Most teams still treat page speed as a developer's chore, something you fix before a Google audit and forget about until the next one. The 2026 data says that's backwards. The Chrome UX Report's May 2026 release shows barely more than half of all sites passing Core Web Vitals, and every recent case study on the record shows the same pattern: the conversion damage is worst exactly where the money changes hands, at the cart, the checkout form and the payment step, not on the homepage hero banner everyone obsesses over.
Page speed now moves conversion more than it moves rankings. In 2026, checkout and form pages lose the most to slow interactions, measured by Interaction to Next Paint, not slow initial loads. Fix INP on your cart and checkout first, add speculative prerendering where the browser supports it, and treat one-tap payment options as a speed shortcut, not a replacement for the underlying fix.
Is page speed really a conversion problem, or is that an SEO talking point?
It's both, but the conversion number moves faster and further than the ranking number ever does. The clearest evidence is the "Milliseconds Make Millions" research, commissioned by Google and run by Deloitte across 37 leading European and American retail sites and more than 30 million sessions. It found that a 0.1 second improvement in mobile speed lifted retail conversion by 8.4% and average order value by 9.2%, with the biggest single jump, 9.1%, happening between the product page and add-to-basket step. That study is a few years old now, but the newer, single-brand case studies published on web.dev in 2026 tell the same story with bigger numbers: Rakuten 24 saw a 53% increase in revenue per visitor and a 33% lift in conversion rate after improving LCP. Vodafone Italy improved LCP by 31% and sold 8% more. RedBus improved INP specifically and saw a 7% increase in sales. None of those are ranking metrics. They're straight revenue outcomes tied to how a page feels under a real thumb.
What does the 2026 Core Web Vitals data actually show?
The latest Chrome UX Report release, covering May 2026 data, found that 55.9% of tracked origins pass all three Core Web Vitals, a small regression of 0.8 percentage points from April. Broken out by metric, roughly 68.6% of origins pass LCP, 81.3% pass CLS, and 86.6% pass INP. That third number sounds reassuring until you separate desktop from mobile: desktop pages pass Core Web Vitals at a noticeably higher rate than mobile pages, and INP is the metric that fails most often industry-wide, with close to 43% of sites missing the 200 millisecond threshold, mostly on mobile devices with weaker processors. Lab data shows why: median Total Blocking Time sits around 67 milliseconds on desktop and over 1,200 milliseconds on mobile. That gap is where checkout forms fall apart, one laggy tap at a time.
Where does speed actually cost you money: the homepage or the checkout?
The checkout, and it isn't close. Baymard Institute's ongoing checkout usability research, based on 50-plus studies, puts average cart abandonment at 70.22%. Separately, Baymard's audits find 65% of ecommerce checkouts rate "mediocre" or worse on usability, with only 2% rated "good." Mobile now drives more than 72% of ecommerce traffic but converts at less than half the desktop rate, and unexpected costs still account for roughly 48% of abandonments. Speed compounds every one of those problems. A shopper filling in a laggy address field on a slow mobile connection is already primed to bail; a 500 millisecond delay on the next tap is often the last straw, not the first one. This is also why generic homepage speed audits miss the point: a fast homepage with a janky payment form is a fast way to lose the sale you already won. If your checkout also carries upsell logic or shipping calculators, as most Shopify AOV setups do, every extra script on that page raises the odds it fails INP before it fails anything else.
| Case study | What changed | Metric fixed | Result |
|---|---|---|---|
| Rakuten 24 | LCP optimization, A/B tested | LCP | +53% revenue per visitor, +33% conversion rate |
| Vodafone Italy | LCP improved 31% | LCP | +8% sales |
| RedBus | Interaction responsiveness improved | INP | +7% sales |
| Ray-Ban | Speculative prerendering on product tiles | LCP, navigation speed | +101% mobile conversion, +156% desktop conversion |
| Tokopedia | LCP reduced 55% | LCP | +23% average session duration |
| iCook | CLS improved 15% | CLS | +10% ad revenue |
These are individually documented case studies published on web.dev, not a blended industry average, so treat the exact percentages as directional for your own site rather than a guarantee. What's consistent across all of them is the size of the effect relative to how small the technical fix usually is: shaving fractions of a second off a load or an interaction, not rebuilding the site.
What is INP and why does it matter more for checkout than LCP?
LCP, Largest Contentful Paint, measures how quickly the biggest visible element loads, which is the right thing to obsess over on a landing page where the first impression is everything. INP, Interaction to Next Paint, measures how long the page takes to visually respond after someone clicks, taps or types, and a good score is under 200 milliseconds at the 75th percentile of visits. A checkout page can have a perfectly fast LCP and still fail badly on INP, because the damage happens later: a shipping-method dropdown that hangs for half a second, a coupon field that doesn't acknowledge a keystroke, a "place order" button that sits there unresponsive while a payment script initializes. Those are INP failures, and they land at the single worst possible moment, while someone is actively trying to give you money.
