Technical SEO · Supporting checklist

Page Speed SEO Checklist

A page speed SEO checklist is an implementation workflow for reducing unnecessary loading and rendering delays. Start with a repeatable test, locate the resource or processing bottleneck, and improve images, fonts, caching, compression or scripts where the measurement supports it. Retest the complete live page and keep lab improvements separate from real-user Core Web Vitals and search rankings.

Free to useNo account neededSaved in your browser
Staggered resource-loading bars between a server stack and a page, with a stopwatch and image and code tiles.
Trace resource loading to find delays, make a focused change and test the page again.
Your working checklist

Work through the checklist

Click a square to mark a task as Pass. Click again to undo. Use the result menu for a problem, a blocker or a task that does not apply. You do not need to enter notes.

18 checks

Find the Actual Loading Bottleneck

Page-speed diagnosis starts with the affected page, visit state and visible delay. Match the next implementation task to a measured cause.

Test first and repeat visits on representative pages

Find out when the slow experience happens.

Use this for: The representative public pages selected for this review.

Choose important landing-page templates and run a mobile and desktop diagnostic where each matters. In browser developer tools, open Network before reloading to see the document and resource requests. Start with a controlled first-visit test, then compare a normal repeat visit with caching enabled.

Keep the URL, test mode, network setting and page state comparable. A warm browser cache can hide a slow first visit, while an always-disabled cache cannot show whether repeat visits benefit from reuse. Do not treat one fluctuating performance score as the entire result.

Example

Hypothetical example: A service page is quick on the owner’s laptop after many visits but slow on a fresh load because its large images have never been cached.

  1. Choose URLs across the important templates.
  2. Compare controlled first-visit and normal repeat-visit behavior.
  3. Keep the setup consistent for later tests.
More help and optional notesHigh · SEO analyst and developer
PassThe baseline distinguishes first and repeat visits with comparable conditions.
FailA cached personal visit or a single score is treated as the full performance picture.

If it fails: Repeat the missing visit state before selecting a fix.

Retest: Reuse the same baseline after implementation.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Choose the resource or delay that blocks the useful page

Fix the observed bottleneck before clearing every warning.

Use this for: The representative public pages selected for this review.

Use the report’s insights and the request timeline to find what delays the main content or the next important action. Note whether the page waits for the first response, a large download, a chain of dependencies or script execution. Ask the developer to explain the specific resource if the report is unfamiliar.

Prioritize a repeatable problem affecting important templates. Estimated savings are diagnostic suggestions, not guaranteed additive gains. Avoid adding every audit estimate together or making cosmetic changes only to chase a score.

Example

Hypothetical example: A slow page waits on a stylesheet imported by another stylesheet. Removing that request chain is a clearer first task than compressing an unrelated footer icon.

  1. Locate the delayed content or action.
  2. Trace the resource or main-thread work that blocks it.
  3. Choose one high-impact cause to repair and retest.
More help and optional notesHigh · SEO analyst and developer
PassThe next fix targets a reproducible delay with a clear expected result.
FailThe queue contains generic optimizations with no connection to the observed delay.

If it fails: Inspect the relevant timing detail and assign a cause-specific change.

Retest: Compare that delay, not only the headline score.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Improve Response and Critical-Resource Delivery

Resource delivery checks inspect the request path, transfer encoding and discovery order. Verify the response visitors actually receive.

Trace slow document responses and unnecessary redirects

Reduce waiting before useful page resources can start.

Use this for: Pages with slow document response or unnecessary redirect paths.

Open the main document request in Network and inspect its redirect path and timing. Check whether an internal link sends visitors through avoidable intermediate URLs before the final page. Preserve necessary redirects and the chosen canonical destination.

If waiting for the response dominates, ask the administrator to investigate the relevant cache miss, server work, region or backend dependency. Time to First Byte includes more than application processing and is not itself a Core Web Vital. Retest the actual audience-relevant route before buying new hosting or adding a CDN.

Example

Hypothetical example: A navigation link uses an old HTTP path that redirects twice before reaching the current HTTPS service page. Updating the link removes those avoidable steps.

  1. Inspect the document redirect path and timing.
  2. Update avoidable internal detours or investigate the slow response stage.
  3. Retest the final route under comparable conditions.
More help and optional notesHigh · SEO lead and developer
PassThe identified avoidable detour or response bottleneck is resolved without changing the intended destination.
FailA slow document response remains unexplained or necessary routing is broken.

