Only 12.6% of the most visited sites running Qualified's chat and pipeline widget passed all three Core Web Vitals on mobile in September 2026, against 46.6% of all sites in Chrome's top million. That is the sharpest gap among the B2B tools we checked in the HTTP Archive's technology data, and it points at the real speed problem on B2B SaaS websites: not the platform the site is built on, but the sales and marketing scripts added after launch.
No public dataset labels a website as B2B SaaS, so this page uses the next best thing. It reads Chrome's real-user data, through the HTTP Archive technology report, for the platforms SaaS marketing sites are built on and the tools they typically carry, then adds Google's thresholds, the Web Almanac and the few studies that tie speed to leads.
The thresholds, and why a B2B SaaS site may have no score at all
Google judges a page on three field measures from real Chrome users: Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness and Cumulative Layout Shift (CLS) for stability. A page is good when 75% of visits reach an LCP of 2.5 seconds or less, an INP of 200 milliseconds or less and a CLS of 0.1 or less, judged separately on mobile and desktop (web.dev). Search Console reads the last 28 days and gives a group of similar URLs the status of its worst metric.
| Metric | Good | Needs improvement | Poor |
|---|---|---|---|
| Largest Contentful Paint | 2.5 s or less | 4 s or less | Over 4 s |
| Interaction to Next Paint | 200 ms or less | 500 ms or less | Over 500 ms |
| Cumulative Layout Shift | 0.1 or less | 0.25 or less | Over 0.25 |
Source: Google Search Console Help, Core Web Vitals report, read 2026-10-20
The B2B catch is traffic. The Chrome UX Report only covers pages and origins that are "sufficiently popular", and Google does not publish the minimum. Its data also leaves out Chrome on iOS and other Chromium browsers such as Edge, which matters when a share of buyers browse from a managed Windows laptop. When a group of URLs lacks data, Search Console falls back to the whole origin, and when the origin lacks data the site does not appear at all. A SaaS demo page with a few hundred visits a month will often have no page-level score, so the number a team watches is usually the origin average, which the blog can dominate.
Pass rates by the platforms SaaS marketing sites run on
The HTTP Archive detects the technology on each origin with Wappalyzer rules and joins it to Chrome's field data. The table shows September 2026, the latest month published on 2026-10-20.
| Platform | Mobile origins | All three, mobile | All three, desktop | Good LCP, mobile | Good INP, mobile | Good TTFB, mobile |
|---|---|---|---|---|---|---|
| All origins | 8,569,446 | 52.8% | 60.8% | 65.6% | 79.4% | 45.9% |
| Webflow | 59,483 | 71.3% | 82.6% | 78.6% | 88.0% | 73.1% |
| Astro | 47,277 | 71.1% | 84.6% | 81.1% | 85.0% | 63.1% |
| Framer | 11,319 | 70.8% | 87.9% | 83.4% | 80.3% | 69.3% |
| HubSpot CMS | 13,904 | 60.0% | 75.1% | 78.3% | 86.2% | 67.3% |
| Sanity | 10,586 | 52.4% | 68.6% | 66.8% | 75.8% | 57.8% |
| WordPress | 2,658,175 | 48.3% | 53.7% | 55.4% | 90.3% | 22.9% |
| Gatsby | 9,278 | 43.7% | 59.5% | 64.4% | 76.0% | 52.7% |
| Contentful | 9,212 | 38.1% | 55.5% | 60.9% | 70.3% | 44.6% |
| Next.js | 395,345 | 35.7% | 56.9% | 53.5% | 59.7% | 45.6% |
| Adobe Experience Manager | 19,174 | 33.9% | 44.3% | 52.3% | 66.8% | 44.1% |
Source: HTTP Archive Core Web Vitals technology report, all geographies, read through its public API on 2026-10-20
Hosted builders lead on server time. Webflow, Framer and HubSpot CMS serve pages from managed hosting with a CDN in front, and between 67% and 73% of their mobile origins have a good time to first byte. Only 22.9% of WordPress origins do, which is where most of its LCP deficit comes from.
Next.js is not one kind of site. Its 395,345 origins include web applications, logged-in dashboards and shops as well as marketing sites, and its weak responsiveness score, 59.7% good INP on mobile, reflects heavy client-side JavaScript across all of them. We build marketing sites in Webflow and applications in Next.js, so we are not neutral on either row. A Next.js marketing site that renders on the server and ships little script can pass; the average one does not.
