What Are Core Web Vitals?
Core Web Vitals are three measurements Google uses to describe how a page actually feels to use: how fast the main content appears, how quickly the page responds to input, and how much the layout jumps around while it loads. They are not the only ranking factor, but for B2B sites that compete on long-tail service terms, they often decide who wins the click.
Why They Matter for B2B Buying Decisions
A slow page does not only annoy the visitor; it changes behaviour. Studies and real-world audits consistently show that pages slower than about three seconds lose a meaningful share of visitors before they read anything. For a B2B company bidding on high-value service terms, a fraction of a second on mobile can be the difference between a qualified inquiry and a bounce that feeds a competitor.
LCP: Loading
Largest Contentful Paint measures how long it takes for the biggest visible element - usually a hero image or headline - to finish rendering. Under 2.5 seconds is the target. On B2B sites the usual culprits are an unoptimized hero photograph, a render-blocking font, and a third-party chat widget that loads before the page content does.
What Counts as Good, Needs Improvement, or Poor
Google groups each metric into three bands. For LCP, under 2.5 seconds is good, 2.5 to 4 seconds needs improvement, and over 4 seconds is poor. Because the score is based on real field data from your actual visitors rather than a lab test, a page can feel fine on your fast office connection and still score poorly for a prospect on a slow mobile network.
Common B2B Causes of a Slow LCP
The most frequent causes we see are large uncompressed hero images exported straight from a designer's tool, fonts loaded from an external provider that block rendering, and multiple third-party scripts - chat, analytics, tag managers, marketing pixels - stacked in the page head before the content. Each one adds delay, and together they easily push LCP past four seconds.
INP: Interactivity
Interaction to Next Paint replaced First Input Delay as the responsiveness metric. It measures the worst delay between a user action - tapping a menu, pressing a button, submitting a form - and the next frame the browser can paint. Heavy analytics bundles, chat scripts, and large JavaScript frameworks all push this number up, and the fix usually starts with removing code rather than adding it.
What Is a Good INP Score
An INP under 200 milliseconds is good, 200 to 500 milliseconds needs improvement, and over 500 milliseconds is poor. For a B2B site this matters because the actions a buyer takes - opening navigation, filtering a product list, starting a quote request - are exactly the moments where a laggy response convinces them to try someone else.
What Typically Slows INP Down
The usual suspects are oversized JavaScript bundles, a tag manager firing too many scripts at once, chatbots that claim the main thread on page load, and poorly written third-party embeds such as maps or calculators. The first diagnostic step is rarely to buy faster hosting; it is to audit what code actually runs on each page and delete what no one asked for.
CLS: Visual Stability
Cumulative Layout Shift scores how much the page visibly moves while loading. A cookie banner that pushes the whole layout down, images without width and height attributes, and late-injected fonts are the classic causes. It is the easiest of the three to fix and the most painful for a visitor to experience.
The Usual Culprits Behind Layout Shifts
The most common shifts come from inserting a banner or interstitial after the page has rendered, images and videos without reserved dimensions, web fonts that swap from a fallback face once they load, and dynamically injected ads or lead widgets. Fixing them is mostly a matter of reserving space in the markup, so the browser knows the size of every element before it paints.
How to Measure and Fix Them
Start with real-user data, not a synthetic lab test. Look at the field report in Google Search Console to find the actual URLs where your real visitors struggle, then reproduce the worst cases on a throttled mobile connection. Fix the one page that sends the most traffic and has the worst score, then move to the next.
Where to Measure From
Field tools such as the Chrome User Experience Report and Search Console show what real visitors experience; lab tools such as Lighthouse show what you can improve. Use the field data to decide what to fix and the lab tools to confirm the fix worked. Never optimize solely against a lab score, because lab conditions rarely match a factory floor on cellular.
Frequently Asked Questions
The questions B2B marketing and IT managers ask most often before starting performance work.
Do Core Web Vitals directly decide Google rankings?
They are one ranking signal among many, not a rule that alone moves positions. Their real effect is indirect but real: pages that load faster and respond better tend to keep visitors longer, engage them more, and convert better, and those behavioural signals are what Google rewards. Fixing the vitals is rarely a magic jump, but ignoring them on competitive terms is a real handicap.
Which of the three should we fix first?
Start with whichever one your field data shows is worst on your highest-traffic pages. In practice, CLS is usually the fastest win because it is mostly markup fixes, LCP follows with image and font work, and INP takes longer because it means removing or deferring JavaScript. Sequencing by effort and impact beats tackling all three at once.
Are Core Web Vitals measured on desktop or mobile?
Google reports them separately for mobile and desktop, and the mobile scores usually matter more for ranking. Since the majority of B2B research now happens on phones, treat the mobile numbers as the ones that count and test your fixes on a throttled 3G connection rather than on a fast office Wi-Fi.
Conclusion
Treat Core Web Vitals as a maintenance routine rather than a project. Measure on real devices, fix the worst page first, and re-check after every redesign or third-party integration. If you would rather hand the routine to someone else, that is exactly what our performance work covers.