Core Web Vitals help answer three practical questions: does the main content appear promptly, does the page respond when someone interacts, and does the layout stay stable? They are useful for finding frustrating website experiences, but they are not a standalone ranking formula.
The current metrics are LCP, INP and CLS. Diagnose the metric that is failing, identify the affected page templates and test the relevant fix. A perfect lab score is not a guarantee of higher rankings, more enquiries or a good experience for every visitor.
What are the current Core Web Vitals?
The three metrics cover loading, responsiveness and visual stability. Google's Web Vitals documentation sets the following good-experience thresholds. Assess them at the 75th percentile, separately for mobile and desktop: all three need to meet the recommended thresholds to pass.
| Metric | What it measures | Good threshold | A symptom to investigate |
|---|---|---|---|
| Largest Contentful Paint (LCP) | When the largest visible image or text block renders. | 2.5 seconds or less | The page opens, but its main content takes too long to appear. |
| Interaction to Next Paint (INP) | Responsiveness to interactions throughout a visit. | 200 milliseconds or less | A menu, filter or button feels slow to respond. |
| Cumulative Layout Shift (CLS) | Unexpected visual movement in the layout. | 0.1 or less | Content moves while someone is reading or trying to tap. |
FID is no longer a Core Web Vital. INP replaced First Input Delay on 12 March 2024. An older report or checklist that still treats FID as the current responsiveness metric needs updating. See Google's INP transition announcement.
How do Core Web Vitals affect SEO?
Google says its ranking systems use Core Web Vitals, but good scores do not guarantee top rankings. Relevance and the overall page experience still matter. Treat performance work as part of a useful website, not a replacement for clear content or an automatic explanation for a traffic change. Google's page experience guidance explains this distinction.
A failing metric alone does not prove why a particular page ranks poorly. Equally, a fast page can still answer the wrong question, contain unsupported claims or have a confusing enquiry process. Our SEO audit guide explains why identifying the actual bottleneck comes before choosing a fix.
Field data and lab scores answer different questions
PageSpeed Insights combines real-user information from the Chrome User Experience Report with a Lighthouse lab test. Field data reflects a rolling 28-day collection period; a lab run captures a controlled test. Check whether the real-user report describes the specific URL or the wider origin, and keep mobile and desktop results separate.
A Lighthouse performance score is not the same as passing the field Core Web Vitals assessment. A single run can also vary. If real-user data is unavailable, that means there is insufficient eligible data for that report, not that the page passed or failed. Google's PageSpeed Insights documentation describes the data sources and limitations.
Keep a simple measurement record: tested URL, device, date, report scope, failing metric and test conditions. That helps prevent comparing a desktop lab result with a mobile field result and calling the difference an improvement.
Turn the failing metric into a specific investigation
| Problem | First checks | How to validate |
|---|---|---|
| Slow main content | Identify the LCP element. Inspect server response, resource discovery and rendering delays. | Repeat comparable tests on the affected template; then monitor field data. |
| Delayed interactions | Reproduce the slow action and inspect main-thread work, event handlers and third-party scripts. | Test the same action on representative mobile hardware, not only a fast desktop. |
| Unexpected page movement | Inspect image dimensions, reserved embed space, font changes and inserted banners. | Check the loading sequence and interactions, not just a static screenshot. |
LCP: find what delays the main content
Start with the actual LCP element rather than assuming it is the hero image. For an image, inspect when its request starts, its transfer size and when it becomes visible. Do not lazy-load the LCP image. Make it discoverable early, use an appropriately sized asset and consider fetch priority where justified. Compressing the file alone may not solve delayed discovery or blocked rendering.
Investigate slow server responses separately from front-end delays. Follow Google's LCP optimisation guide for a breakdown of the loading stages. Avoid enabling every optimisation at once: you need to know which change helped and whether anything broke.
INP: test real interactions
Try the actions that matter on the page: opening navigation, filtering products, expanding an accordion or submitting a form. Long JavaScript tasks can delay the response. Review expensive handlers, unnecessary work and third-party scripts; split or defer work where appropriate. A fast initial load does not establish that later interactions are responsive.
Google's INP optimisation guide covers input delay, processing and presentation. After a script change, also verify analytics, consent handling and form behaviour. Removing an essential function to improve a number is not a successful fix.
CLS: reserve space before content arrives
Set image dimensions or a stable aspect ratio, reserve appropriate space for embeds and avoid inserting unexpected content above what the visitor is reading. Review font swaps and dynamically added banners. Some layout shifts happen after loading, so test a normal browsing session rather than relying only on the initial viewport.
Use Google's CLS optimisation guide to investigate the source. Check narrow screens, long headings and delayed assets as well as the desktop layout.
A practical example: fixing a service-page template
Illustrative workflow, not a measured client result: a business finds that several service pages share a slow main image, a delayed mobile menu and a shifting enquiry banner. Because the pages use one template, it makes sense to investigate the shared components before editing every page individually.
- Record one representative page and reproduce each symptom.
- Confirm which element is LCP and whether its loading is unnecessarily delayed.
- Record a trace while opening the mobile menu and identify the expensive task.
- Reserve space for the enquiry banner so its appearance does not move surrounding content.
- Test the revised template on several pages with different image sizes and heading lengths.
- Check navigation, enquiry forms and tracking before releasing the change.
This process produces a testable improvement plan. It does not establish a percentage gain in traffic or conversions. Those outcomes need their own evidence and reporting period.
Prioritise fixes without chasing a perfect score
- Start with important journeys: service enquiries, product selection and checkout deserve attention when visitors encounter delays.
- Group shared templates: a common component can affect many URLs, but verify that the same problem actually occurs across them.
- Fix observed causes: do not upgrade hosting or remove plugins merely because a generic checklist says to.
- Preserve functionality: test forms, menus, accessibility and measurement after performance changes.
- Separate immediate tests from field trends: the rolling real-user window takes time to reflect a release.
For broader delivery and asset considerations, read our website speed and performance guide. This article focuses on interpreting the three Core Web Vitals and choosing the investigation each one calls for.
How to verify an improvement
Keep the original measurements and record exactly what changed. Repeat tests under comparable conditions and check more than one run. Test important interactions manually and review real-user trends once enough post-change data is available. Do not attribute every traffic movement to the release: content, search demand and other website changes can affect results too.
Use the Search Console troubleshooting guide for the wider reporting context. A Core Web Vitals investigation and a crawl or canonical investigation answer different questions; one should not be treated as proof of the other.
Build performance into the website process
Performance is easier to maintain when image sizing, interaction testing and stable layouts are part of the design and development process. Add checks for new templates and third-party integrations instead of treating optimisation as a one-time plugin installation.
Brainvative's website design and development services and SEO services address different parts of that work. A useful brief begins with affected URLs, reproducible symptoms and agreed validation steps, not a promised ranking position.