Headless content platforms sit in the middle. Sanity and Contentful are content stores, so their rows describe whatever front end each team built. Contentful origins pass less often than Sanity ones, and both trail the hosted builders.
The platform advantage narrows on sites with real traffic
The all-origins figures are dominated by small sites with simple pages. Filtering to the top million origins by Chrome traffic, closer to the size of a funded SaaS company's site, changes the picture.
| Platform | Mobile origins | All three, mobile | All three, desktop | Good INP, mobile |
|---|---|---|---|---|
| All origins in the top 1 million | 747,521 | 46.6% | 61.9% | 65.7% |
| Webflow | 2,635 | 49.8% | 72.3% | 68.5% |
| Astro | 3,087 | 54.5% | 74.5% | 68.8% |
| HubSpot CMS | 704 | 45.2% | 64.8% | 65.2% |
| Framer | 359 | 42.3% | 74.5% | 54.2% |
| WordPress | 110,537 | 50.1% | 55.8% | 79.4% |
| Sanity | 1,043 | 35.3% | 60.4% | 54.6% |
| Contentful | 2,259 | 28.0% | 49.8% | 52.1% |
| Next.js | 39,764 | 25.5% | 49.3% | 44.7% |
Source: HTTP Archive Core Web Vitals technology report, rank filter Top 1M, read 2026-10-20
On mobile, Webflow's lead over the average falls from 18.5 points across all origins to 3.2 points in the top million, and HubSpot CMS drops slightly below it. Desktop holds up better for the hosted builders. The cause is visible in the INP column: busier sites carry more scripts, and responsiveness on a mid-range phone suffers whatever the platform. The Web Almanac's 2025 performance chapter found the same pattern by page type, with 45% of mobile home pages passing against 56% of secondary pages, because the home page carries the hero media and most of the tags.
The trend is still in the hosted builders' favour. Webflow's mobile pass rate rose from 56.3% in January 2024 to 71.3% in September 2026, Framer's from 54.0% to 70.8% and HubSpot CMS's from 48.1% to 60.0%. Next.js went from 33.0% to 27.4% in January 2025 and back to 35.7%, while all origins moved from 46.8% to 52.8% (HTTP Archive, same report).
The scripts B2B SaaS sites add cost more than the platform
Most B2B SaaS sites carry a tag manager, a marketing automation script, a consent banner, session recording, an intent or account identification tool, a chat or meeting router and an A/B testing script. The HTTP Archive tracks each of them.
| Tool | Mobile origins | All three, mobile | All three, mobile, top 1M | Good INP, mobile | Median mobile JS (KB) |
|---|---|---|---|---|---|
| All origins | 8,569,446 | 52.8% | 46.6% | 79.4% | 714 |
| Google Tag Manager | 2,547,920 | 46.5% | 44.2% | 77.7% | 1,173 |
| HubSpot (tracking and forms) | 228,372 | 43.9% | 36.7% | 71.9% | 1,415 |
| Salesforce Account Engagement (Pardot) | 53,778 | 48.2% | 41.1% | 79.5% | 1,174 |
| Marketo | 6,013 | 38.8% | 33.2% | 70.7% | 1,611 |
| OneTrust | 107,248 | 42.5% | 38.0% | 75.6% | 1,560 |
| Hotjar | 176,745 | 41.9% | 38.1% | 73.1% | 1,601 |
| VWO | 21,224 | 37.6% | 32.3% | 65.3% | 1,866 |
| Segment | 22,508 | 33.3% | 31.7% | 61.0% | 1,875 |
| Intercom | 21,758 | 37.8% | 28.1% | 60.2% | 1,803 |
| HubSpot Chat | 27,039 | 34.5% | 32.1% | 75.4% | 1,882 |
| 6sense | 2,578 | 29.9% | 21.8% | 61.2% | 1,882 |
| Qualified | 1,642 | 29.3% | 12.6% | 57.7% | 2,145 |
Source: HTTP Archive Core Web Vitals and page-weight technology reports, read 2026-10-20. KB is 1,024 bytes; JavaScript is the median per page across all origins with the tool.