If it fails: Repair the specific link, cache or backend cause with the responsible administrator.

Retest: Repeat the same route and relevant first-visit state.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Verify compression on larger text responses

Reduce transferred HTML, CSS and JavaScript bytes.

Use this for: The representative public pages selected for this review.

In Network, inspect a substantial HTML, CSS or JavaScript response. Check its Content-Encoding header and transferred size. Ask the hosting provider or developer to enable supported compression such as Brotli or gzip when the response is uncompressed and compression is appropriate.

Check the response that visitors actually receive, including the CDN or proxy layer. Minification and transfer compression are different: a minified script can still be sent uncompressed. Already-compressed media and tiny files may not benefit from the same treatment.

Example

Hypothetical example: A minified stylesheet is still transferred without compression. Enabling the host’s supported text compression reduces transfer bytes without changing the CSS rules.

  1. Inspect headers and transfer size for larger text resources.
  2. Enable appropriate supported compression at the serving layer.
  3. Verify the live encoded response and correct content.
More help and optional notesHigh · Developer or hosting administrator
PassThe selected compressible text responses are transferred with an appropriate encoding and work correctly.
FailLarge compressible resources are sent uncompressed or a setting is assumed to work without response checks.

If it fails: Correct the serving-layer compression and verify the actual response.

Retest: Retest the resource after deployment and cache refresh.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Make the important resource discoverable early

Avoid waiting for optional scripts before finding the main content.

Use this for: The representative public pages selected for this review.

Inspect when the main image, critical stylesheet or required font is requested. If the important image is introduced only after JavaScript runs or is buried behind a stylesheet, ask the developer whether it can be discovered in the initial HTML or needs a carefully chosen resource hint.

Use preloads and high fetch priority selectively after confirming the bottleneck. Do not preload the whole page. Make sure a preload matches the resource and responsive candidate actually used; otherwise it can waste bandwidth or create a duplicate request.

Example

Hypothetical example: A category banner is added after an optional animation script. Placing its image reference in the initial markup lets loading begin earlier.

  1. Find the critical resource request and its initiator.
  2. Remove the unnecessary discovery delay or test a selective hint.
  3. Check for earlier loading without duplicate or competing requests.
More help and optional notesHigh · SEO lead and developer
PassThe needed resource starts earlier without unnecessary duplicate or high-priority work.
FailThe resource remains late-discovered or broad preloading makes it compete with unrelated downloads.

If it fails: Correct discovery and limit hints to proven critical resources.

Retest: Compare the request timeline and visible result.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Image Delivery Checklist

Image delivery checks balance downloaded dimensions, transfer size and useful visual detail. Verify the candidate and file the browser actually uses.

Serve image dimensions that fit the layout

Avoid downloading a desktop-sized image for every phone.

Use this for: Templates with raster content images.

Choose a large image from the affected template and inspect its actual downloaded file on a phone-sized viewport. Compare the file’s pixel dimensions with its displayed size and device pixel density. Ask the developer to configure appropriate responsive candidates and sizes where the component needs them.

Check the candidate the browser really requests, not only the filenames in the CMS. Preserve sufficient quality for high-density screens, meaningful crops and any product detail users need to examine. The smallest possible file is not automatically the correct image.

Example

Hypothetical example: A card displayed about 350 CSS pixels wide downloads the same 4000-pixel original as the desktop gallery. Responsive candidates reduce the unnecessary transfer.

  1. Inspect the downloaded image and its display size.
  2. Configure suitable candidates and a correct layout-size hint.
  3. Confirm the requested candidate at phone and desktop widths.
More help and optional notesHigh · SEO lead and developer
PassThe browser selects appropriately sized images for the tested layouts and density.
FailEvery small component downloads an unnecessarily large original or an unusably small image.

If it fails: Correct the image variants and sizes configuration.

Retest: Inspect actual requests after the change, including high-density display.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Compare format and compression without losing useful detail

Choose a smaller image that still does its job.

Use this for: The representative public pages selected for this review.

Export a representative image with suitable format and quality options, then compare file size and visible quality at its intended display size. AVIF or WebP can help, but photographs, logos, transparency and detailed diagrams have different needs.

Check the selected file in the browsers you support and keep appropriate fallback behavior where required. Preserve the subject, readable labels and important product detail. Avoid promising a fixed percentage saving or converting every image to one format without comparison.

Example

Hypothetical example: A product photo compresses well, but the same setting blurs a technical diagram’s labels. The team uses different exports for the two content types.

