Technical SEO
Core Web Vitals Optimisation
Pass Core Web Vitals on real user data, not just in a lab test.
Most Core Web Vitals work optimises the wrong number. A perfect Lighthouse score is a lab result on a simulated device, while Google ranks on field data from real Chrome users. Those two figures routinely disagree, and only one of them counts.
You probably need this if
- Lighthouse shows 100 but Search Console still reports failing URLs
- Interaction to Next Paint regressed after adding a tag manager
- Layout shifts as images and ads load in
- Largest Contentful Paint is dominated by a hero image or web font
What you get
- Field data analysis from CrUX, segmented by device and template
- LCP element identification and optimisation per template
- INP profiling to find the specific handlers blocking the main thread
- CLS elimination through explicit dimensions and font loading strategy
- Regression monitoring so gains are not lost at the next deployment
How we approach it
-
Start with field data
Pull real user metrics from CrUX and segment by template, since one bad template can fail the whole origin.
-
Fix LCP at the source
Identify the actual LCP element per template and address it directly, usually through image format, preloading and render-blocking removal.
-
Profile INP properly
Trace long tasks to the scripts responsible, which in practice is very often third-party tags rather than your own code.
-
Lock in the gains
Add performance budgets to the build so a future deployment cannot silently undo the work.
Core Web Vitals: common questions
Is Lighthouse score a ranking factor?
No. Lighthouse is a lab tool producing a diagnostic score on a simulated device. Google uses field data from real Chrome users, reported in Search Console and CrUX. You can score 100 in Lighthouse and still fail Core Web Vitals.
How much do Core Web Vitals affect rankings?
They are a genuine but modest signal, and act mainly as a tiebreaker between pages of comparable relevance. The stronger commercial case is conversion: faster pages convert better, which is usually worth more than the ranking effect.
What replaced First Input Delay?
Interaction to Next Paint replaced FID in March 2024. INP is considerably harder to pass because it measures the full latency of all interactions through the page lifecycle, rather than only the delay on the first one.
Related services
Find out what is actually holding you back
A short conversation and a look at your site is usually enough to tell you where the real problem is. No obligation, and no pressure to sign anything.