Core Web Vitals for business websites, explained without jargon

Your developer said 'Core Web Vitals are fine' and you nodded politely. Here is what was actually being measured, in words a business owner can use — because you should be able to check this without trusting anyone's dashboard, including ours.

The three numbers Google measures

• LCP — how fast the main thing appears. The moment the visitor sees the meaningful content: your hero photo, your headline, the product image. Google's threshold: 2.5 seconds or faster, measured on real phones, not your office laptop.
• INP — how fast the page reacts when tapped. A menu opening, a filter applying, a button responding. Threshold: 200 milliseconds or less. INP replaced the older FID metric in March 2024 — if a report still says FID, it is out of date.
• CLS — how much the page jumps while loading. The 'I tapped the wrong thing because the page moved' score. Threshold: 0.1 or less — effectively, nothing should jump.

These thresholds are Google's own, published on web.dev, stable for years. Nothing here is a Codeloom opinion — it is the same yardstick Google uses.

Field data, not lab data — this trips everyone

The numbers that matter come from real Chrome visitors, collected at the 75th percentile across your actual traffic — Google calls this field data, and it shows in PageSpeed Insights and Search Console. A Lighthouse score, on the other hand, is a lab simulation on one test device with one connection. A site can score 98 in the lab and fail in the field, because your customers use slower phones and patchier networks than the test robot. When anyone quotes you a Lighthouse number as a promise, ask for the field data instead — that is the version your customers actually experience.

Why a business owner should care

Two reasons, in order of importance. First, customers: on a slow phone, every extra second before your products appear is people leaving — and they rarely come back. Second, search: page experience is one of many signals Google uses, and a fast site removes a handicap; it does not buy rankings by itself. Anyone promising rank #1 because of a speed score is selling something Google does not sell. We treat speed the way we treat spelling: expected, measured, never marketed as a miracle.

The three fixes that move the needle

• Right-size your images. On photo-heavy sites — furniture, interiors, food — oversized images are the usual cause of slow LCP. Served properly, a photo-heavy catalog still opens fast on phones; that is exactly what we did for CozyNest India, whose furniture catalog opens instantly and whose team told us it boosted local bookings.
• Show content before everything else loads. Your visitor should see the headline and the first products immediately, even while fonts, chat widgets and analytics are still arriving behind them.
• Reserve space for images and embeds. Fixed sizes and dimensions for media mean nothing jumps while loading — which is most of CLS right there.

How to check yours in five minutes

Open PageSpeed Insights, enter your URL, and scroll past the big score to the 'Diagnose performance issues' field section — that is the real-visitor data. Look at each Core Web Vital: green, needs improvement, or poor. Google Search Console also has a Core Web Vitals report showing this per page type for your whole site. Check on your phone too: open your site on the network you get at a railway station. That is your customer's network.

One honest limit: chasing a perfect 100 usually costs design quality for no customer benefit. If your field data is green and your customers are on fast-enough pages, stop optimizing — spend the budget on photos and content instead, which move conversions more than the last few milliseconds.

Send us your website URL on WhatsApp. We will read your field data — not the lab score — and tell you the one fix that matters, or tell you honestly that your site is fine and you should spend the money elsewhere.