How do we reduce the cost of supporting our legacy applications?

Measure what each application really costs, stabilise it so the same faults stop recurring, consolidate and retire what is unused, and only then modernise where a saving justifies it; a fixed-scope support arrangement replaces open-ended contractor time with a predictable fee. This page is for an IT manager at a medium or large organisation who pays for legacy business applications and has been asked to bring the cost down without putting them at risk.

By Graham Church, founder and managing director. Updated 15 October 2026.

Where the money usually goes

A contractor or supplier paid by the day with no scope. The invoice reflects hours, not outcomes, and nobody has checked in a year what those hours produced. The rate rises because the person is the only one who knows the system.

Hosting sized years ago. Servers specified for a peak that never came, or for a user count the business has since moved past in the other direction. In the cloud, oversized machines are paid for every month; on premises, they come with maintenance contracts and the cost of replacing them.

Licences for things no longer used. Database editions above what the system needs, reporting and component licences renewed out of habit, user licences for staff who left, and the second system that still does the job the first one took over.

Firefighting instead of fixing. The same fault restarted every Monday, the same report re-run by hand each month-end. Each incident is cheap; the pattern is not, and it keeps a contractor billing.

Knowledge held by one person. The cost is hidden until they are ill, leave or retire, when a routine fault becomes an emergency rate and a week of downtime.

Our page on the rising running costs of an older system goes through each of these in more detail, including extended security updates.

The order to attack it

1. Measure. Gather every invoice and contract that touches each application: support, hosting, licences, hardware maintenance, extended updates. Add the time staff spend on workarounds and re-keying, which never appears on an invoice. Then ask what the support spend bought in the last twelve months: a list of changes made and faults fixed. For many organisations this list alone is the biggest saving, because it shows what is being paid for that nobody uses or that delivers nothing.

2. Stabilise. Fix the root causes of the recurring incidents, in order of how often they happen. Get the source code into version control and prove it builds and releases, so a fix no longer depends on one person’s machine. Prove the backups restore. This step costs money before it saves any, and it is what makes the support arrangement cheaper afterwards, because a stable system needs fewer hours.

3. Consolidate. Retire what is not used: servers for jobs that moved elsewhere, reports nobody opens, licences nobody signed in with. Where two systems do overlapping jobs, keep one, move the other’s data and users onto it, then switch the other off completely; our guide to consolidating overlapping business systems covers how. Resize hosting from measured use rather than the original specification.

4. Modernise. Only now, and only where it pays. Moving a database or framework off an out-of-support version removes extended-update fees and the premium a supplier charges for old technology. Replacing the part of the system that generates most of the support calls removes those calls. Each of these is a one-off cost justified by a measured saving, not a programme for its own sake.

Doing these in the wrong order is the common mistake. Modernising first spends money on a system whose costs were never measured, and consolidating before stabilising moves users onto a system that still fails every Monday.

What a fixed-scope support arrangement changes

A support arrangement with a written scope and a monthly fee changes three things compared with a contractor paid by the day.

The cost becomes predictable. The fee covers an agreed set of work (faults, security updates, keeping platform versions current, backup checks) and anything beyond it is quoted first, so the budget holds and the surprise invoices stop.

The incentive reverses. A supplier on a fixed fee loses money on a system that keeps breaking, so it is in their interest to fix root causes rather than bill for the same restart each week.

The knowledge stops sitting with one person. More than one engineer knows the system, documentation exists because the supplier needs it, and absence is covered. That is the single-person risk paid as a known fee instead of an unknown one.

It is not automatically cheaper by the hour. A contractor usually costs less per hour, and for a system that rarely needs touching that can be the right choice. The comparison is set out on our support contract or contractor page.

What not to cut

Some lines look like savings and are not.

Backups and restore testing. A backup that has never been restored is a hope. Cutting the time spent testing restores saves hours a year and risks the whole system.

Security updates. Stopping patching, or letting a server or database run past the end of vendor support with no plan, converts a monthly cost into an uninsured one. Extended security updates are a bridge while you upgrade, not a place to live.

Monitoring. Knowing that a job failed overnight costs little and is the difference between a quiet fix and a day of staff doing the job by hand.

The one person who knows it. If a contractor is the only knowledge you have, the saving is to get that knowledge out of their head and into code, documentation and a second pair of hands before changing the arrangement, not after.

Where to start

Measuring is the step you can do yourself this month. For the stabilise step, CodeFirst, a UK software company that maintains business systems other people built, starts with a fixed-price code audit that records the system’s condition, what it depends on and what could be retired, and then takes it on under a maintenance arrangement with a scope agreed in writing.

Related questions

Is a managed support contract cheaper than our contractor?

By the hour, usually not; a contractor's rate is normally lower. By the year, it can be, because a fixed-scope contract caps the cost, fixes root causes rather than billing for repeat incidents, and removes the emergency rates you pay when the one person who knows the system is unavailable. Compare the last twelve months of contractor invoices with a quoted annual fee for the same scope.

How do we find out what each application really costs?

Start with the invoices: support, hosting, licences, hardware maintenance and extended security updates. Then ask the teams that use each application how much time they spend on manual steps and workarounds. Put both in one list against the application each cost belongs to. The total is usually larger than the support contract alone suggests.

Can we cut costs without touching the code?

Often, yes. Retiring unused servers, resizing hosting from measured use, dropping licences nothing uses and ending the contract on a second system that overlaps the first need no development. Fixing recurring faults, removing manual steps and merging two systems usually do.

Should we stop paying for support and call someone when it breaks?

For a system the business depends on, no. The call-out will come at an emergency rate from someone who has never seen the system, and in the meantime security updates and backup checks stop. Reduce the scope of support to what the system needs rather than removing it.

Want the answer for your system?

Tell us what the software does and what you are trying to decide. We will reply with a straight answer and what we would need to know to be more precise.

Tell us about your system 0800 433 7990 Monday to Friday, 9am to 5pm. A first 20-minute call is free, and we reply to every enquiry within one working day. What happens after you get in touch