Core Web Vitals are three measurements with three thresholds. LCP, how long before the largest visible element appears: under 2.5 seconds. INP, how long the site takes to react when you tap: under 200 milliseconds. CLS, visual stability while the page loads: under 0.1. Google measures all three on your real visitors, on mobile, over a rolling 28 days — not on the test you just ran from your desk.
For a small business, the point of these numbers is not ranking first of all: it is the money lost during loading. A site taking six seconds on mobile loses a meaningful share of its visitors before the offer has even appeared. Here is what matters, in what order to fix it, and what not to expect from it.
The three measurements, in plain words
LCP — time until the page looks loaded. The moment the biggest element on screen appears: usually the hero image or the main heading. It is the easiest one to improve and the one that weighs most on how the visitor experiences the site.
INP — reaction time. You tap a button; how long before something happens. This measurement replaced the old FID in 2024, and it is stricter: it watches every interaction in the visit, not just the first. On WordPress it nearly always degrades because of script overload.
CLS — layout shift. You go to tap, an image finishes loading, the button moves, you tap the wrong thing. It is the least known of the three and often the simplest to fix: declare image dimensions and reserve space for banners.
The score that counts is your real mobile visitors, not your desktop test.
The five causes of slowness on WordPress
1. Images. The number one cause, by a wide margin. A 4 MB photo uploaded straight off a camera can take a page from two seconds to eight on its own. Resizing, compressing and serving modern formats solves half of all LCP problems by itself.
2. The number of plugins. Each plugin adds scripts and stylesheets, loaded on every page whether they are needed there or not. A small-business site rarely needs more than a dozen active plugins; we routinely find installations running thirty-five, half of them overlapping.
3. Cheap shared hosting. Hosting at $4 a month shares a server with hundreds of sites. Server response time — the very first brick of LCP — becomes unpredictable at peak hours. Moving to decent hosting often beats three days of optimisation work.
4. Web fonts. Three families with five weights each, loaded from an external service, cost several hundred milliseconds and cause text to jump. Two families, two weights, hosted locally: the gain is immediate and invisible to the visitor except for the better.
5. Carousels and background video. A five-image carousel loads five images to show one. A hero background video loads several megabytes before the first readable word. Both are almost always aesthetic decisions with no measurable effect on conversion, and they are expensive.
The order to fix things in
Biggest gain to smallest, no exceptions: images first, hosting second, then a clear-out of plugins, then fonts, then caching. Caching comes last on purpose: it is the most common reflex, and it is a dressing over a wound when the four causes above are untreated. A slow site with caching is still slow for the first visitor to every page.
The orders of magnitude we see on a small-business WordPress site: image compression recovers 30 to 50% of load time, decent hosting 20 to 30%, plugin clear-out 10 to 20%. One well-aimed day of work moves most sites out of the red and into the green. What remains after that is development work, not maintenance.
What not to expect from it
Going from 60 to 95 on a performance test will not move you from position twenty to position three. Speed is a tiebreaker between pages of comparable quality, not a ranking lever in its own right. A slow page that answers the question exactly will always beat a fast, empty one.
Where speed genuinely pays is conversion, and the effect there is far more direct than in search. Every second saved on mobile recovers visitors who would have left before seeing your offer — that is stage 1 to stage 2 of the sales funnel, and a leak that leaves no trace in your form data, since those people never reached the form. Spotting that kind of loss is covered in conversion funnel leaks.
Two moments justify a real speed push: right after a rebuild, alongside the checks described in site migration without losing your SEO, and as part of a broader assessment, as described in what a free SEO audit should contain.
The rest is upkeep: images get heavier, plugins accumulate, and a site optimised once is slow again within eighteen months. That is exactly what our maintenance plan, from $120 a month, covers — Canadian hosting and monitoring included. To find out where your site stands today, book the call: we open your Core Web Vitals report together and rank the fixes by expected gain.