Page Experience is a ranking signal
Google explicitly uses Core Web Vitals as a ranking signal. Pages with green CWV outrank equivalent pages without it, all else equal.
Page Speed Optimization
Page Speed Optimization for Real Devices and Real Users
A SeoMata speed engagement gets your Core Web Vitals green on a mid-range Android over 4G — not just on a desktop in the office. We measure, fix and lock in performance budgets so the gains stay green after every release.
<2.5s
LCP target
real-device, 75th percentile
<200ms
INP target
on key interactions
<0.1
CLS target
no layout shift after first paint
90+
Lighthouse perf
on production URL
What "page speed optimization" actually means
Page speed optimization at SeoMata is not "we ran PageSpeed Insights and applied the recommendations". It is a structured audit + fix + lock-in process built around the way Google actually measures Core Web Vitals — Field data from real users on real devices, weighted at the 75th percentile.
We work on Next.js, WordPress, Shopify, Magento and custom stacks. Most engagements deliver green CWV on production within 4–8 weeks and ship a regression-blocking CI check so the gains do not erode after the next release.
4–8 wk
Typical engagement
audit + fix + monitor
5 pts
Lighthouse regression
budget enforced in CI
0
"It was fast on my laptop" answers
we measure on real devices
Why
Six reasons page speed pays back
Conversion drops as LCP rises
Every additional second of LCP correlates with roughly a 7% drop in conversion rate. Real, measured, repeatable.
Ad spend gets cheaper
Quality Score on Google Ads factors in landing-page experience. Slow pages literally cost more per click for the same position.
SEO signals stack
Green CWV pages get faster indexing, more frequent recrawl, and stronger AI-search citation eligibility. Speed is one of the smaller-effort levers with compounding gain.
Mobile is most of your traffic
58–82% of organic traffic on service-business sites is mobile. If the site is slow on mid-range Android, most of your traffic is having a bad time.
Maintenance cost drops
Slow sites often = poorly architected sites. The fixes that ship CWV also fix the brittleness causing your engineers nights and weekends.
Targets
Targets we hold on every engagement
Google's thresholds and ours. We aim for "Good" across all three on the 75th percentile of Field data — not Lab data.
| Google "Good" threshold | SeoMata target | |
|---|---|---|
| LCP (Largest Contentful Paint) | < 2.5s | < 2.0s p75 on mid Android 4G |
| INP (Interaction to Next Paint) | < 200ms | < 150ms p75 on representative interactions |
| CLS (Cumulative Layout Shift) | < 0.1 | < 0.05 — no shifts after FCP |
| TTFB (Time to First Byte) | < 800ms | < 300ms via edge + cache |
| Lighthouse Performance | > 50 = green | > 90 in CI |
| Total JS budget (above-fold) | Not specified | < 170kb gzip on key templates |
When a target cannot be met (e.g. heavy third-party requirement), we document why and propose mitigations rather than silently shipping a red number.
Scope
Six things we actually tune
Image pipeline
WebP/AVIF, responsive sources per breakpoint, eager only on LCP, explicit width/height to lock CLS.
JS bundle
Tree-shake, code-split, dead-code removal, route-level prefetch budget, dynamic import where it pays.
Third-party scripts
Audit every chat / analytics / A/B / heatmap script; defer or remove anything that touches LCP path.
Fonts
Self-host, `font-display: swap`, `size-adjust`, subset, preload only the critical face.
Edge + cache
Vercel/Cloudflare edge for HTML, ISR or full-page cache where appropriate, smart cache invalidation.
CSS critical path
Inline critical CSS, defer the rest, eliminate render-blocking link tags, kill unused Tailwind / Bootstrap classes.
Process
Six-stage speed engagement process
Baseline measurement
Field data from CrUX + RUM (PostHog or web-vitals.js), Lab data from Lighthouse + WebPageTest. Output: baseline report + prioritised opportunity list.
