Contact

Software maintenance and operations: what happens after launch?

A Japanese woodblock print: a towering, foaming wave rises over three boats, Mount Fuji in the background. (Katsushika Hokusai, Under the Wave off Kanagawa (The Great Wave), c. 1830–32)
Katsushika Hokusai: Under the Wave off Kanagawa (The Great Wave), c. 1830–32 · The Metropolitan Museum of Art · Public domain
Category
Project
Published
Reading time
~3 min read
Author

Summary

Software is not finished on launch day: that is when its life begins. Operations and maintenance are predictable if responsibilities are agreed in advance.

Why does finished software need ongoing attention?

Software does not wear out like physical machinery, but its environment changes constantly. Browsers and mobile operating systems get updated, new security vulnerabilities are discovered in third-party libraries, connected APIs evolve and regulations shift. Left untouched, a system becomes risky over time even if not a single line of its own code has changed.

Your business changes too: new products, new workflows, new team members. Well-maintained software evolves alongside them.

What does post-launch work involve?

Types of post-launch responsibilities
ActivityWhat it coversWhen it happens
OperationsRunning the system: servers, hosting, SSL certificates, backupsContinuously
MonitoringTracking uptime, errors and performance; alerting on issuesContinuously
Security maintenanceUpdating dependencies, libraries and the runtime environmentOn a regular schedule, plus urgently when critical vulnerabilities emerge
Bug fixingResolving defects uncovered during day-to-day usePrioritised by severity
Further developmentAdding new features and adapting to changing business workflowsDriven by business needs

It makes sense to treat these categories separately because they have different rhythms and cost structures: operations are an ongoing baseline cost, while feature development is a planned investment.

What should a support agreement cover?

  • Responsibilities: who runs the infrastructure, who fixes defects and who is the point of contact when an issue arises.
  • Response times: how quickly work begins on an incident, tiered by severity.
  • Backups and recovery: how often backups run, where they are stored and whether restoration has actually been tested.
  • Update schedule: how and how often security patches and dependency updates are applied.
  • Access and ownership: who holds access to the source code repository, servers and third-party service accounts.
  • Billing terms: what is covered by the monthly retainer and what counts as a separate work order.

Expected response times and availability targets depend on how critical the system is: an internal directory and a 24/7 high-volume online store require very different service levels.

How to avoid vendor lock-in

As a client, you are safe when the system does not live solely inside one person's or one agency's head. A few straightforward principles ensure this.

  1. The source code lives in a repository that you own and can access.
  2. Domain names, hosting accounts and third-party subscriptions are registered in your company's name.
  3. Up-to-date documentation explains the architecture and how to deploy the system.
  4. Deployments are automated rather than relying on manual steps known only to one engineer.

Insisting on these is not a sign of distrust. A professional development team works this way by default because it makes operations safer and cleaner for everyone.

How to budget for post-launch costs

Post-launch costs depend on system complexity, traffic volume, availability requirements and how rapidly your business evolves, so there is no universal price tag. What is certain is that you should factor them in when making the initial investment decision: total cost of ownership is initial development plus multi-year operations.

Ask vendors to break out post-launch operational and maintenance costs separately in their proposals. What you can see in advance, you can plan for.

Key takeaways

  • A software system's environment changes even when the code does not; that is why ongoing maintenance is essential.
  • Operations, security patching, bug fixing and feature development are distinct activities with different rhythms.
  • Response times and availability targets should match how business-critical the system is.
  • Keep source code repositories and third-party service accounts in your own company's hands.
  • Total cost of ownership combines build cost and multi-year operations: ask for both in the initial proposal.

Frequently Asked Questions

Do we have to sign a maintenance contract with the original developer?

No, it is not mandatory, but someone must be accountable for the system. That can be the original developer, your internal IT team or an external operations partner; what matters is that ownership is unambiguous.

What happens if we do not update the system?

For a while, nothing visible. Beneath the surface, however, unpatched security vulnerabilities accumulate, compatibility with new browsers and devices degrades, and deferred upgrades eventually turn into a much larger, riskier overhaul.

Can a new development partner take over a system built by someone else?

Usually yes, starting with a technical handover audit: assessing code quality, checking documentation and verifying how deployments work. The audit shows how much preparation is needed to maintain the system safely.

Facing a similar decision? We are happy to talk it through.

Let's talk