What does it really cost to keep your current IT?

Keeping the current system often looks like the cheapest option. There is no procurement to run, no migration to plan and no disruption to users. The support contract comes up for renewal, the supplier sends a quote and the agreement is extended for another term.

The trouble is that the cost of “keeping” is rarely measured properly. In its March 2023 report, Digital transformation in government: addressing the barriers to efficiency, the National Audit Office (NAO) found that outdated IT systems and ageing data are a key source of inefficiency and a “major constraint” on improving and modernising government services. It also said their limitations and poor-quality data increase the cost of services.

For a public body facing a renewal on an ageing system, that leaves a practical question: should you maintain it, renegotiate it or replace it, and do you know enough about what you currently pay to decide?

What the NAO said about legacy systems

The report looks across central government, but several findings apply to any organisation running older technology:

• Keeping legacy is not a neutral choice. The NAO says tackling legacy is costly and complex, but continuing to use it is risky and commits government to even higher future costs and risks, including service disruption, operational failure and cyber-attacks.
• Maintenance costs and skills are a growing problem. It describes the rising cost of maintaining legacy systems, and the loss of the specialist skills needed to support them, as unsustainable in the long term.
• A large share of spend goes on standing still. Government estimated that nearly half of all technology expenditure across government in 2019 went on “keeping the lights on” for outdated legacy systems.
• Many costs are hidden. The NAO points to manual workarounds for missing functionality, extra complexity from additional interfaces, and the costs of third-party support agreements. It found departments often do not know the full business cost of running legacy services.
• Funding rules can make it worse. Departments told the NAO it is easier to bid for capital funding for new developments than for resource funding to maintain and update existing services.

The report also notes how government defines a legacy system: one that is out of support, not cost-effective, hard to maintain, above an acceptable risk threshold or an end-of-life product. That is a useful test for any public body looking at its own estate.

Where the cost of keeping legacy IT sits

The contract price is usually only part of the picture. When we look at ageing IT agreements, the cost of keeping a system tends to fall into a few areas:

• Extended or premium support. Once a product passes its standard support dates, suppliers may charge more for continued patches and fixes, or offer limited support at the same price.
• Scarce skills. Older platforms need people who know them. Whether that knowledge sits with the supplier, contractors or a few internal staff, it tends to cost more over time.
• Workarounds. Manual processes, spreadsheets and extra checks fill gaps the system cannot. These costs rarely appear on an IT budget line.
• Services that no longer match need. Support hours, service levels, environments and modules may have been set years ago for a different organisation.
• Duplicated licences. As new tools arrive, older licences are often kept “just in case”, so the organisation pays twice for similar capability.
• Risk. Security exposure and the chance of an outage are real costs, even if they only show up when something goes wrong.

Signs an ageing agreement should not simply be extended

• The product is out of mainstream support, or will be during the next term.
• The renewal price has risen but the scope, volumes or service levels have not changed.
• Licence counts or capacity are based on user numbers or services that no longer exist.
• Only the incumbent supplier, or one or two individuals, understand how the system works.
• Staff rely on manual workarounds for tasks the system should handle.
• Nobody can say with confidence what the system costs to run in total, not just the contract value.

If several of these apply, a straight extension is likely to lock in costs that could be reduced, and to push a harder decision a few more years down the road.

Maintain, renegotiate or replace: making a fair comparison

These three options are not always exclusive. A common route is to renegotiate a shorter, better-priced agreement that keeps the service stable while a replacement is planned. Whatever the outcome, the comparison should be made on the same basis for each option. A practical approach looks like this:

1. Build the commercial baseline. Gather invoices, order forms, variations and uplifts to find the true annual cost of the current arrangement, not just the headline contract value.
2. Review licences and usage. Compare what you are licensed for with what is actually used. Look for inactive accounts, unused modules and overlapping products.
3. Test support against operational need. Check whether service levels, support hours and environments still match how the service runs today.
4. Benchmark the price. Compare unit rates and terms with current market pricing, so you know whether the incumbent’s offer is fair for what it covers.
5. Add the costs outside the contract. Estimate internal effort, workarounds and the cost of scarce skills, even if the figures are approximate.
6. Cost replacement honestly. Include migration, data cleansing, transition support, training, business change and any period of running both systems in parallel.
7. Weigh the risk. Consider the likelihood and impact of failure under each option, including supplier support, security and skills.

What each option typically involves

The table below sets out the main cost categories to consider for each option. It is a checklist, not a scoring model.

This is a general illustration, not an assessment of any particular system. The right answer depends on the contract, the technology, the market and the organisation’s wider plans.

Why renegotiation often comes first

Replacement takes time. Even where it is the right long-term answer, most organisations need the current system to keep running while a new one is procured and implemented. That makes the terms of the current agreement important. Paying for unused licences, outdated service levels or a long term you do not need can absorb money that could fund the replacement itself.

A well-evidenced renegotiation can remove costs that no longer serve a purpose, align the term with a realistic replacement timeline, and include exit and transition support that will be needed later. It also gives you a clearer baseline, which makes any replacement business case more credible.

Frequently asked questions

What counts as a legacy IT system?
The NAO’s 2023 report notes that government treats a system as legacy if it is out of support, not cost-effective, hard to maintain, above an acceptable risk threshold or an end-of-life product. A system can be legacy even if it still works day to day.

Is it always cheaper to replace an ageing system?
No. Replacement brings its own costs, including migration, transition support, business change and a period of running both systems. The point is to compare all options on the same full-cost basis rather than assume either answer.

Can we renegotiate a contract for a system we plan to replace?
Often, yes. Rescoping support, removing unused licences and aligning the term with a realistic replacement timeline can reduce costs while keeping the service stable during the transition.

Where should we start?
With the commercial baseline: what the current arrangement actually costs, what is being used and what the contract commits you to. Without that, any comparison between options rests on assumptions.