Software maintenance budgets are usually set by subtraction: whatever remains after licences, salaries and new projects becomes the maintenance envelope. That method has a predictable result. The work that prevents incidents is deferred until an incident forces it, and the cost arrives later as an emergency rather than as a planned line.
This comparison sets out five ways to fund the ongoing care of business software, the criteria that separate them, and the cost lines any of them must expose. It is written for an SME that owns a custom application, an integration or a small portfolio of internal tools and needs a budget it can defend.
The five funding models
Fixed percentage of build cost. The annual envelope is set as a stated percentage of the original build cost or replacement value. The percentage is a policy choice, not an industry fact; it should be reviewed against the actual work log each year.
Prepaid hours retainer. The business buys a block of hours each month, with a stated response time and a defined scope for what counts as maintenance. Unused hours may roll over or expire, which changes the real value.
Outcome-based service level. The provider is paid to keep defined services available and within agreed performance targets, with credits or remedies when they are missed. It aligns incentives around reliability but can leave improvements and small changes outside the contract.
Time and materials per incident. Work is authorised case by case. It is transparent and flexible, and it makes preventive work hard to fund because nothing is authorised until something breaks.
Internal team plus escalation partner. Salaried staff handle routine changes and first-line support, with an external partner for specialist work or overflow. It retains knowledge but depends on hiring, cover and documentation.
Criteria that matter
- Predictability: whether the monthly cost is stable enough to budget and explain.
- Coverage: which work is included, especially security patching, dependency upgrades, monitoring and small enhancements.
- Incentive alignment: whether the provider benefits from preventing incidents or from responding to them.
- Administrative load: the hours your team spends approving, tracking and reconciling the work.
- Scalability: how cost and service change when another application or integration is added.
- Knowledge retention: whether documentation and understanding stay with the business.
- Auditability: whether you can see what was done, when and by whom.
Funding models side by side
| Criterion | Fixed percentage | Hour retainer | Outcome SLA | Time and materials | Internal plus partner |
|---|---|---|---|---|---|
| Predictability | High | Medium | High | Low | High for salary, variable for overflow |
| Coverage | Depends on definition | Depends on hours | Defined services only | Only what is authorised | Broad but capacity-limited |
| Incentive | Weak unless reviewed | Weak at renewal | Strong for uptime | Reactive | Mixed |
| Admin load | Low | Medium | Medium | High | High |
| Scalability | Easy to recalculate | Easy to adjust | Requires renegotiation | Immediate but costly | Slow, through hiring |
| Knowledge retention | Depends on documentation | Depends on documentation | Often provider-held | Provider-held | Strongest |
| Auditability | Annual review | Timesheets | Service reports | Invoices | Internal records |
Cost lines every model must expose
- Corrective fixes: defects, failed jobs and user-reported errors, with a severity definition.
- Security patching: framework, library and platform updates, including testing time.
- Dependency upgrades: payment, messaging and infrastructure providers that change their APIs.
- Monitoring and response: alerts, log review, on-call time and the escalation path.
- Small changes: the requests that are too small for a project but too frequent to ignore.
- Licences and environments: hosting, tooling, testing environments, certificates and backups.
- Documentation and handover: runbooks, architecture notes and access records.
- Contingency: a stated reserve for the incident that does not fit any category.
Build a defensible envelope
Start from the work, not from a percentage. Estimate annual hours by category, multiply by a blended rate, add pass-through costs and a contingency, then compare the result with what the last twelve months actually consumed.
Annual envelope = (corrective hours + patch hours + upgrade hours + monitoring hours + small-change hours) × blended rate + licences and environments + contingency
A percentage-of-build heuristic is useful as a cross-check, not as the primary method. If the formula and the heuristic disagree, the work log is the better evidence. Review the envelope quarterly and reforecast when a major dependency changes.
Choose the fixed percentage when
The application is stable, the scope is easy to define and the business wants a single number for planning. Write a service definition that lists what is included, and review the actual work annually so the percentage does not become a fee for doing very little.
Choose the hour retainer when
Demand is uneven and the business needs access to a team without hiring. Track utilisation and rollover rules, and require a monthly summary of work by category. A retainer without reporting is a subscription, not a maintenance plan.
Choose the outcome SLA when
Availability and recovery matter more than a list of tasks, and the provider can genuinely influence them. Define the service, the measurement method and the remedy. Keep enhancements in a separate agreement so reliability work is not starved by feature requests.
Choose time and materials when
The system is approaching replacement, demand is genuinely unpredictable, or the business needs a short bridge. Authorise work in small batches, require estimates before starting, and keep a running log. This model is a tool for a transition, not a stable annual plan.
Choose an internal team plus partner when
The application carries knowledge that cannot leave the building and the business can employ or assign capable staff. Budget for documentation, cover during leave and a partner who can step in without a long ramp-up.
Contract questions that protect the budget
- What exactly is excluded, and what is the rate for excluded work?
- Who owns the code, documentation and deployment configuration?
- What is the response time for a critical defect, and how is critical defined?
- How are third-party licence increases passed through?
- What notice is required to change the model or end the agreement?
- What evidence is provided each month to show the work performed?
We maintain and extend business systems under Custom Software & Integrations, with the maintenance scope written separately from new build. The pricing page explains how retainers and project work are quoted. For the wider planning context, read how SME IT budgeting changes year by year.
Sources
- NIST, SP 800-40 Rev. 4, Guide to Enterprise Patch Management Planning, accessed 17 November 2026.
- ISO, ISO/IEC 20000-1:2018, Service management system requirements, accessed 17 November 2026.
- FinOps Foundation, FinOps Framework, accessed 17 November 2026.
- Flexera, 2026 State of the Cloud report, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.