Maintain a website after launch
Launch is the beginning of a service, not the end of a project. A manageable maintenance plan names an owner, separates urgent fixes from planned improvements and records how the team will verify that the site still serves its users.
What to remember.
- Assign a named owner and a route for reporting faults.
- Test restoration and important user journeys, not only backup creation.
- Review content and accessibility alongside software and performance.
Name the service owner and the decision route
Record who owns the domain, hosting, code, content, analytics and supplier relationship. Give one person responsibility for prioritising changes and approving publication. Keep an up-to-date contact and handover record so a supplier change or absence does not leave the site unmanaged.
Maintain a small operational inventory
List the main page types, forms, integrations, libraries, renewal dates, third-party services and data flows. Mark each item with an owner and the consequence of failure. This inventory helps the team assess updates and makes incidents easier to investigate.
Plan updates, backups and restoration
Set a routine for software and dependency updates, with a test environment where changes can be checked before publication. Keep backups separate from the live service and periodically restore a copy to verify that it works. Define how to roll back a release and who can trigger that decision.
Review the information people rely on
Prioritise contact details, offers, dates, eligibility, source links and high-traffic guidance. Record a review owner and interval appropriate to each item. Correct or retire obsolete pages, then update internal links, language equivalents and the sitemap when the preferred URL changes.
Check key journeys and accessibility
Repeat a short set of real tasks on desktop and mobile: find a service, use a filter, submit a form and reach a relevant answer. Include keyboard navigation, focus visibility, text enlargement and meaningful image alternatives. An automated scan can help, but a human must still confirm that the task is understandable.
Use a small set of useful signals
Review availability, server errors, broken links, form delivery, search discovery and a few performance measures. Investigate trends in context rather than reacting to every metric change. Keep a dated record of incidents, fixes and decisions so recurring faults can be addressed at their source.
Create a sustainable review rhythm
Assign a quick weekly check for critical functions, a monthly review of updates and content, and a deeper quarterly review of user journeys, accessibility and suppliers. Adjust the frequency to the risk and pace of change. Each review should end with a short list of owners and due dates.