Core Web Vitals for marketing sites: the smallest set of changes that matter

The three metrics, and the exact numbers

Core Web Vitals are three field metrics that describe how a page feels to a real person. Largest Contentful Paint (LCP) measures loading, and the good threshold is 2.5 seconds or less. Interaction to Next Paint (INP) measures responsiveness, and the good threshold is 200 milliseconds or less. Cumulative Layout Shift (CLS) measures visual stability, and the good threshold is 0.1 or less. Those are the numbers to keep in your head. Anything past them is yellow or red.

Why people still get this wrong

If you read older performance advice, you will see First Input Delay (FID) treated as the responsiveness metric. That changed. On March 12, 2024, INP became a stable Core Web Vital and replaced FID, and FID was removed from Google's tooling later that year, on September 9, 2024. The difference matters because FID only measured the delay before the browser started handling your first interaction, while INP measures the full latency of interactions across the whole visit. A page that looked fine on FID can fail INP, because INP catches the slow handlers and the janky re-renders that FID never saw. If your audit or your mental model still says FID, it is out of date.

How the score is actually judged

Core Web Vitals are assessed on real visits at the 75th percentile, segmented across mobile and desktop. That means a quarter of your visitors can have a worse experience than your reported number, so the metric is deliberately unforgiving of your slowest devices and connections. A page passes only when all three metrics hit good at that 75th percentile. One red metric fails the whole page. This is why a fast laptop on office wifi tells you almost nothing. The field data, not your local Lighthouse run, is what gets measured.

The short list that moves the needle

For a typical marketing site with modest JavaScript, a handful of changes take you from yellow or red into green, and they map cleanly onto the three metrics. For LCP and CLS, serve next-gen image formats with explicit width and height attributes so the hero image loads fast and nothing jumps as it arrives. For LCP and render-blocking, handle web fonts deliberately with preloading and a sensible font-display so text is not held hostage by a font file. Ship your CSS, Bootstrap or otherwise, as a single minified file with the critical above-the-fold CSS inlined, so the page can paint without a round trip. And pull render-blocking third-party scripts out of the top of the page, because tag managers, chat widgets, and analytics loaders are the usual reason a simple page is slow.

The lever most performance checklists skip

Older versions of this advice stopped there, and that was fine when responsiveness barely counted. With INP in the picture, you need at least one lever aimed at it. The main cause of bad INP on a marketing site is too much JavaScript running at the wrong time, especially above the fold. So defer or trim the scripts that do not need to run during the initial load, and break up long tasks that block the main thread, so the browser can respond to a tap or click instead of freezing for a beat. You rarely need a heavy framework on a brochure page at all. Less JavaScript is the cheapest INP fix there is.

What rarely matters

On marketing sites with modest JavaScript, the fashionable optimizations are usually noise next to the list above. Code-splitting, hydration tuning, edge rendering experiments, or swapping the framework wholesale tend to cost a lot of engineering time and move the field numbers very little. Profile first with real field data, find the one or two metrics that are actually red, and fix those. Do not refactor before you measure, and do not chase a perfect Lighthouse score on a synthetic run that no real visitor will ever experience.

A reality check

Performance is a floor, not a strategy. Google has been explicit that there is no single page-experience ranking signal, and that Search always seeks to show the most relevant content even when the page experience is sub-par. Where a lot of genuinely helpful content exists, a good page experience can help, but it acts as a tiebreaker among comparable pages, not as a substitute for being the better answer. So get into the green and stop there. A 95 score on a thin page does not beat a 75 on a deep, useful one. Fix the floor, then go spend your time on the content, because that is what the ranking is actually about.

Sources

External sources are provided for verification. NavoTech is not affiliated with and does not endorse the organizations cited.

Written by the team at NavoTech Digital Solutions. Have a project or counter-example? Get in touch.