  1. Compare suitable format and quality exports.
  2. Inspect the actual display size and important details.
  3. Verify delivery and supported-browser behavior.
More help and optional notesHigh · Content editor and developer
PassThe selected image reduces unnecessary bytes while preserving the needed visual information.
FailCompression harms useful detail or a blanket format rule creates a broken image.

If it fails: Choose a better format, quality level or fallback for that asset.

Retest: Compare the live file and visible result after replacement.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Font Loading Checklist

Font loading checks remove unnecessary requests and keep text useful while the chosen typeface arrives. Verify the visible transition as well as the network.

Remove unnecessary font files and variants

Load the typefaces the page actually uses.

Use this for: The representative public pages selected for this review.

Filter Network to font requests and compare the downloaded families, weights and styles with the visible page. Look for unused variants, duplicated providers and fonts loaded on templates that do not need them.

Ask the developer to remove unused requests or use suitable subsets while keeping the characters and languages your content requires. Check headings, body copy, bold text and any supported non-Latin content after the change. A smaller subset is not useful if it drops required characters.

Example

Hypothetical example: A service template downloads six font weights but uses two. Removing the unused variants reduces requests while preserving the design.

  1. List the font files requested by the template.
  2. Match them to actual families, weights and languages.
  3. Remove unused requests and retest representative text.
More help and optional notesHigh · Developer and designer
PassThe tested template downloads the necessary font resources without missing required characters or styles.
FailUnused variants remain or subsetting breaks real text.

If it fails: Correct the font inventory and subset coverage.

Retest: Inspect font requests and representative page content again.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Check readable text while fonts load

Balance early reading with stable layout.

Use this for: The representative public pages selected for this review.

Test a fresh font load on a slower connection and watch whether the text stays invisible, appears in a fallback or changes line wrapping when the brand font arrives. Ask the developer to choose an appropriate font-display strategy and a compatible fallback for the design.

Preload only a genuinely critical font when the trace supports it. Match the requested resource and configuration so the hint is reused. Do not assume font-display: swap prevents layout shifts; verify the actual fallback-to-web-font transition.

Example

Hypothetical example: A heading appears quickly with a fallback but gains a line when the web font arrives. A better-matched fallback improves stability while keeping the heading readable.

  1. Observe a fresh load with the font delayed.
  2. Adjust the display strategy and fallback where needed.
  3. Check wrapping, visibility and any critical font preload.
More help and optional notesHigh · Developer and designer
PassText remains useful during loading and the chosen font transition avoids the reproduced harmful delay or shift.
FailText disappears unnecessarily or the font strategy creates unresolved shifts and duplicate downloads.

If it fails: Revise the fallback, display choice or selective preload.

Retest: Retest fresh font loading and the visible layout.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Caching Checklist

Caching checks verify safe reuse, freshness and consistent releases. Public static assets and personalized pages need appropriate, different policies.

Check browser reuse for versioned static assets

Avoid downloading unchanged files on every normal repeat visit.

Use this for: The representative public pages selected for this review.

Choose a stable stylesheet, script or image and inspect its caching response headers. Reload normally with developer tools’ Disable cache option off, then check whether the browser reuses or revalidates the resource as intended.

Ask the developer to use suitable cache lifetimes for the asset’s update strategy. Long freshness is a good fit for content-versioned filenames only when changed content gets a new URL. Do not apply a long lifetime to a frequently replaced file at the same address without a refresh plan.

Example

Hypothetical example: A versioned stylesheet is downloaded on every repeat visit because its cache policy prevents reuse. The team corrects the policy and checks the next release’s filename.

  1. Inspect a static asset’s cache headers.
  2. Test a normal repeat visit with caching enabled.
  3. Match its freshness policy to the filename/versioning strategy.
More help and optional notesHigh · Developer or hosting administrator
PassUnchanged assets are reused or revalidated appropriately and updates have a reliable path.
FailRepeat visits redownload unnecessarily or a long-lived URL can trap old content.

If it fails: Correct the policy and content-versioning approach.

Retest: Test repeat use and a changed asset version.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Separate public caching from personalized responses

Improve delivery without exposing one visitor’s content to another.

Use this for: Sites using shared caches, CDNs or personalized responses.

Have the developer identify which responses are safe for a shared CDN cache and which depend on sign-in, account, cart or other personal state. Check the actual cache policy and variation behavior on those response types using approved test accounts.

Do not enable blanket public caching for every page to improve a score. Private and no-store directives serve different purposes, and cookies alone are not a complete shared-cache safety policy. The responsible developer should set the policy for the application and verify isolation.

