Core Web Vitals and Conversion: What a Slow Site Actually Costs You

Speed is not a vanity metric. Here is how each second of load time maps to lost revenue, and the five fixes that recover most of it.

Performance work is easy to postpone because the cost is invisible. Nobody emails to say they abandoned your checkout because the page felt sluggish on a hotel wifi connection. They simply leave, and the analytics record a bounce indistinguishable from disinterest.

The relationship between load time and conversion is one of the most consistently replicated findings in ecommerce and lead generation. The exact figures vary by industry, but the direction never does: slower pages convert worse, and mobile amplifies the effect.

The three metrics that matter

Largest Contentful Paint (LCP)

How long until the main content appears. Target under 2.5 seconds on mobile. The usual culprits are an oversized hero image, a render-blocking font, and — on modern animated sites — content that fades in, which is measured as painting late even though the asset arrived on time.

Interaction to Next Paint (INP)

How quickly the page responds when someone taps or clicks. Target under 200 milliseconds. Heavy JavaScript running on every frame — smooth-scroll libraries, parallax, cursor effects — is the most common cause, and it never shows up on a developer's high-end laptop.

Cumulative Layout Shift (CLS)

How much the page jumps while loading. Target under 0.1. Images without reserved dimensions, late-injected banners and swapped fonts move content under the user's thumb — the single most infuriating failure mode on mobile, and the easiest to fix.

The five fixes with the highest return

1. Right-size and modernize images. Serve WebP or AVIF, generate responsive sizes, and lazy-load everything below the fold. This alone frequently halves page weight.
2. Preload the hero asset and the primary font, and never animate the LCP element's opacity from zero.
3. Reserve space for every image, embed and dynamic banner so nothing shifts once it arrives.
4. Defer non-essential third-party scripts. Chat widgets, heatmaps and tag managers are usually the largest single cost on a marketing site.
5. Budget your animation. Scroll libraries are wonderful until they run work on every frame; profile them and disable heavy effects on low-power devices.

• <2.5s LCP on mobile
• <200ms INP
• <0.1 CLS

Field data beats lab data

A synthetic audit score is a diagnostic tool, not a goal. What Google uses for ranking, and what actually reflects your customers, is field data from real visits. A site can score 98 in the lab and fail in the field because real users arrive on older phones, over slower networks, with an ad blocker and twelve tabs open.

Do not sacrifice the brand for the score

Stripping every animation to reach a perfect score is the wrong trade for a premium brand. The objective is a site that feels immediate and remains distinctive. In practice this means loading the fast, static shell first, and layering the expressive parts once the content is already usable — the difference between a site that performs and a site that merely benchmarks well.