Image + font pass
Image pipeline rebuild, font strategy tuned. Usually accounts for 40-60% of LCP improvement.
JS audit + defer
Bundle analysis, third-party defer, dynamic import. Accounts for most INP improvement.
Edge + cache
Edge HTML caching, ISR where appropriate, smart invalidation. Drops TTFB and improves repeat-visit metrics.
CI lock-in
Lighthouse CI added to PR pipeline, regression threshold set. Speed wins do not erode after the next release.
Monitor + hand off
Field data dashboard set up, weekly check for 30 days, then handed off or kept on retainer.
Tools
Tooling we use to measure and enforce
Chrome UX Report (CrUX)
Field data — what real Chrome users experienced. The source of truth for ranking signal eligibility.
Lighthouse + Lighthouse CI
Lab data + regression budget enforced in PR pipeline.
WebPageTest
Filmstrip + waterfall + multi-location, multi-throttle testing. Essential for diagnosing edge cases.
PostHog / web-vitals.js
Per-route Field data from your own users, not just CrUX. Find regressions in 24h, not 28 days.
Bundle analyzer
Next.js Bundle Analyzer / Source Map Explorer / Vite Bundle Visualizer. Surfaces JS bloat in seconds.
Sentry Performance
Transaction-level performance tracing for slow APIs, slow renders, and N+1 surprises in production.
Portfolio
Recent speed wins
Scope
Standard scope of a speed engagement
Baseline measurement
CrUX + Lab data + RUM if present. Documented benchmark.
Image pipeline rebuild
AVIF/WebP, responsive sources, LCP eager-load, explicit dims.
Font strategy
Self-host, swap, size-adjust, subset, critical preload.
JS bundle audit + defer
Tree-shake, code-split, dynamic import on heavy components.
Third-party script defer
Audit and defer or remove non-essential third-party.
Edge + cache strategy
Edge HTML, ISR where it pays, smart invalidation.
CI lock-in
Lighthouse CI on the PR pipeline with regression budget.
30-day monitoring
PostHog or equivalent dashboard, weekly check, hotfix if regression appears.
Common page-speed questions
How long until we see results?
Lab data improves within 4 weeks of starting work. Field data — what counts for the Google ranking signal — moves on a 28-day rolling window so meaningful CrUX improvement shows in week 6–8.
Will this break the site?
No. Changes go through PR review on a staging URL. We do not deploy directly to production. Visual regression tests on key templates ensure design integrity.
Can you work on our existing tech?
Yes — Next.js, WordPress (with caveats), Shopify, Magento, Squarespace (limited), Webflow (limited), custom stacks. Some platforms have ceilings we will be honest about up front.
Are these the same fixes as your competitors offer?
The fixes are well-known. The difference is in the discipline — we measure with Field data, we lock in with CI, and we explicitly hand off the runbook so your team can maintain the gains.
What if the issue is the hosting?
We will tell you. About 20% of speed engagements end with a hosting recommendation as part of the fix (Cloudflare in front of WP, Vercel for Next.js sites currently on a slow VPS, etc.).
Do you do ongoing performance retainers?
Yes. About half our speed clients move to a retainer after the initial engagement so someone owns the dashboard and addresses regressions before they hurt traffic.
Ready to scope a speed engagement?
Start with a free 30-minute speed audit. We will measure Field + Lab data on your top templates and write back with a prioritised opportunity list and rough scope — usually within 3 business days.
Related services
Explore related services
Technical SEO
Connect design, development, and SEO into one growth program.
Web Development
Connect design, development, and SEO into one growth program.
Website Maintenance
Connect design, development, and SEO into one growth program.
Free SEO Audit
Connect design, development, and SEO into one growth program.
Next step
Get a Free Page Speed Audit
Request a free page speed audit. SeoMata will benchmark your Core Web Vitals against category leaders, highlight the technical bottlenecks slowing your highest-traffic pages, and prioritize fixes by impact.