Stop auditing speed page by page and start auditing it step by step through the funnel. Run PageSpeed Insights or a real-user monitoring tool specifically on your cart, checkout and lead-capture form, not just the homepage, and prioritize INP fixes on those pages before anything else, since it's the most commonly failed Core Web Vital and the one directly tied to form abandonment. A homepage that scores 95 and a checkout that scores 40 is not a passing site, it's a leak with good PR.
Is the Speculation Rules API worth adopting in 2026?
For the right site, yes, and it's the most concrete new tool in this space. The Speculation Rules API lets a browser prerender a page a visitor is likely to click next, so the next page feels instant by the time they land on it. Ray-Ban's documented case, published on web.dev, used it to prerender product pages triggered by hover on desktop and by the most-clicked tiles on mobile, cutting LCP by 43% and lifting mobile conversion 101% and desktop conversion 156%. That's a genuinely large result from a browser feature, not a redesign. The catch is support: it works in Chrome and Edge, ships in Safari 26.2 but disabled by default, and Firefox hasn't shipped it yet, though it has a positive standards position for the prefetch portion. Unsupported browsers just ignore the tag, so it's safe to add as an enhancement. It is not, on its own, a reason to skip fixing the underlying page weight and script execution that's slowing things down for everyone else.
Do one-tap payments solve the speed problem instead?
They solve a different, related problem: they remove the slowest part of checkout, manual data entry, rather than fixing page performance. Shopify's own reporting shows Shop Pay converting meaningfully higher than standard guest checkout for eligible shoppers, and PayPal's one-touch option shows a real lift among the roughly one third of US shoppers who already have an account saved. Stripe Link has closed much of that gap for stores not on Shopify. These tools are worth enabling. But they mask a slow site for the fraction of buyers who use them and do nothing for everyone else typing a card number into a laggy form. Treat one-tap payment as a genuine conversion win layered on top of a fast checkout, not a substitute for one.
If you're on Shopify, enable Shop Pay and keep guest checkout as the visually dominant option, since Baymard's audits still find half of ecommerce sites burying it behind account creation. If you're on a custom stack, add Stripe Link or an equivalent before you consider a full checkout rebuild. It's a lower-cost move with a faster payback, and it buys you time to fix the INP problems underneath it properly.
Frequently asked questions
Does page speed actually affect conversion rate, or just SEO ranking?
Both, but the conversion effect is larger and faster to show up than the ranking effect. Case studies published on web.dev show LCP fixes producing conversion lifts of 8 to 156 percent, and a Deloitte and Google study across 37 retail sites found a 0.1 second speed improvement lifted retail conversion 8.4 percent. Treat speed as a revenue metric first and an SEO metric second.
What is INP and why does it matter more for checkout than LCP?
INP, Interaction to Next Paint, measures how long a page takes to visually respond after a click, tap or keystroke, and a good score is under 200 milliseconds. LCP measures how fast the first view loads, which matters most on landing pages. INP matters most on checkout, because every field, dropdown and button press on a payment form is an interaction that can lag.
What percentage of websites currently pass Core Web Vitals?
In the May 2026 Chrome UX Report release, 55.9 percent of tracked origins passed all three Core Web Vitals, down 0.8 points from April. INP is the most commonly failed metric, with roughly 43 percent of sites missing the 200 millisecond threshold, mostly on mobile.
What is the Speculation Rules API and is it worth using?
It is a browser feature that lets a site prerender the next page a visitor is likely to click, so the page feels instant when they land on it. Ray-Ban's case study on web.dev reported a 101 percent lift in mobile conversion and 156 percent on desktop after adopting it. It currently works in Chrome and Edge, ships disabled by default in Safari 26.2, and is not yet supported in Firefox, so it should be added as an enhancement, not a dependency.
Do one-tap payment options like Shop Pay or PayPal actually replace speed work?
They remove the slowest part of checkout, manual form entry, rather than fixing the underlying page performance. Shopify reports Shop Pay converting significantly higher than standard guest checkout, and PayPal's one-touch option shows a real lift among shoppers who already have an account. They are a strong complement to speed work, not a substitute for it.
Where should a small ecommerce or lead-gen site start with speed and conversion work?
Start on mobile, on the page where the transaction actually happens: cart, checkout, or the lead form, not the homepage hero. Fix INP on form interactions first, since that is the most commonly failed metric and the one directly tied to abandonment while someone is trying to pay or submit.
The takeaway
The Core Web Vitals story in 2026 isn't that the web got faster, it's that it stalled, and the pages where it matters most, checkout and lead forms, are the ones still failing INP at the highest rate. Every documented case study this year points the same direction: fix the interaction that happens right before someone hands over their money, and the revenue moves before the rankings do. If your last speed audit was a homepage screenshot and a PageSpeed score, you audited the wrong page. Go measure the cart, then fold the fix into your broader conversion optimization and performance marketing work, since a faster checkout makes every paid click you're already buying worth more.