A maintenance handover checklist for agencies using external developers

A maintenance handover checklist for agencies using external developers

By DOT IT Services

A website can be straightforward to launch and difficult to maintain. The original developer understands where the code lives, which integration sends notifications and why a deployment needs a particular setting. The agency taking responsibility for support may receive little more than an admin login.

A useful handover transfers the knowledge needed to operate the site. It should help an authorized person make a routine change, investigate a failed enquiry and identify who owns a problem without relying on private messages from the original build.

The checklist below is intended for agencies working with an external development team. Adapt it to the systems actually used rather than filling every field with generic documentation.

1. Identify what is being handed over

Record the production website, staging environment, code repository and deployment destination. State which parts of the system the external team built and which remain under another supplier's control.

List the important integrations, such as forms, payments, appointment tools, CRM connections or transactional email. For each one, record its purpose, account owner, support contact and where the configuration is documented.

Keep passwords and tokens out of the handover sheet. Refer to the approved secure location or access process instead. Documentation should explain how authorized people obtain access without exposing credentials to everyone who can read project notes.

2. Assign a business owner and a technical owner

The person who approves a change may not be the person who implements it. Identify both roles, including a backup contact for time-sensitive issues.

For an agency development partnership, also record who can communicate with the end client. A technical incident should not create uncertainty about whether the developer or account manager sends the update.

Agree which decisions require approval: changes to service information, paid integrations, production deployments or work outside the support scope. This prevents a request to investigate an issue from being treated as unlimited permission to alter the live site.

3. Demonstrate one routine change

Choose a low-risk example, such as updating a test paragraph in staging. Ask the incoming team to follow the written instructions, preview the result and explain how it would be approved for production.

A successful demonstration can reveal missing steps more quickly than reading a long setup document. If the process depends on a particular machine or undocumented configuration, resolve that dependency before treating the handover as complete.

Record how a release is deployed and how the team identifies the version currently running. Include the approval step and the intended response if the release fails.

4. Trace a test enquiry through the system

Use fictional details and confirm where the test submission arrives. Inspect the receiving record rather than relying only on the website's success message.

Record which team owns the website form, notification delivery and receiving system. A failure can occur between them, so “contact the developer” may be too vague to help the person handling support.

Check what evidence can be collected without exposing customer data. An internal test identifier, timestamp and error description may be enough to start investigation. Real enquiry contents should not be copied into public issue trackers.

5. Review dependencies and recurring charges

Document hosting, domain renewal, paid plugins, third-party services and any licences required to keep the site operational. State who pays each charge and what happens if a supplier relationship ends.

For platforms with automatic updates, identify who monitors failures and checks important journeys afterwards. For custom applications, identify how dependency updates are reviewed and released. Avoid promising that a website will remain secure or compatible simply because it was working at handover.

If the agency is buying ongoing developer capacity, make maintenance responsibilities explicit in the work queue. Available development time does not automatically mean someone is monitoring incidents, reviewing updates or approving releases.

6. Document recovery instead of assuming it

Record what is backed up, where the backup is held, its retention arrangements and who can restore it. Distinguish website files from databases, uploaded media and configuration that may be stored elsewhere.

Where a recovery test is agreed, record its actual outcome and limitations. Do not describe a restore as tested merely because a backup job reports success. A recovery process needs to be understandable to the people expected to use it.

Also consider new activity received after a backup was taken. Restoring an older system may require reconciling recent orders or enquiries. The incident owner should know that this is a decision to investigate, not an automatic consequence to ignore.

7. Agree support boundaries and close the handover

List known issues, unfinished work and the period covered by any agreed defect support. Separate defects against the original requirements from maintenance and new features. Record how additional work is approved and charged under the actual agreement.

Use a final sign-off table:

Check Evidence or owner
Authorized access works Named reviewer and check date
Routine change demonstrated Staging example or internal record
Test enquiry received Test identifier and receiving system
Release process documented Current instructions and owner
Recovery arrangements recorded Backup scope and any test outcome
Remaining issues accepted Issue list, owners and next actions

The handover is useful when the incoming team can operate the website and explain its limitations. A folder of files or a collection of logins is only part of that result.