Example

Hypothetical example: A public article can be shared-cache eligible, while an account page needs a different policy. The team tests both before changing the cache configuration.

  1. Separate public and personalized response types.
  2. Review their actual cache headers and variation rules.
  3. Verify approved test accounts cannot receive each other’s content.
More help and optional notesHigh · Developer or hosting administrator
PassThe cache policy improves eligible public delivery while preserving personalized-response isolation.
FailA blanket cache change could reuse personal content across visitors or serves the wrong variant.

If it fails: Stop the unsafe shared-cache behavior and have the responsible developer correct the policy.

Retest: Repeat isolated approved-account and public-page tests.

Not applicable: There is no shared caching or personalized response behavior in scope.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Test that releases replace stale content correctly

Keep speed gains from serving the wrong version.

Use this for: The representative public pages selected for this review.

After a template or asset change, compare a fresh visitor, a repeat visitor and the relevant public delivery layer. Check that the page references the current versioned assets and that cached HTML does not keep pointing to missing files.

Use the host’s documented refresh or invalidation path when needed. Verify the live page rather than assuming a purge request completed. Avoid clearing every cache routinely; target the stale resource or route and preserve a rollback option for the release.

Example

Hypothetical example: A cached article template references an old script that was removed during deployment. Keeping compatible assets or refreshing the HTML fixes the repeat-visitor error.

  1. Open the changed page as a fresh and repeat visitor.
  2. Check HTML and asset versions for consistency.
  3. Refresh the specific stale resource through the documented process.
More help and optional notesHigh · Developer or hosting administrator
PassFresh and repeat visitors receive a working, consistent release.
FailOld HTML and missing or mismatched assets break the live page.

If it fails: Correct the refresh/versioning sequence and retain compatible assets where needed.

Retest: Repeat both visit states after the correction.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Lazy Loading Checklist

Lazy loading checks defer suitable offscreen work while protecting critical media and crawler-accessible content.

Lazy-load offscreen media without delaying the main image

Save optional work while protecting the first useful screen.

Use this for: The representative public pages selected for this review.

Identify the image likely to become LCP and the other media visible at the initial viewport. Keep those necessary early resources from waiting on lazy loading. Use selective lazy loading for genuinely offscreen images or embeds where it improves the measured experience.

Reserve dimensions for the delayed media and scroll through the page to confirm it loads before it is needed. Native browser loading thresholds can fetch offscreen images in advance; a request before the image is visible is not automatically a failure.

Example

Hypothetical example: A theme lazily loads every image, including the hero. Removing that delay from the hero while keeping it for lower gallery images improves the intended loading sequence.

  1. Separate initial-screen media from lower-page media.
  2. Remove lazy loading from the critical image.
  3. Test delayed loading and reserved space while scrolling.
More help and optional notesHigh · SEO lead and developer
PassImportant early media loads promptly and deferred media appears reliably without unexpected movement.
FailThe hero waits unnecessarily or offscreen media remains blank when needed.

If it fails: Correct the loading assignment and reserved dimensions.

Retest: Retest initial loading and a complete scroll-through.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Keep important lazy-loaded content available to crawlers

Avoid making the only useful answer depend on a click.

Use this for: The representative public pages selected for this review.

After changing deferred loading, inspect the rendered page in a relevant Google test and compare its main content with the browser. Important text or images should not require a visitor to click, type or swipe simply to fetch them.

For a long listing, keep persistent crawlable pagination where needed rather than exposing later items only through an endless-scroll action. This check protects content access; it does not require loading every offscreen image immediately.

Example

Hypothetical example: A speed plugin delays the service description until an accordion click. The team restores its content to the rendered page and defers only suitable media.

  1. Compare the expected content with the rendered test output.
  2. Remove interaction-only loading for important search content.
  3. Check persistent links for later listing content.
More help and optional notesHigh · SEO lead and developer
PassThe rendered page retains important content and a crawlable route to later items where applicable.
FailAn optimization hides useful content behind an action the crawler will not perform.

If it fails: Repair the deferred-loading trigger or persistent pagination.

Retest: Repeat the rendering test and normal visitor journey.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Reduce Unnecessary CSS, JavaScript and External Work

Code optimization checks the actual rendering or execution cost and preserves the states visitors need. A smaller file alone does not prove a working improvement.

Reduce blocking CSS without breaking other page states

Shorten the styling path safely.

