The honest answer about CWV in 2026

If you run a B2B SaaS marketing site, your weak Core Web Vital is almost certainly INP, not LCP. Industry benchmarks put SaaS at 345ms mobile INP, worse than e-commerce, news, and finance (Linkgraph, 2026). And the threshold for "good" sits at 200ms. So most SaaS sites are roughly 1.7x over the line on the metric that matters most.

The other piece of honest news: Google has not added a new metric, has not tightened any threshold, and has confirmed LCP, INP, and CLS remain unchanged going into 2026 (Whitehat SEO). Anyone selling you "Core Web Vitals 2.0 readiness" is selling air. The problem is not that the rules changed. The problem is that B2B SaaS sites carry more JavaScript than almost any other category, and INP punishes JavaScript.

Core web vitals B2B SaaS benchmarks for 2026

Here is what the 2026 reality looks like across industries, measured at the 75th percentile of real-user mobile sessions. SaaS sits at the bottom.

Mobile INP by industry (75th percentile, milliseconds)

Source: Linkgraph INP benchmarks, 2026. Anything above 200ms is "needs improvement". Anything above 500ms is "poor". SaaS as a category lives in the yellow zone.

Why is SaaS the worst? Three structural reasons we see on every audit:

• Marketing sites built on the same component library as the product app, shipping app-grade JavaScript to a brochure page

• Heavy third-party scripts: HubSpot, Marketo, Drift, Clearbit, GA4, Hotjar, Segment, plus a CMP banner that loads on every interaction

• Animation-heavy hero sections (Framer, Webflow Lottie embeds) that delay the main thread on the first user click

None of these show up in PageSpeed Insights as a single villain. They show up as a death-by-a-thousand-cuts INP score.

Why INP hurts B2B more than B2C

The Core Web Vitals threshold for passing is the 75th percentile of real-user data, meaning a page passes only if at least 75% of visits achieve "good" scores (Parachute Design). For consumer sites, mobile traffic dominates and the 75th-percentile sessions are usually mid-tier Android phones on flaky 4G. That sounds bad for B2C, and it is.

But B2B has its own quirk. Decision-makers research on mobile, then convert on desktop. Parachute Design's data shows 72% of product research starts on mobile (source). If your mobile INP is 400ms, the buyer forms a "this site feels slow" impression during research, then never returns to the desktop demo form. The conversion lost on Tuesday's desktop session was actually killed on Sunday's mobile session.

This is also why "but our desktop INP is fine" is the most common bad excuse we hear from SaaS founders. Google grades you on the worst form factor your real users hit. Mobile-first indexing means mobile-first INP scoring.

For more on how mobile shapes ranking signals beyond just performance, see our breakdown of mobile-first indexing.

What we actually find in Gravidy audits

We have run technical audits on B2B SaaS sites in the 50-300 employee range, mostly DACH and Nordics. The pattern repeats. Below is what we typically find driving INP into the yellow or red zone:

1. Cookie consent banner blocking the main thread. Cookiebot, OneTrust, Usercentrics. They all add 200-400ms of script execution before the first user click registers. The fix is not "remove consent". The fix is to defer the analytics scripts behind the banner, not the banner itself.

2. HubSpot embed forms that ship 180KB of JavaScript to render a five-field demo form. We replace these with native HTML forms posted to the HubSpot API. INP usually drops 80-120ms on the contact page alone.

3. Framer or Webflow sites with auto-playing video heroes. The video itself is fine. The lazy-loaded JS player that hydrates on scroll is what spikes INP when the user clicks the nav menu mid-scroll.

4. Tag managers loading 15+ marketing pixels. LinkedIn Insight Tag, Meta, G2, Capterra, Bing UET, plus three retargeting pixels nobody remembers approving. Each adds 20-60ms of long tasks.

5. React-based marketing sites doing client-side routing for static content. The whole point of a marketing page is server-rendered HTML. Hydrating a 400KB bundle to render text and an image is malpractice.

We are not above making mistakes here. On one audit we recommended removing a chat widget that turned out to drive 18% of qualified pipeline. Speed matters, but speed without conversion data is just vanity. Always measure the dollar impact, not only the millisecond impact.

