Can you explain where your IT budget goes?

Ask most public bodies what they spend on IT and you will get a number. Ask what drives that number, service by service, and the answer is usually much less certain. Which contracts, licences and suppliers make up the total? Which of them are growing? Which are paying for things nobody uses any more?

The National Audit Office (NAO) has made the same point about government services more widely. In its September 2025 report, Improving government’s productivity through better cost information, it found that departments often lack data at the level of detail needed to analyse the costs of specific business processes. That, the NAO says, limits their ability to identify inefficiencies, prioritise areas for cost reduction and measure improvement. It stresses that understanding costs is “not just about knowing the total of the cost base”.

The report looks at the cost of public services as a whole, not IT specifically. But the lesson carries straight across to IT budgets. So the practical question is simple: could you explain, with evidence, where your IT money goes and why?

What the NAO found

The report is an initial assessment rather than a full audit of cost data across government, and the NAO says so. Even so, its findings are relevant to anyone who owns an IT budget:

• Totals are known, detail is not. The NAO says most government organisations generally have a good understanding of their total costs. The gap is in the granular detail: some departments do not know the costs of individual services, nor where excess costs arise from staff working around old systems and poor-quality data.
• Without cost drivers, you cannot prioritise. The NAO found that poor data often leaves departments unaware of the underlying cost drivers. Without detailed cost insights, it says, government cannot prioritise the most expensive or inefficient areas, or assess the return on investment in digital transformation.
• Top-down models hide the problem. Costing models were more typically based on a top-down allocation of overheads using assumptions, rather than bottom-up analysis. Attempts to calculate cost per transaction often left out some operational costs.
• Good practice is more granular. The NAO points to activity-based costing approaches that can reveal duplications, bottlenecks and other specific drivers of cost.
• Past efforts have stalled. The cross-government ‘Top 75’ services programme, which ran from 2022 to 2025, closed with only 29 of the 75 services assessed as ‘great’. The NAO links this partly to a lack of sponsorship to improve data in departments.

The NAO’s conclusion is that public bodies should “act with urgency” to build a clearer, more consistent understanding of service costs. With the 2025 Spending Review setting a target of at least 5% efficiency savings by 2028-29, that urgency applies to IT as much as any other area.

Why IT spend is hard to see clearly

IT spending tends to be spread across many budgets, suppliers and contract types. That makes the total easy to report and the detail hard to explain. Common reasons include:

• Spend sits in several places. Central IT, departments, projects and shared services may all buy technology separately, often coded differently in the ledger.
• Invoices don’t match contracts. The contract describes what was agreed. Invoices show what was actually charged, including uplifts, variations and extras.
• Licence data sits somewhere else. Entitlements, usage and user numbers are held in portals and tools that finance rarely sees.
• Costs are allocated, not traced. Like the top-down models the NAO describes, IT costs are often spread across services by formula, which hides what each service really costs to run.

Common blind spots

• Duplicated tools or licences bought by different teams for the same purpose.
• Licences and subscriptions for leavers, retired systems or services that have moved to the cloud.
• Annual uplifts and indexation that have compounded without review.
• Low-value, high-volume purchases that never appear in contract registers.
• Contracts approaching renewal that nobody has flagged because the owner has moved on.

None of these are unusual. They are what happens when spend grows over years without a single, reliable view of it.

How to build an IT spend baseline you can act on

A good baseline does not need to be perfect on day one. It needs to be accurate enough to show where the money goes and where to look first. A practical approach has five steps:

1. Gather the sources. Pull accounts payable and invoice data, purchase orders, the contract register and the contracts themselves, plus licence entitlements and usage data from suppliers and admin portals.
2. Clean and match. Normalise supplier names, link invoices to contracts and flag spend that has no contract behind it. This step usually reveals the first surprises.
3. Categorise consistently. Group spend into clear categories, such as end-user software, infrastructure and hosting, cloud, network and telecoms, managed services, and support and maintenance. Use the same categories every year so trends are meaningful.
4. Link cost to drivers. Where you can, relate each category to what drives it: number of users, devices, services, transactions or capacity. Cost per user or per service is far more useful than a single total, and it is closer to the granular view the NAO recommends.
5. Add the timeline. Record renewal dates, notice periods and uplift clauses, so the baseline doubles as a forward plan of when each saving can realistically be pursued.

Separate potential savings from delivered savings

Once the baseline exists, opportunities start to appear. It is important to be honest about what stage each one is at. A simple pipeline helps:

• Identified: an opportunity has been spotted, such as unused licences or a contract priced above market, but it has not been tested.
• Validated: the evidence has been checked and a realistic range of value has been agreed with the budget holder.
• Agreed: a contract change, renegotiation or new procurement has been approved and signed.
• Delivered: the lower cost is visible in invoices or the ledger, compared with the original baseline.

Only the last stage is a saving. Reporting the pipeline as if it were delivered is how savings targets lose credibility. The NAO also notes that poor data limits the ability to measure improvement, which is another reason to fix the baseline first.

What is 1% worth?

A useful way to judge scale is to ask what a small reduction would mean on your own baseline. Every 1% of a £5 million annual IT spend is £50,000 a year. The table below shows the same arithmetic across different levels of spend.

This is an illustration, not a forecast. Figures are annual, shown before fees and the cost of achieving them. What is actually achievable depends on the evidence in your baseline, your contracts and your market.

The arithmetic shows the scale. The evidence decides what is realistic. That is why the baseline comes first.

Frequently asked questions

What is an IT spend baseline?
It is a consolidated, categorised view of what your organisation actually spends on IT, built from invoices, contracts and licence data. A good baseline shows not just the total but what drives it, such as cost per user or per service.

What data do we need to get started?
Usually accounts payable or invoice data, the contract register and key contracts, and licence and usage information. It does not need to be perfect at the start. Gaps in the data are often findings in their own right.

What is the difference between identified and delivered savings?
An identified saving is an opportunity that has not yet been acted on. A delivered saving is a reduction you can see in invoices or the ledger against the original baseline. Only delivered savings should be reported as achieved.

Did the NAO report look specifically at IT spending?
No. It looked at the cost of government services in general. Its findings on granular cost information and cost drivers apply just as well to IT budgets.