Yes, website speed affects conversion rate, and the relationship is stronger than most people expect. This isn't a marginal factor buried among dozens of others — it's one of the more directly measurable levers in conversion-rate optimization. If you've ever wondered whether a faster site is worth the engineering time, this post lays out exactly how website speed affects conversion rate and why the effect is so consistent across studies.
Does website speed really affect conversion rate?
Yes. Every major study that has measured it finds conversion rate falling as load time rises. In Portent's 2022 analysis of more than 100 million page views, B2B lead-generation pages that loaded in one second converted at roughly 40%, falling to about 29% at three seconds — a one-second page converted three times as well as a five-second page and five times as well as a ten-second page. For B2C e-commerce, one-second pages converted 2.5 times as well as five-second pages. And Deloitte's "Milliseconds Make Millions" study found that a 0.1-second mobile speed improvement lifted retail conversions by 8.4%.
That's the direct answer, and it's worth reading twice, because the drop isn't a cliff at some arbitrary threshold — it's a steady slide that starts immediately. There's no "safe" load time where speed stops mattering; every additional second costs you something.
Why does a few seconds of delay change buying behavior?
It comes down to expectation and impatience, both of which have gotten less forgiving over time, not more.
- Expectations are set by the fastest sites people use. In a 2009 Forrester survey commissioned by Akamai, 47% of online shoppers already expected a page to load in two seconds or less — and expectations have only risen since. Your site isn't competing against your last redesign; it's competing against whatever app a visitor used five minutes earlier.
- Abandonment happens before your message lands. The same survey found 40% of shoppers will wait no more than three seconds before abandoning a site, and Google's 2018 mobile benchmarks found the probability of a visitor bouncing rises 123% as load time goes from one second to ten. Your headline, your offer and your trust signals don't matter if the page hasn't rendered yet.
- Mobile networks amplify the problem. A large share of commercial traffic is mobile, often on cellular connections with inconsistent speed, so a page that loads acceptably on office wifi during development can be meaningfully slower for real visitors.
Is this just about e-commerce, or does it apply to any site?
It applies broadly, though the stakes vary by business model. For online stores the link is close to direct: Deloitte's study measured higher conversions and larger average orders from faster retail pages, and former Amazon engineer Greg Linden has described internal Amazon tests from 2006 in which every added 100 milliseconds cost about 1% of sales (widely quoted, though never published as a formal study). For lead-generation sites, service businesses, and SaaS signup flows, the mechanism is the same even though the "conversion" being measured is a form fill or a demo request instead of a purchase: slower pages mean fewer completed actions, full stop.
What does Google's own ranking signal tell us?
Google factors page experience, including Core Web Vitals, into how it ranks pages — a signal Google has published and maintained for several years now. That matters for conversions in two ways. First, a slow site can rank lower, meaning fewer visitors reach you in the first place. Second, the same technical issues that hurt your Core Web Vitals score — a slow Largest Contentful Paint, layout shift, sluggish interactivity — are usually the exact same issues suppressing your conversion rate. Speed and SEO aren't two separate projects; they're the same fix seen from two angles.
"A few seconds doesn't really matter" — is that true?
No, and it's one of the most persistent myths in web performance. The data says otherwise at every threshold measured:
| Claim | What the data actually shows |
|---|---|
| "Users won't notice a few extra seconds" | Google's mobile data shows bounce probability rising 123% from a one-second to a ten-second load |
| "Speed only matters for huge e-commerce stores" | Portent's data shows the same pattern for B2B lead-generation sites, where a one-second page converted 3x as well as a five-second one |
| "Our site feels fast enough to us" | Developer machines and office connections are rarely representative of real visitor conditions |
| "Speed and design are a tradeoff" | Most slowdowns come from unoptimized assets and scripts, not necessary design choices |
Does this hold true across industries, or just for retail?
The underlying research skews toward e-commerce because purchases are the easiest conversions to measure cleanly, but the mechanism isn't retail-specific. Any site asking a visitor to complete an action — filling out a quote request, booking a consultation, starting a free trial, submitting a contact form — is subject to the same abandonment pressure from slow load times. A B2B services site that takes five seconds to load its contact page is losing exactly the kind of impatient, high-intent visitor who was ready to reach out, simply because the page didn't render fast enough to let them.
If anything, the cost can be more painful for lead-generation businesses than for e-commerce, since a lost lead often represents a much larger potential deal value than a single lost cart. The visitor doesn't leave a trace of what they almost did — they just don't show up in your CRM, and there's no abandoned-cart email to win them back with.
Does speed matter differently for new visitors versus returning ones?
Yes, somewhat. First-time visitors have no prior relationship with your brand and no accumulated trust to draw on, so a slow, stuttering first impression does disproportionate damage — there's nothing offsetting it yet. Returning visitors who already trust your business may tolerate a slow page more, but that tolerance isn't unlimited, and repeat friction is still a real driver of long-term churn even among visitors who like what you sell. Either way, speed works in your favor, never against it — there's no visitor segment for whom a slower page converts better.
How do you know if your site's speed is actually hurting conversions?
Guessing isn't useful here — the fix depends on knowing your actual numbers. Run a real audit against your live URL, look at your Largest Contentful Paint and overall load time, and compare that against the conversion-by-load-time pattern above. If you're loading in three-plus seconds on mobile, you likely have measurable, recoverable conversion rate sitting on the table, and it's usually addressable with targeted fixes to images, scripts, and hosting rather than a full rebuild. For a deeper breakdown of specific causes, see Why Is My Website So Slow?.
Two related reads: PageSpeed Insights vs GTmetrix vs Pingdom, and what slow load times cost a business.
