VDNX
ESLet’s talk

Maintenance & security·4 min read

How to define a useful website maintenance plan

Website maintenance means checking that the site still serves its purpose. Agree responsibilities, evidence and priorities before an incident occurs.

Start with what an interruption would put at risk

A website maintenance plan should explain what is checked, who acts and how completion is demonstrated. A generic list of updates is not enough to decide whether a service fits your business. Priorities depend on what the website allows people to do and what happens when part of that journey stops working.

Identify the important actions first: sending an enquiry, checking a price, signing in or completing a booking. Record who would notice a failure and what alternative the visitor would have. An information site and an application with user accounts need different plans. This classification helps you agree a sensible review frequency without suggesting every incident can be anticipated.

Create an inventory someone else can understand

List the domain, hosting, repository, publishing system and external providers. For each item, record its purpose, its owner and where access is managed. Add renewal dates where relevant and name the responsible contact. The inventory should not contain passwords: it only needs to identify where authorised access to them is managed.

Include less visible components, such as an email service, booking tool or script embedded in a page. OWASP explains that third-party JavaScript can introduce changes outside your control and expose information. Each integration should therefore have a clear purpose and a plan for removal when the function is no longer needed.

Pair updates with checks and a recovery option

For each change, record what is being modified, why and how it will be checked. An urgent correction calls for a different decision from an optional improvement. Do not let a routine calendar postpone the assessment of a relevant advisory. A responsible person needs to evaluate its impact and choose a proportionate response.

OWASP's dependency guidance describes testing updates in a separate environment and checking affected functions. Translate that into specific journeys in your plan: open key pages, send a test enquiry and check publishing. Keep a recoverable version as well, and record any data changes that could prevent a straightforward rollback.

Check the entire contact journey

A form can display confirmation while the message never reaches the team. A useful check follows the enquiry from the browser to the agreed mailbox or system. Use test details, identify the test submission and avoid real personal information. Also check that the visitor receives a useful explanation when an external provider fails.

Agree who monitors error reports and what each status means. A provider accepting a message does not prove it reached the recipient's inbox. Record only what is needed to investigate the journey and define a retention period. Maintenance can then support incident investigation without turning operational logs into an unnecessary second copy of every enquiry.

Treat access and service keys as ongoing work

Review accounts when a team member or supplier changes. Give each role the permissions it needs and avoid a shared account for every task. A clear handover should establish who controls the main accounts and who can recover the service if one person is no longer available.

OWASP describes service secrets as having a lifecycle that includes creation, rotation, revocation and expiry. Your plan should explain how to replace a key without putting it in code or logs. If a credential is exposed, removing the visible text does not revoke it: invalidate it at the provider and check the replacement configuration.

Ask for a recovery exercise, not just a backup

Define what needs to be recovered: code, images, content, configuration and data where applicable. Then agree where each item is stored and who can access it. A code backup does not replace an application's data. A content export does not, by itself, explain how to publish a working website again.

Propose a recovery exercise in a separate environment and document the steps. The aim is to discover missing dependencies or access before they are needed. Its scope and frequency should reflect the site's importance and rate of change. A successful exercise is useful evidence, but should not be described as a guarantee of recovery from every possible incident.

Turn reporting into clear decisions

Ask for a short record of changes, checks, open issues and the next action. Anything still outstanding should have an owner and a reason for its priority. A count of completed tasks is less informative than knowing whether enquiries work, a dependency needs attention or a renewal requires action.

Before arranging maintenance, use these questions to define the service. They also help distinguish recurring work from an additional intervention and establish what evidence you can expect to receive after the work is done.

  • Which parts of the website and which providers does the agreement cover?
  • How are incidents detected, reported and prioritised?
  • Which checks accompany an update?
  • Who holds the access and performs a recovery?
  • What evidence will you receive, and which actions need a separate estimate?

Sources and documentation

Content created with AI assistance. Check the sources, date and scope of each guide.

What does your website need next?

Tell us what runs your website, who maintains it and what you would like to improve.

Let’s talk about your idea