Spend analysis collects, cleanses, categorises and examines all third-party expenditure to establish who you buy from, what you buy, how much you spend and where opportunity sits. It is the prerequisite for every other procurement improvement, because nothing else can be prioritised without it.
Companies running a first spend analysis after years of unmanaged buying consistently discover supplier counts several times higher than internal estimates, with the majority holding no contract and never having been competitively evaluated.
Nobody was careless. Spend accumulates quietly across business units, expense systems and purchasing cards, and no single report ever displayed the whole picture.
The four steps of spend analysis
| Step | Work involved | Common obstacle |
|---|---|---|
| 1. Collect | Pull transactions from every purchasing channel | Sources nobody remembered existed |
| 2. Cleanse | Deduplicate suppliers, normalise names, fix coding | The same vendor under several spellings |
| 3. Categorise | Map every transaction to a spend taxonomy | Generic ledger codes hiding real categories |
| 4. Analyse | Rank by category and supplier, identify opportunity | Producing a report instead of a plan |
Expect the data to be poor on the first pass. Vendor names appear inconsistently, entities merge, ledger codes group unrelated purchases under headings chosen for accounting convenience rather than category logic.
Cleansing is most of the work and it is not optional, because categorisation built on dirty data produces conclusions nobody trusts and therefore nobody acts on.
Six findings that surface almost every time
| Finding | Action it points to |
|---|---|
| Supplier count far above estimate | Consolidation and deactivation programme |
| Duplicate supplier records | Merge and check for duplicate payments |
| Categories never competitively sourced | Prioritise for sourcing events |
| Significant off-contract spend | Compliance enforcement, easier catalogue access |
| Same item bought at different prices across units | Volume consolidation |
| Dormant suppliers still active in the system | Deactivate |
The fifth is usually the most immediately actionable, because it requires no market engagement at all. Two sites paying different prices for an identical item can be aligned to the lower price within weeks.
The data traps that distort the first analysis
Four issues recur often enough to check for deliberately before drawing conclusions.
Intercompany transactions inflate apparent third-party spend. Payments between entities in the same group appear as supplier spend but are not addressable, and including them overstates the opportunity.
Pass-through costs distort category totals. Freight, duties and taxes billed through a supplier appear under that supplier’s category rather than separately, making some categories look larger and more sourceable than they are.
One-time capital purchases skew annual figures. A single equipment purchase can push a category into the top ranking even though it will not recur, wasting sourcing effort on spend that no longer exists.
Payments to non-suppliers pollute the count. Tax authorities, regulated utilities, insurers and statutory bodies appear as vendors but cannot be sourced, inflating both the supplier count and the perceived opportunity.
Turning findings into a plan
Rank categories on two variables: annual spend, and time since the last competitive test.
Categories that are large and untested form the first sourcing wave. Categories that are large and recently sourced need compliance monitoring rather than another event. Small fragmented categories need process controls rather than sourcing at all, because the cost of an event exceeds the available saving.
That ranking is the entire value of the exercise. A spend analysis producing a report nobody acts on has cost money and delivered nothing, and this is the most common outcome.
The analytical infrastructure supporting this is covered in connected procurement platforms.
Cadence and ownership
Run a full analysis annually with quarterly refreshes.
Spend patterns drift, new suppliers appear, and compliance in previously sourced categories erodes. Organisations treating it as a one-off project rebuild the same picture from scratch three years later, having lost the intervening opportunity entirely.
Ownership matters as much as cadence. Where the analysis belongs to a project team that disbands, the refresh never happens. Where it belongs to a named role with the reporting built into a regular cycle, it persists.
Prioritisation output feeds directly into supply chain planning.
Supplier concentration revealed by the analysis is examined in managing supply chain risk.
Freight and logistics categories often surface as opportunities, covered in reducing transportation cost in logistics.
For continuing spend management coverage, see our ongoing procurement analytics reporting.
Choosing a taxonomy
The categorisation structure determines what the analysis can tell you, and changing it later means redoing the work.
Build it around how the market is organised rather than how your ledger is organised. Suppliers compete within market categories, and a taxonomy grouping spend by accounting treatment rather than by what is actually being bought cannot support sourcing decisions.
Three levels is usually sufficient: a broad family, a category, and a sub-category granular enough that a single sourcing event could address it. Deeper structures look thorough and become impossible to maintain.
Presenting the results
The audience for a spend analysis is rarely procurement, and the presentation should reflect that.
Finance wants the addressable figure and the expected recovery. Budget holders want to know which of their suppliers will be affected and when. Executives want the headline opportunity and the resource required to capture it.
A single deck attempting all three usually satisfies none. Producing the analysis once and framing it three ways takes an additional afternoon and materially improves the chance that anything follows.
What to do in the first week after
Momentum decays quickly once the analysis is delivered, so the first actions should require no approval.
Deactivate the dormant supplier records. Merge the duplicates. Circulate the list of categories where identical items are bought at different prices, with the lower price attached. None of these needs a business case, and together they demonstrate that the exercise produces action rather than documentation.
The larger sourcing programme can then be proposed against a track record rather than a promise, which materially changes how it is received.
A note on tooling
Dedicated analytics software helps at scale and is unnecessary for a first exercise. A competent analyst with a spreadsheet can categorise several thousand suppliers in a fortnight, which is sufficient to identify the opportunity and justify the tooling later if the volume warrants it.
Frequently asked questions
What is spend analysis?
The process of collecting, cleansing, categorising and analysing all third-party expenditure to establish who the organisation buys from, what it buys, how much it spends and where savings opportunity sits. It is the foundation for category strategy, sourcing prioritisation and supplier consolidation.
What data sources are needed?
Every channel through which money reaches suppliers: the core finance system for order-based spend, payables for all invoiced spend including non-order purchases, purchasing card systems for direct staff purchases, and expense management for employee-submitted vendor payments. Omitting any one produces a materially incomplete picture.
Why is data cleansing the hardest part?
Because the same supplier appears under multiple spellings, legal entities, acquired names and abbreviations, while ledger coding often groups unrelated purchases under generic headings. Until those are resolved, category totals are wrong and the conclusions drawn from them cannot be defended.
How often should spend analysis be run?
A full analysis annually with quarterly refreshes. Spend patterns shift, new suppliers enter, and compliance in previously sourced categories erodes over time. Treating it as a one-time project means rebuilding the same picture years later having missed the opportunity in between.