These are correlations, not tests. A site that buys 6sense and Qualified is usually a larger company with a longer list of other scripts too, and the table cannot separate one tool's cost from the stack around it. Comparing within the top million removes some of the size effect, and every tool on the list still sits below the 46.6% average, most of them by 8 to 34 points.
The common thread is JavaScript. Pages carrying these tools hold 1.6 to 3 times the median page's script weight, from 1,173 KB with the tag manager to 2,145 KB with Qualified. Script that runs when the page loads competes with the visitor's first tap, which is why INP, not LCP, is the measure that falls furthest. The Web Almanac's 2025 performance chapter puts the median mobile page's total blocking time at 1,916 milliseconds in lab tests, and its third-parties chapter finds about 90% of pages load at least one third party, with a median of 79 third-party requests per mobile page.
Chat and booking widgets can wait. Google's Lighthouse guidance names Intercom and Drift among the embeds that suit a facade: a static button that looks like the widget and loads the real script only when a visitor clicks. Video players are the other common candidate, and the guide's example swaps a 540 KB YouTube player for a 3 KB placeholder.
What speed does to B2B leads
The evidence that ties speed to B2B pipeline is thin and old, and it is worth knowing exactly what each study measured. Our Website Speed Statistics 2026 post sets the controlled ecommerce tests from Vodafone and Rakuten 24 alongside these; here are the two that cover lead generation.
Deloitte and 55 for Google. The study followed 37 European and American brands and more than 30 million sessions, monitoring mobile load times for 30 days at the end of 2019. On lead-generation sites, a 0.1 second improvement across four speed measures came with a 21.6% rise in visitors moving from the first form step to the submission page, and a 5.5% rise from the home page to the first form step. It is correlational, and one of its four measures, First Meaningful Paint, is now deprecated.
Portent. Its 2022 analysis of about 100 million page views across 20 sites, 14 of them B2B lead-generation sites, found a page loading in one second converted three times as well as one loading in five seconds, and five times as well as one loading in ten. Portent does not name the load metric, and each site defines its own conversion.
Rankings. Google's documentation, updated in September 2026, says Core Web Vitals "are used by our ranking systems", that there is no single page experience signal, and that Search shows the most relevant content even when page experience is poor. For a SaaS site competing on comparison and category terms, speed breaks ties; it does not replace a better page.
The five-page speed test order for a B2B SaaS site
Field data tells you which pages fail; it cannot tell you why. We would test in this order, because it follows where buyers land and where a slow page costs a lead. It complements the conversion test order in our B2B SaaS Website Conversion Rate Statistics 2026 post.
| Order | Page | Why it goes here | What to check first |
|---|---|---|---|
| 1 | Demo or contact page | Every lead passes through it, and its form and scheduler scripts are the heaviest on the site | INP on mobile with the form, router and chat loaded |
| 2 | Pricing page | High-intent visitors compare here, often on a phone | LCP and layout shift from plan toggles and tables |
| 3 | Home page | The page most likely to fail, with hero video and the most tags | LCP image priority, hero video, tag count |
| 4 | Top paid landing page | Paid clicks cost money on every slow load | LCP on mobile, scripts loaded before consent |
| 5 | Top organic blog post | Usually the largest share of origin traffic, so it sets the origin score | Embeds, ads, consent banner shift |
Source: Octo Agency recommendation, based on the HTTP Archive and Web Almanac findings above
For each page, read the 28-day field data in Search Console or the Chrome UX Report first, then use a lab tool on a throttled mobile profile to find the cause. Run the lab test twice, once as a new visitor with the consent banner and once with every script accepted, because many B2B tools load only after consent.
What this means for a B2B SaaS marketing site
The platform choice buys a head start. Webflow, Framer and HubSpot CMS origins pass more often than the web as a whole, and the hosted builders answer the first request faster than most self-hosted stacks. That head start shrinks to a few points on sites with real traffic, and it disappears when a team adds an intent tool, a chat router and session recording without a budget.
So set the budget before the tools arrive. Agree an INP of 200 milliseconds and an LCP of 2.5 seconds on mobile for the five pages above, list every third-party script with an owner and a reason, load chat and booking widgets on interaction, and check field data monthly rather than once at launch. On our four-week marketing-site process we would write that script list into the brief, because it is easier to keep a tool off a page than to remove one a sales team relies on. Our marketing websites service page explains how we scope a build.