Use this for: The representative public pages selected for this review.

Inspect the stylesheets needed before the page can render and look for avoidable CSS import chains or very large shared bundles. Ask the developer to minify, split or remove genuinely unused styles where the trace shows useful savings.

Check multiple templates, responsive widths, open menus, validation errors and interactive states before deleting code marked unused in one recording. If critical CSS is inlined, maintain it through the build and test the rest of the stylesheet loading; stale or incomplete inline styles can create visual defects.

Example

Hypothetical example: A coverage recording labels menu styles unused because the menu was never opened. Testing that state prevents an overaggressive cleanup from breaking navigation.

  1. Identify the blocking stylesheet or request chain.
  2. Review unused coverage across relevant routes and states.
  3. Apply the scoped change and compare rendering.
More help and optional notesHigh · Front-end developer
PassThe selected CSS delay is reduced and the relevant layouts and states remain styled correctly.
FailA one-page coverage result causes missing styles or the blocking chain remains.

If it fails: Restore required styles and scope the optimization to the proven cause.

Retest: Recheck the affected templates and states after the build.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Load and execute only the initial JavaScript you need

Reduce startup work while preserving the visitor’s action.

Use this for: The representative public pages selected for this review.

Use the trace and bundle information to identify scripts that delay initial rendering or responsiveness. Ask the developer to remove unused dependencies, split optional features or defer noncritical execution where appropriate. Preserve the code needed for the first meaningful interaction.

Choose async or defer based on script dependencies and execution order rather than applying either to every file. Code splitting should keep optional work out of the initial path without adding a long new wait when the visitor opens that feature. Test both startup and first use.

Example

Hypothetical example: A booking widget’s full editor loads on every article even when no booking form is present. Loading it only on the relevant route removes unnecessary startup work.

  1. Identify costly scripts and the features they support.
  2. Remove or postpone work outside the initial need.
  3. Test startup and the first use of each deferred feature.
More help and optional notesHigh · Front-end developer
PassThe measured startup work is reduced and important interactions still work at the right time.
FailDeferral breaks execution order or moves the same unacceptable delay into a core interaction.

If it fails: Correct dependency ordering or the feature-loading boundary.

Retest: Repeat the initial load and affected interaction trace.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Review the cost of external widgets and tags

Keep useful features without unnecessary background work.

Use this for: The representative public pages selected for this review.

List externally supplied scripts and embeds such as chat, video, maps, analytics and marketing tags. Match each to a business owner and a current purpose, then inspect its network and execution cost. Compare a controlled test with a suspect feature disabled only in a safe test setup.

Remove obsolete integrations with the responsible owner’s approval, or load a useful feature when it is actually needed. Keep consent behavior and required measurement intact. A third-party script loaded asynchronously can still compete for bandwidth and main-thread time.

Example

Hypothetical example: Two old campaign tags load the same library twice. The marketing owner confirms which integration is still needed before the duplicate is removed.

  1. Inventory external scripts and their owners.
  2. Measure a suspect feature in a controlled comparison.
  3. Remove or reschedule unnecessary work with the appropriate owner.
More help and optional notesHigh · Developer and integration owner
PassThe chosen external-work change reduces the measured cost while preserving required functionality and consent behavior.
FailTags are removed blindly or async loading is treated as zero execution cost.

If it fails: Restore required behavior and narrow the change to the measured unnecessary work.

Retest: Retest the page, affected integration and relevant consent states.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Retest the Live Page

A speed change is complete only when the intended improvement survives deployment and the affected page still works.

Verify the complete page after the speed change

Check that the improvement survived deployment.

Use this for: The representative public pages selected for this review.

Repeat the original public-page tests after deployment, including first and repeat visits. Compare the resource or timing you changed, then use the page as a visitor: read the main content, open navigation and test the approved form or product action.

Check for new layout shifts, missing images, broken deferred features and stale assets. Keep the implementation result separate from field Core Web Vitals confirmation, which needs comparable real-user data. A faster diagnostic run does not promise a ranking increase.

Example

Hypothetical example: A smaller script bundle improves the load trace but prevents a product option from opening. The release needs a functional repair before the speed task is complete.

  1. Repeat the matched live tests and compare the targeted delay.
  2. Complete the affected visitor journey.
  3. Check the later field outcome through the Core Web Vitals workflow.
More help and optional notesHigh · Developer and QA reviewer
PassThe live change improves the intended diagnostic result without breaking content or the journey.
FailA score improves while the live page has a functional, content or visual regression.