The thresholds that matter (and the ones that don't)

The "good" thresholds for 2026 are unchanged: LCP under 2.5s, INP under 200ms, CLS under 0.1. Google's own pass-rate data shows 55.7% of origins now hit good across all three (Whitehat SEO, December 2025). That is a strong improvement over the 2022 baseline. It also means the bar to outrank competitors on technical performance is rising, not falling.

What does NOT matter, despite what some SEO tools imply:

• Lighthouse scores. They are lab data. CrUX (Chrome User Experience) is what Google actually uses for ranking.

• TTFB targets below 200ms. Helpful, not required. Most B2B sites win or lose on INP, not server response.

• Perfect 100/100 scores. Passing the threshold matters. Beating it by 50ms does not.

The actual ROI math from real case studies is unambiguous. Vodafone improved LCP by 31% and saw an 8% increase in sales plus a 15% rise in lead-to-visit rate (Growmatix). Rakuten 24 improved CWV and recorded a 53.37% increase in revenue per visitor (same source). A one-second delay can reduce conversions by up to 7% (Parachute Design).

For B2B SaaS specifically, sites passing all three Core Web Vitals are 24% less likely to experience user abandonment (Parachute Design). On a $50K ACV product, that is not a rounding error.

How to fix INP on a SaaS marketing site

The fix order we recommend, in priority sequence:

1. Audit your tag manager. Remove every pixel without a named owner and a documented use. Expect to cut 30-50% of tags on the first pass.

2. Replace embed forms with native HTML posting to vendor APIs. HubSpot, Marketo, Pardot all support this. Your dev team will fight you on it. Win the fight.

3. Move all non-essential third-party scripts behind interaction. Chat widgets load when the user clicks the chat button, not on page load.

4. Audit your CMS theme. If you are on Webflow or Framer with a stock template, you are shipping CSS and JS for components you never use. Strip them.

5. Use content-visibility: auto on below-the-fold sections. It is a one-line CSS change that often shaves 200-400ms off LCP.

This is the kind of work that compounds with everything else in technical SEO. You cannot earn rankings with content alone if the page is slow enough to hurt the user signal.

Frequently Asked Questions

How important is INP for B2B SaaS specifically?

Critical. SaaS averages 345ms mobile INP, the worst across industries tracked, versus a 200ms "good" threshold (Linkgraph). If you fix only one Core Web Vital, fix INP. The JavaScript-heavy nature of SaaS marketing stacks (forms, chat, analytics, animations) makes INP the bottleneck on almost every audit we run.

Do Core Web Vitals affect B2B differently than B2C?

Yes, in two ways. First, B2B has higher absolute conversion value, so each abandoned session costs more. Second, B2B buyers research on mobile and convert on desktop, so a poor mobile INP score can kill a desktop conversion that never happened. Sites passing all three CWV see 24% less abandonment (Parachute Design).

Are the CWV thresholds changing in 2026?

No. LCP under 2.5s, INP under 200ms, and CLS under 0.1 remain unchanged. No new metrics have been added since INP replaced FID in March 2024 (Whitehat SEO). Anyone marketing "CWV 2.0 services" is selling repackaged consultancy, not new requirements.

Is Lighthouse score the same as Core Web Vitals?

No, and confusing them is a common mistake. Lighthouse runs in a controlled lab environment and produces a synthetic score. Google ranks pages using CrUX field data from real Chrome users at the 75th percentile. A site can hit Lighthouse 95 and still fail CWV in the field. Always check Search Console's Core Web Vitals report, not just Lighthouse.

What this means for your site

Most B2B SaaS marketing sites we audit have 3-5 INP-killing scripts running for no business reason, plus a HubSpot or Marketo embed that single-handedly fails the contact page. None of it requires a redesign. It requires an honest audit and someone willing to remove things.

If you want to know which scripts are draining your INP and which fixes will move the ranking needle, book a free SEO audit call. Thirty minutes, specific findings on your actual CrUX data, no slide decks.