumes, duplicate payments
- Internal benchmarking: Compare prices across sites, business units and countries for the same category
Step 5 - Actionable insights
Turn the analysis into concrete recommendations with a quantified impact.
🏢 Typical opportunity table
| Opportunity | Category | Current spend | Estimated savings | Action |
|---|---|---|---|---|
| Supplier consolidation | Office supplies | EUR 2.4M | EUR 360K (15%) | Go from 45 to 3 suppliers |
| Contract renegotiation | Transport | EUR 8.1M | EUR 650K (8%) | Grouped multi-site tender |
| Standardisation | Packaging | EUR 5.2M | EUR 780K (15%) | Reduce references from 120 to 35 |
| Maverick elimination | IT software | EUR 3.7M | EUR 555K (15%) | Centralise SaaS purchases |
💡 Key takeaway: The 5-step methodology (Collection, Cleansing, Categorisation, Analysis, Insights) is sequential. The classic mistake is skipping steps 2 and 3 to jump straight to analysis, which produces unreliable results and flawed recommendations.