If it fails: Repair or roll back the defective change, then repeat the same checks.

Retest: Review comparable field data after sufficient post-release visits.

Not applicable: No pages or features in the stated scope require this check.

Optional: the URL or template, what you found and the next action. Keep confidential data out of shared exports.

No result recorded yet.

Changing this result updates your checkmark and progress immediately. Notes are optional.

Reset this checklist?

This removes saved statuses, evidence and result dates for the checks on this page. Other checklists keep their records. Export a CSV first if you need a copy.

A useful result, every time

Check a task. Choose a result.

This is a manual SEO workflow. It helps you decide what to investigate and keep a usable record. It does not crawl your website or verify your answers automatically.

  1. Choose your scope. Pick the relevant task group and representative pages you are authorized to inspect.
  2. Run the stated check. Follow the steps, then choose your result. Click the square for Pass, or use the result menu for another status. You do not need to type anything.
  3. Assign the next action. Fix a problem when you can, or ask the right person for help. Optional notes can remind you what to do next. Check the result again after a fix.

What your progress means

Review progress counts passes and failures against applicable checks. Failed work still counts as reviewed. Blocked and untested checks remain unfinished. This is not an SEO score.

Untested
You have not checked this yet.
Pass
You checked it and it works as described.
Fail
You checked it and found something to fix.
Blocked
You need access, information or someone’s help.
Not applicable
This task does not apply to your website.

Export CSV downloads the checks currently shown and any optional notes. Print expands the instructions for that view. Saved results stay in this browser; there is no account sync.

Apply the workflow

What Should You Check After a Page-Speed Change?

Use the examples and related workflows below to separate implementation success, real-user outcomes and wider SEO checks.

What Does a Page-Speed Implementation Record Look Like?

A page-speed implementation record connects a real delay to a specific resource change and a repeatable public-page test. Keep the lab result and later field outcome in separate columns.

These hypothetical examples show the format. Use your own measurements; this page does not run a speed test or create a performance report when you export its checklist results.

Hypothetical implementation choices and verification.
FindingSpecific changeImmediate checkLater outcome
Phone card requests a 4000-pixel photo.Add suitable responsive image candidates.Inspect the downloaded candidate, bytes and visual quality.Review the relevant real-user loading trend.
Stylesheet discovered through an import chain.Expose the necessary stylesheet earlier.Compare request timing and all affected layouts.Review the same field population.
Repeat visitor receives old HTML with missing scripts.Correct asset retention and cache refresh order.Test fresh and repeat visits after deployment.Monitor the affected journey for recurring errors.

How Is Page Speed Different From Core Web Vitals?

Page-speed work changes how resources are delivered, processed and displayed. Core Web Vitals measure particular outcomes: loading through LCP, responsiveness through INP and visual stability through CLS.

A smaller download can help, but it does not establish that all three outcomes improved. Use the relevant metric and a matching test to confirm the result.

What Page-Speed Score Should You Aim For?

Aim to remove delays that affect your visitors and important journeys, then use repeatable measurements to verify the repair. A Lighthouse score summarizes a particular lab run; it is not a universal loading-time target or a guarantee about real users.

There is no single page-weight or total-load-time number that proves every website is good for SEO. Use the Core Web Vitals thresholds for those specific field metrics and keep diagnostic budgets appropriate to your site.

Should You Install a Speed Plugin or Add a CDN First?

Choose a plugin or CDN only after identifying the bottleneck and checking what your host already provides. The right tool depends on whether the problem is response delivery, image generation, unused code, rendering or a specific integration.

Test one controlled change at a time and keep a rollback path. Overlapping cache, minification or lazy-loading features can change behavior unexpectedly, so verify the actual page rather than assuming that more optimization settings are better.

Should You Remove Useful Content to Make the Page Faster?

Preserve the answer and the visitor’s main action. Start by removing unnecessary transfer, duplicate dependencies and avoidable processing, then consider a simpler presentation if it still serves the same need.

An optimization that hides important text from the rendered page, breaks navigation or makes an image unreadable is incomplete even if its performance score increases.

Which Checklist Should You Use After the Speed Fixes?

Choose the next workflow based on the result that still needs verification.

Continue With a Focused Technical Review

Use the workflow that matches the discovery, rendering or delivery issue you are investigating.

A clear process, with honest limits.

Follow the steps, choose the result that matches what you found, and check again after changes. Notes are optional. A completed check does not guarantee indexing, traffic or rankings. How the checklists work.