AegisRunner
All posts
Article· 10 min read

Website QA Checklist: What to Test After a Client Update

AegisRunner Team·October 20, 2026
Website QA Checklist: What to Test After a Client Update

After a client update, start with what changed and the journeys it could affect. Then record what you tested, what failed, and what remains unchecked. A clear scope gives clients more useful evidence than a blanket “site passed” label.

The post-update QA checklist

  • Confirm changed pages, routes, links, and redirects.
  • Recheck affected customer journeys and relevant failure states.
  • Test relevant browsers and mobile screen sizes.
  • Review accessibility around changed controls.
  • Check indexing signals where relevant.
  • Compare performance if assets, scripts, or layout changed.
  • Record outcomes, evidence, and anything not checked.

Illustrative seven-step website QA process: routes, journeys, responsive checks, accessibility, indexing, performance, and handoff

Illustrative checklist graphic, not a test report or AegisRunner interface.

Set the scope before testing

This guide covers post-update QA for websites managed by agencies. It is not a full security audit, compliance review, or native iOS or Android app release checklist.

Before you begin, confirm the environment, account, authorization, and permissions for each test. Use approved test data. Identify actions that must stay out of scope, such as real payments, bookings, or destructive changes.

Start with two questions:

  • What changed? List affected pages, templates, forms, navigation, content, scripts, and integrations.
  • What could break? Identify customer journeys that rely on those pages or components.

A shared navigation component, for example, can affect more routes than the page where its code changed. Choose checks based on the release and its dependencies.

Coverage also has limits. It depends on the pages, accounts, and environments made available, and on what the test actually reaches. A report describes its scope; it does not prove that every path on the site was tested. See how AegisRunner works for an overview of website scanning and test workflows.

Open every changed URL directly. Then reach it through the site’s navigation. Confirm that each page loads at the intended address and that its updated links, buttons, images, and downloadable files work as expected.

If a route changed, test both the old and new URLs. Check that the new URL loads the intended page and that the old URL redirects to the right destination, not a 404, a redirect loop, or an unrelated page.

Pay special attention to changes shared across templates, such as headers, footers, navigation, and reusable page components. A small shared change can affect many pages.

2. Recheck affected journeys and failure states

Test complete actions, not just page appearance. Follow the journey from its entry point to the expected outcome. For more on this approach, see automated UI testing.

Examples include:

  • Contact page → form submission → confirmation
  • Sign-in page → authorized account login → expected destination
  • Search or filter → results → selected detail page
  • Booking or checkout → safe test path → expected confirmation

Check relevant failure states too. Try an empty required field, an invalid value, or a search with no results. Confirm that the site explains the problem and provides a sensible next step.

Handle forms and integrations safely

Before submitting test data, confirm that emails, payments, bookings, and connected services use approved test destinations. A staging address does not guarantee that connected integrations are in test mode.

For a changed contact form, submit approved test data to a designated test recipient. Check the confirmation. Where you have access, verify that the submission reaches the intended inbox or system. Record anything you could not verify.

Never perform a real payment, booking, or destructive data action without authorization. If a step cannot be tested safely, mark it not tested or inconclusive and explain why.

Illustrative diagram of a website form connected to approved test destinations, with a live-service action blocked

Illustrative safety diagram, not a product screenshot.

For repeatable user journeys, keep a regression set and rerun relevant checks after future changes. To automate regression testing, start with the journeys that matter most to the client. This regression testing example shows how a targeted test can catch a change that breaks an existing flow.

3. Review changed layouts in relevant browsers and sizes

Check the browsers and screen sizes that matter to the client’s audience. At minimum, review the changed interaction on desktop and a small-screen viewport.

Look for:

  • Clipped text or overlapping controls
  • Hidden buttons, menus, or form fields
  • Broken line breaks or horizontal overflow
  • Dialogs that are difficult to close
  • Layout changes that obstruct the next action

A responsive website check is not a native iOS or Android app test. If the client also changed a mobile app, plan a separate device-specific test and report its setup and coverage separately.

4. Review accessibility around changed controls

Use the keyboard to reach and operate changed controls. Check that focus stays visible and follows a sensible order. For changed forms, review labels, instructions, and error messages. For updated images, check that meaningful content has appropriate text alternatives.

Automated checks can flag certain issues. They do not prove that every interaction works for every person, and they do not certify compliance. Pair automated findings with human review, especially for keyboard behavior and screen-reader output.

Use the W3C WCAG 2.2 Quick Reference to find relevant success criteria. Learn more about AegisRunner accessibility testing, including the distinction between automated findings and human evaluation. For a fuller list of checks, see the WCAG accessibility testing checklist.

5. Check search and indexing signals when affected

If the update changes a URL, page template, or indexing setting, review the page’s status, canonical URL, title, description, and internal links. Look for unintended restrictions on important pages.

Keep these two controls distinct:

  • robots.txt manages crawler access to URLs. It does not, by itself, keep a page out of search results. See Google’s robots.txt introduction.
  • Page-level noindex directives tell search engines not to show a page in results. See Google’s robots meta tag documentation.

A page can be accessible to visitors and still carry an unintended noindex directive. Check crawling restrictions and indexing directives separately.

Correct metadata does not predict rankings. This check catches unintended technical changes; it does not forecast search performance.

6. Compare performance when the update could affect it

Recheck performance when a release changes large images, fonts, scripts, layout, or third-party tools. Compare the same page under similar conditions, and record the device and test environment.

Where field data is available, consider real visitor experience alongside individual test runs. Google’s Web Vitals guidance describes three Core Web Vitals:

  • LCP measures loading performance.
  • INP measures responsiveness to user interactions.
  • CLS measures unexpected layout shifts.

These metrics help identify areas to investigate. A single score does not establish the performance of every page or every visitor’s experience.

Prepare a client-ready handoff

Give the client a release-specific record. Use clear outcomes: passed, failed, inconclusive, or not tested. Keep evidence tied to the exact environment and test scope.

Handoff detail What to record
Scope Release identifier, URLs, and journeys checked
Environment Environment, browser, viewport, and test-account role; never credentials
Outcome Passed, failed, inconclusive, or not tested
Evidence Expected vs. observed result and screenshot/run reference
Follow-up Owner, next action, and retest result

Illustrative graphic showing a defined set of tested pages and journeys beside an evidence record with pass, warning, failed, and unknown markers

Illustrative scope-and-handoff graphic, not a real AegisRunner report or interface.

Describe only the coverage the evidence supports. If a page, account, browser, or integration was unavailable, say so. A report can document a useful sample without implying that the whole site passed.

Keep the checklist with the release

Repeat the relevant checks after each client update. Keep the checklist with release notes, and report what passed, failed, and remains unverified.

Want to reduce repetitive checking? Explore how AegisRunner helps teams generate, review, and repeat website journey tests, or scan a website.