Fast. Simple. No meetings. No delays.
Home / Blog / UK / Ongoing Software Maintenance for a UK SME: What to Budget For

Ongoing Software Maintenance for a UK SME: What to Budget For

A realistic view of ongoing application upkeep — updates, security, hosting, small changes — and how to plan it as a predictable cost.

Cover graphic for the article: Ongoing Software Maintenance for a UK SME: What to Budget For

Budget for software maintenance as a recurring operating cost, not a one-off surprise. For a UK SME, that means setting aside time every month for updates, security, hosting and small changes, plus a modest reserve for the unexpected. This guide shows how to break the cost down so you can plan it.

Why software is never finished

A custom application is not a printed brochure. The day it goes live, it starts depending on things you do not control: the programming language and framework it runs on, the libraries it imports, the operating system and database underneath, the browsers your staff and customers use, and third-party services such as payment providers, email senders and accounting tools. All of those change on their own schedule.

Meanwhile, your business changes too. A new product line, a different supplier, a changed approval step or a new reporting request all land on software that was built for last year's process. If nobody owns that gap, the system slowly becomes the thing people work around, usually with a spreadsheet on the side.

The practical consequence is simple: if you only budget for the build, you have budgeted for the first chapter. If you are still choosing a supplier, our guide on how to vet an outsourced development partner includes questions about what happens after launch.

The four buckets of upkeep

It helps to split maintenance into four buckets, because each one behaves differently and is priced differently.

  1. Keeping it running (corrective). Investigating and fixing faults: an error after a data import, a report that stopped matching, a form that fails on one browser. Volume is unpredictable, but a stable system produces fewer of these over time.
  2. Keeping it current (adaptive). Updating the framework, language version, libraries and server software so they stay supported, and adjusting when an external service changes its interface. This is the bucket most often skipped, and skipping it is how systems become expensive to rescue.
  3. Keeping it safe (security). Applying security patches, reviewing access, rotating credentials, checking backups can actually be restored, and responding when a vulnerability is announced in a component you use.
  4. Keeping it useful (small improvements). Small requests from the people who use it daily: a new field, a changed label, an extra export, a tweaked permission.

The first three are largely driven by technology and the calendar. The fourth is driven by your business. Treat them as separate lines in your plan so that a busy month of improvements does not quietly eat the time meant for security patches.

Hosting and infrastructure

Hosting is usually the most predictable part of the budget, but it has more moving pieces than the monthly invoice suggests. When you plan, list each of these and who pays for and manages it:

  • The server or platform the application runs on, and any separate database service.
  • Backups: how often, where they are stored, how long they are kept and who tests a restore.
  • Domain names, DNS and SSL/TLS certificates, including renewal dates and who receives the reminders.
  • Email sending, file storage and any queue or search service.
  • Monitoring and alerting, so someone hears about downtime before your customers do.
  • Licences for commercial components or third-party APIs, which may be priced per user or per volume.

Ask who holds the accounts. A sensible default is that the hosting and domain accounts are in your company's name, with your supplier given access, so you are never locked out of your own system. Costs here scale with usage and with how much resilience you need, so ask for a range tied to your expected traffic rather than a single number.

Small changes vs projects

Much budget friction comes from not agreeing where a small change ends and a project begins. A useful working rule: a small change is something a developer can understand, build, test and release without redesigning data structures or workflows. Renaming a status, adding a column to a table or adjusting an email template are small. Adding a new approval stage, a customer portal or a new integration is a project and deserves its own estimate.

To plan small changes, agree a monthly allowance of time, keep a visible backlog, and review it together. If the backlog keeps growing past the allowance, that is a signal either to raise the allowance or to turn the largest items into a scoped project. If you have a system you are considering replacing instead, modernizing a bespoke PHP system covers when upkeep stops being the right answer.

Software maintenance SME UK budgets: a yearly plan

A defensible yearly budget has a few lines rather than one lump. Build it like this:

LineWhat drives itHow to plan it
Hosting and servicesTraffic, storage, resilience, licencesQuote from current invoices plus a growth allowance
Updates and securityAge of the stack, number of dependenciesFixed monthly time block or scheduled quarterly work
FixesComplexity and quality of the original buildReserve drawn on as needed
Small improvementsHow actively your team uses the systemMonthly allowance with a backlog
Larger projectsYour business plansBudgeted separately, estimated when scoped

As a rule of thumb, newer, well-tested systems with few dependencies need less than older ones with many integrations, but there is no honest universal ratio of maintenance to build cost, and anyone quoting one without knowing your system is guessing. The better approach is to ask your supplier for an itemised estimate based on your actual stack, then compare it with a year of real usage and adjust.

Also decide what happens to unspent time. Some arrangements roll it over, some expire it. Either is fine if it is stated up front.

Who does it: in-house, supplier or both

You have three realistic options. An in-house developer gives you fast context and availability, but a single person is a continuity risk and may not cover security, hosting and development equally well. A supplier gives you a broader skill set and cover for holidays and sickness, but needs clear documentation and a defined response process. A hybrid, with an in-house owner of priorities and a supplier doing the technical work, is common in smaller firms. The trade-offs are similar to those in in-house versus outsourced development for a UK SME.

Whichever route you choose, insist on three things: the source code and documentation are accessible to you, deployments are repeatable rather than done by hand from one laptop, and at least two people know how the system is released. Those three habits make any later change of supplier far less painful. For a look at how a UK-focused delivery team works with clients, see our software development for UK businesses page.

Reporting and review

Maintenance you cannot see is hard to value. Ask for a short monthly or quarterly report that covers at least:

  • Work done, split by the four buckets, with time spent against the allowance.
  • Updates applied and anything deferred, with the reason.
  • Incidents: what happened, how long it took to notice and fix, and what changed to prevent a repeat.
  • Backup and restore checks, and any expiring certificates, domains or licences.
  • Upcoming end-of-support dates for the language, framework or server software you rely on.
  • The current backlog and a recommendation for what to do next.

Review this once or twice a year with someone who can make budget decisions. The goal is not more meetings but a short list: what is working, what is getting expensive, and what should change.

Putting it together

Start with an inventory of what you run, who owns each part and when each component reaches the end of support. Turn that into the four buckets, agree a monthly allowance and a reserve, and ask for a simple report. If you want a second opinion on an existing quote or an application you have inherited, you can get in touch and describe it in plain terms.

Frequently asked questions

What percentage of build cost is maintenance?

There is no reliable single figure, and we would not trust one quoted without context. It depends on the age of the stack, the number of integrations, how actively the system changes and how well it was built. Ask for an itemised estimate for your own application, then check it against a year of real usage.

How do I plan small changes?

Agree a monthly allowance of developer time, keep a shared backlog and review it regularly. Define a small change as work that needs no redesign of data or workflow. Anything larger becomes a scoped project with its own estimate, so routine requests do not crowd out security and update work.

What should reports show?

A useful report lists work done by category against the time allowance, updates applied or deferred, incidents and how they were resolved, backup and restore checks, upcoming end-of-support dates and expiring certificates or domains, plus the current backlog with a recommended next step.

Related articles

Business Technology Australia

Replacing Spreadsheets With a Custom Web App

How to recognise when spreadsheets have become the bottleneck, and a low-risk way to move the process into a purpose-built web application.

5 min read

🚀 Start Now