The procurement teams that consistently outperform on software cost have one thing in common: a clean, current view of vendor spend. The data is rarely missing — it sits across AP, P-card, marketplace, embedded-vendor billing and SaaS pass-through. The work is categorisation, classification and run-rate analytics. This guide walks through what to build and why each layer matters.
The single biggest blind spot in software spend management is the assumption that AP is the source of truth. AP captures roughly 60-75% of true software spend in a typical mid-size enterprise. The remainder hides across cloud marketplace billing (third-party software billed through AWS/Azure marketplace), P-card and expense (long-tail SaaS), embedded billing (software bundled into platform contracts), pass-through (software the systems integrator buys on the customer's behalf), and free-trial conversions that never made it through procurement.
The primary spine. Captures large, contracted vendor spend. Misses small-ticket SaaS, P-card spend and marketplace pass-through. Categorisation here is often inconsistent — same vendor billed under three different supplier IDs is the rule, not the exception.
AWS Marketplace, Azure Marketplace, GCP Marketplace. Third-party software billed through cloud invoices, frequently uncategorised at the cost-management layer. Often 8-15% of total software spend in a cloud-heavy estate.
SaaS subscriptions purchased by business users, charged to personal credit cards and expensed. Typically 3-8% of total software spend; the share is rising as SaaS sprawl grows.
Software the SI procures on the customer's behalf as part of a larger services engagement. Frequently re-marked at 12-25% above customer-direct pricing. Rarely visible in spend analytics.
Software bundled into platform contracts — Oracle Database inside an Oracle Cloud subscription, Microsoft Defender inside an E5 SKU. Hard to extract from the platform line item, but material for category-level run-rate tracking.
We design vendor-spend frameworks buyer-side. No platform commissions.
Categorisation is where most spend-analytics programmes fail. The classic mistake is to categorise by vendor name rather than by capability. A vendor-name taxonomy makes it impossible to compare alternatives or identify overlap; a capability taxonomy makes both straightforward.
The classification scheme we recommend is two-axis. Axis one — capability category: collaboration, productivity, infrastructure, security, data and analytics, developer tools, business applications, AI platform, low-code, monitoring, etc. Axis two — commercial vehicle: perpetual + support, subscription, consumption, marketplace pass-through, P-card SaaS. The cross-tabulation surfaces the most actionable views: same capability across multiple vendors (consolidation candidates), same commercial vehicle across many vendors (renewal-cycle aggregation), high-growth consumption categories (cloud commit optimization opportunities).
A practical guide to finding, categorising and optimising the long tail of SaaS and software spend.
Total annual spend is a useful summary metric and a poor decision metric. The decision metric is run-rate trajectory — what does this vendor cost the enterprise per quarter, and how is that number changing? The trajectory captures the renewal-rate inflation, the consumption-growth inflation, and the M&A growth that compound across budget cycles. Customers who manage spend by trajectory routinely capture 6-12% additional savings per cycle versus customers who manage by annual total.
We design analytics frameworks buyer-side. No platform commissions, no partnerships.
Weekly compliance intelligence for IT leaders.