What Is Spend Analysis and Why Does It Matter?

August 24, 2026

What Is Spend Analysis and Why Does It Matter?

Monday, 24 August 2026

Spend analysis is the process of collecting, cleaning, classifying and examining an organisation’s expenditure data to work out what it actually buys, from whom, and how often. It matters because organisations cannot negotiate, consolidate or control spending they cannot see.

Key takeaways

Most of the work is cleaning and classifying, not analysing. The insight is easy once the data is trustworthy.
Spend analysis reveals behaviour, not just cost. Australian National Audit Office analysis found 26 per cent of short term Commonwealth contracts start in June, the final month of the financial year.
It also reveals commitments that grew quietly. Of Commonwealth contracts amended in value, 49 per cent at least doubled.
Supplier names are unreliable identifiers. The ANAO used Australian Business Numbers instead, because the same supplier appears under multiple names.

What is spend analysis?

Start with an example. A finance team is asked a simple question: how much do we spend with our largest supplier? Three weeks later the answer is somewhere between four and seven million dollars, because that supplier trades under two names, one division codes them to consulting and another to professional services, and nobody is certain whether the licensing renewals belong to the same entity. The organisation is not short of data. It is short of a version of the data that can answer a question.

Spend analysis exists to close that gap. The Chartered Institute of Procurement & Supply (CIPS) defines it as the process of collecting, classifying and analysing expenditure data, applied to historical spend to answer questions about visibility, compliance and control. The output is not a report so much as a reliable picture: what was bought, from whom, by which part of the organisation, under what arrangement and at what price.

Two things follow. The first is that spend analysis is descriptive rather than prescriptive. It tells you what happened, and the decisions about what to do next belong to category strategy, sourcing and contract management. The second is that most of the effort goes into preparation rather than insight. Once the data is clean and consistently classified, the findings are often obvious within an afternoon. Getting to that point is the job.

Why is spend analysis harder than it sounds?

The obstacles are unglamorous and they are the reason most attempts stall. Consider the identification problem alone. Preparing its analysis of a decade of Commonwealth procurement, the Australian National Audit Office (ANAO) noted that supplier names self-reported by entities varied enough that the same supplier appeared under multiple names, making supplier name unreliable as the only identifier. It used the Australian Business Number instead, while noting that some suppliers are structured across several ABNs and others are exempt from having one.

Then there is classification. Australian Government entities code each contract to a United Nations Standard Products and Services Code, which the Department of Finance uses to sort spending into goods, services and higher level categories. A standard taxonomy helps, though it only works to the extent the person entering the contract picks the right code. In organisations without a standard, the general ledger becomes the default classification, and general ledgers are built for financial reporting rather than for understanding markets. Accuracy is the third obstacle, and the most consequential, because analysis built on unreliable records produces confident conclusions that happen to be wrong.

What does spend data reveal that nobody meant to disclose?

This is where spend analysis becomes interesting, because expenditure records behaviour rather than intention. Analysing 824,178 Commonwealth contracts worth $564.5 billion over ten years, the ANAO found that while July is the most common month to start a contract, 26 per cent of all short term contracts began in June, the final month of the financial year. Nobody writes a policy encouraging end of year spending. The pattern is visible anyway, because thousands of individually reasonable decisions leave a shape in the data.

The same analysis surfaced something arguably more significant about how commitments grow. Where a contract had been amended to increase its value, 49 per cent of those contracts had their value increased by at least 100 per cent, meaning the commitment at least doubled from what was originally approved. Contract amendments accounted for roughly a third of all committed value over the decade, rising from $4 billion in 2012 and 2013 to $28 billion in 2021 and 2022. Anyone looking only at original contract values would be reading less than half the story.

These findings are useful beyond government because the mechanisms are ordinary. Budgets create timing incentives, approved values drift upward through variations that individually seem minor, and spend concentrates on a small number of suppliers without anyone deciding that it should. None of that is visible in a single transaction. It is only visible in aggregate, which is precisely what spend analysis assembles.

What can go wrong with the underlying data?

Reliability deserves particular attention, because it determines whether the analysis is worth acting on. Examining a sample of 155 contract notices, the ANAO found that only 26 per cent had all the basic contract details correctly specified and were reported within the mandated timeframe. Reported contract values matched the actual contract in 65 per cent of cases, start dates in 51 per cent and end dates in 53 per cent. That audit is now some years old and reporting practice has improved since, with Finance advising that it has reviewed and corrected more than 360,000 contract notices since 2018.

The transferable lesson is about how much confidence to place in a first extract. Duplicate records, variations recorded as new commitments, and arrangements split across several entries all inflate or fragment the picture in ways that survive into the analysis unless someone checks. A practical habit is to validate any striking finding against a handful of source documents before it appears in a paper. If the number is real it will hold, and if it is an artefact you will find out privately rather than in a meeting.

What do you actually do with the results?

Analysis is only worth its cost if it changes a decision. The joint CIPS and NIGP practice standard on spend analysis is direct about the intended uses: reducing costs through informed sourcing strategies, eliminating duplicate suppliers, improving contract compliance and using contract pricing to create savings. Each of those is a specific action rather than a general aspiration.

The most common findings tend to repeat across organisations. Fragmented spend, where one category is scattered across more suppliers than the market requires. Off contract buying, where purchases sit outside an arrangement negotiated precisely to cover them. Price variation, where different parts of the organisation pay different rates for the same thing. Concentration, where more depends on one supplier than anyone realised. Tail spend, where a large share of transactions accounts for a small share of value and consumes a disproportionate amount of administrative effort.

The genuine difficulty is not producing the analysis. It is the conversation afterwards, when the numbers say a business unit has been buying outside the arrangement for two years and that unit has good reasons, or thinks it does. Turning a finding into a changed decision requires evidence people trust and the ability to hold a position under challenge. That is what AcademyGlobal (AG) built its Procurement Analysis in the Public Sector training around, working with the spent data participants actually have rather than tidy sample sets.

The test of a spend analysis is not whether it produced a dashboard. It is whether anyone bought differently because of it.

Frequently asked questions

What data do you need for a spend analysis?

At minimum, accounts payable transaction data covering supplier, amount, date and the part of the organisation that bought. Purchase order and contract data make the analysis considerably stronger, since they show what was committed rather than only what was paid. Card transactions matter too, because they often hold the fragmented spend nobody has looked at.

How often should you run a spend analysis?

At least annually for the full picture, with more frequent refreshes for categories under active management. Data degrades continuously as suppliers change names, merge and are recorded, so a one off exercise ages faster than people expect. Building the cleansing into a repeatable process matters more than the frequency you choose.

What is the difference between spend analysis and spend analytics?

Spend analysis is the underlying process of collecting, cleansing, classifying and examining expenditure data. Spend analytics usually describes the tooling and more advanced techniques applied on top, including automated classification and predictive modelling. The distinction matters less than data quality, since sophisticated analysis of unreliable data is still unreliable.

Why does supplier name matching cause so many problems?

Because names are entered by people. The same company appears with and without its legal suffix, under a trading name, with typographical errors, or as separate subsidiaries. The ANAO addressed this by matching on Australian Business Numbers rather than names, and any serious analysis needs an equivalent unique identifier.

Can spend analysis identify savings on its own?

It identifies opportunities rather than savings. Finding that a category is fragmented across twelve suppliers indicates a possible consolidation, but whether that produces value depends on market structure, switching costs and how the sourcing is run. Treating an opportunity as a booked saving is a common way for reported figures to lose credibility.