Lighthouse 100
is it still
realistic?
I hear the question in every kick-off: "Can you hit 100?" The honest answer: we can — but it's rarely worth it. Here are the numbers on why 95+ is the sweet spot and where the marginal benefit of perfectionism stops.
// What does Lighthouse measure?
// Why 95+ is enough
// The cost of the last 5 points
// What Google really wants
What Lighthouse actually measures.
Lighthouse is Google's audit tool that scores four areas: Performance, Accessibility, Best Practices, SEO. Each area gets a score from 0 to 100. Most clients see only the numbers and think: "The higher, the better." That's true, but with one important caveat.
The scores aren't linear. From 60 to 80 the improvements are often easy: image optimisation, caching, compression. From 80 to 95 it gets more demanding. From 95 to 100 it gets partly pathological — you optimise for milliseconds no one notices, and often sacrifice functionality for it.
Important to understand: Lighthouse is a synthetic test, not a measurement of real users. The score depends on the hardware of the testing device, the simulated connection and even the time of day. Two tests of the same page can vary by ±5 points.
Why 95+ is the sweet spot.
In our projects we consistently deliver Lighthouse scores of 95 to 100 — depending on the page, content and integrations. Reaching 95+ means in practice:
- Load time under 1.5 seconds on a 4G connection
- Core Web Vitals green — LCP < 2.5s, CLS < 0.1, INP < 200ms
- No render-blocking from large CSS/JS files
- Semantic HTML that screen readers interpret correctly
- All images in WebP/AVIF at appropriate sizes
- HTTPS, CSP headers, valid certificates
For Google, 95+ is effectively the same as 100. In ranking terms there's no measurable difference. The algorithm doesn't ask "what's the Lighthouse number", but "are the Core Web Vitals in the green". At 95+ they already are.
What it costs to chase the last 5 points.
The marginal benefit flattens drastically at 95. To get from 95 to 100 you often have to drop or rebuild things the user actually needs:
Google Analytics / Tracking
The GA4 snippet drags performance down by 2 to 3 points. Alternative: server-side tracking or Plausible — but that costs extra. Question: do you want 100 points or proper tracking?
External fonts
Google Fonts causes an extra DNS lookup. Self-hosting is the clean solution, but font licences then often become an issue — and sometimes the designs aren't interchangeable.
YouTube-Embeds
An embedded YouTube player can ruin the score by 5 to 10 points. The facade technique (a lite preview first, then the full player on click) solves it — but not every CMS supports it natively.
Interactive features
Chatbot widget, cookie banner, live chat — each of these features costs points. You can lazy-load them, but that needs custom code.
Our rule of thumb: the last 5 points cost as much development time as the first 50. Work out for yourself whether that serves your business goals.
Why Core Web Vitals do matter.
Lighthouse and Core Web Vitals are often confused. Lighthouse is the synthetic test, Core Web Vitals are the real measurements from Chrome user sessions (CrUX data). Google uses the latter for ranking, not Lighthouse.
The three current Core Web Vitals:
- LCP (Largest Contentful Paint): when does the largest element become visible? Target: < 2.5 seconds
- CLS (Cumulative Layout Shift): how much does the layout shift while loading? Target: < 0.1
- INP (Interaction to Next Paint): how fast does the page respond to clicks? Target: < 200ms (the replacement for FID since March 2024)
You'll find the Core Web Vitals in Google Search Console under "Experience → Core Web Vitals". There you see how users really experience your page — not how an audit tool simulates it.
Our recommendation.
Set yourself these realistic targets:
- Lighthouse Performance: 95+ on the homepage, 90+ on all subpages
- Lighthouse SEO: 100 — it costs little and Google values it
- Lighthouse Accessibility: 95+ — increasingly relevant legally too (BFSG)
- Best Practices: 95+ — usually automatic with a clean code base
- Core Web Vitals: all three green — the only thing Google really uses for ranking
Anyone selling you Lighthouse 100 as a marketing argument has either built little, left a lot out, or has no plan. Anyone who explains why 95+ with working tracking and a full feature set is the better deal knows what they're doing.
95 is the new 100.
Lighthouse is a useful indicator, not a competition. The score is a means, not the goal. The real goal is: fast pages, happy users, good ranking, working business processes.
In our projects we deliver a guaranteed 90+ — and build in tracking, chatbots, font embedding, cookie banners as normal. Because it matters to the client. We skip the magical last 5 points — and thereby save our clients real money.