Site Speed and Rankings: How Much Speed Really Matters

Speed is a ranking factor. It’s also a conversion factor. A slow site can still rank if the content is uniquely strong — and still lose the lead because the form never became usable. Treat speed as both: Page Experience, and the first impression a buyer gets on a phone.

The Direct Impact

Google has used page experience — including Core Web Vitals — as a ranking input for years. It is not the strongest signal. Relevance and links still dominate. It is a tie-breaker you control, and it is binary enough that “good” vs “poor” shows up in Search Console.

Core Web Vitals

Metric	Good	What usually breaks it
LCP	Under 2.5s	Huge hero image, slow TTFB, render-blocking CSS
INP	Under 200ms	Heavy JavaScript, third-party tags, long tasks
CLS	Under 0.1	Images without dimensions, injected banners, web fonts

A page can rank with a poor LCP. It ranks more easily, and keeps more of the clicks it earns, when those three are in range. Measure field data in PageSpeed Insights when Chrome UX Report has enough traffic; use lab tests to debug. See Core Web Vitals for B2B for which report to trust.

Mobile

Indexing is mobile-first. The phone test is the real test. Desktop-only “90 Lighthouse” while mobile LCP sits at 5s is a false win. Page Experience is evaluated on the URL Google actually uses — the mobile version.

Crawl efficiency

Faster responses let Googlebot fetch more URLs in the same crawl budget. That matters on large catalogs and ecommerce facet trees. A 2-second TTFB means fewer product or article URLs get recrawled after you update them. Speed doesn’t “unlock indexing” by itself, but it stops the crawler from stalling on templates that time out.

The Indirect Impact

Rankings are only half the math. The other half is what users do after the click.

Bounce rate

People leave slow pages. Google does not publish a bounce-rate ranking factor. What you can observe: if LCP is 6s, a large share of sessions never see the H1. Those sessions don’t convert. That’s lost demand, whether or not position 4 holds.

Time on page

Fast pages get read. Slow pages get abandoned mid-scroll. Engagement is messy as a ranking theory; as a business metric it’s obvious. If the spec table is below a 3MB hero, buyers never reach it.

Conversion

Industry tests have shown conversion dropping as delay rises — often in the range of several percent per extra second on landing pages. You don’t need a branded case study to act on that. If rankings stayed flat and the site got 1s faster, you would still want the faster site. Forms, configurators, and quote CTAs are where INP shows up: a 400ms tap delay feels broken.

What to Fix First

Don’t start with a new animation library. Start with the request path.

1. Server response (TTFB)

Target TTFB under 600ms. Hosting, PHP/Node work, database queries, and a CDN in front of static assets. Cache HTML where the page is public and stable. Origin in one region with buyers in another is a CDN problem, not a “compress the logo” problem.

2. Render-blocking CSS and JS

Anything in <head> that must download before first paint delays LCP. Inline critical CSS for the first screen. Defer or async non-critical JS. Kill tags you can’t name a job for — leftover chat widgets and three analytics pixels are a common INP tax.

3. Images

Usually the largest contentful element. Resize to display size, use WebP, lazy-load below the fold, never lazy-load the hero. Walk through technical SEO foundations before you rewrite copy.

How much it matters

Speed will not outrank a better answer on a competitive term. It will keep you in the running when two pages are close, and it will stop you from paying for clicks that bounce. Fix TTFB, then the critical path, then images. Recheck field CWV in two to four weeks after a template change.

---

We build sites that hit 90+ Lighthouse. Plan your rebuild.