Performance

How Website Speed Quietly Kills Your Conversion Rate

Speed is the tax every visitor pays before they see your offer. Most teams never connect the soft quarter to the four-second load.

Most teams find out their site is slow the same way. A soft quarter, a marketing spend that is not paying back, and nobody can point to why. The homepage looks fine. The copy is fine. The product is fine. So the speed never gets blamed, because speed is the one thing you do not notice when it is missing. You just leave.

We design and build these sites for a living, and website speed and conversion rate are joined at the hip more tightly than almost anything else we touch. Not because speed is glamorous, but because it is the first impression a visitor forms, made before a single line of your copy is read.

The number that should bother you

Deloitte’s “Milliseconds Make Millions” study put a figure on it that is hard to unsee. A 0.1 second improvement in mobile load time lifted retail conversions by around 8 percent. A tenth of a second. Google’s own data has shown bounce probability climbing fast as load time stretches from one second to three, then falling off a cliff past five.

Here is the part people miss. You do not lose the slow-page visitors at checkout, where you would see them in the funnel. You lose them at the first paint, before they enter any funnel you are measuring. They never become a data point. So your analytics look healthy while a third of your traffic bails in the gap between “clicked the ad” and “saw anything”. That invisibility is exactly why slow sites stay slow.

What “slow” actually means now

Fast stopped being a vibe and became a measurement. Google’s Core Web Vitals are the three you are graded on, and they map to how a real person experiences the page:

• LCP (Largest Contentful Paint): how long until the main thing, usually your hero image or headline, shows up. Aim for under 2.5 seconds. This is the big one.
• INP (Interaction to Next Paint): when someone taps a button, how long before something happens. Under 200 milliseconds. Sluggish menus and laggy add-to-cart buttons live here.
• CLS (Cumulative Layout Shift): how much the page jumps around while loading. You have felt this. You go to tap a link, an ad shoves it down, and you tap the wrong thing.

These are not just experience niceties. They are a confirmed Google ranking input, so a slow page loses twice: fewer people arrive from search, and the ones who do convert worse.

Where the time actually goes

After enough audits you start seeing the same culprits. In rough order of how often they are the real problem:

Images. Almost always images. A 2 MB hero photo that could have been 180 KB. PNGs where WebP or AVIF would have cut the weight by 70 percent with no visible difference. Images with no width or height set, so the layout reflows as each one loads. Fixing images alone often takes a site from embarrassing to fine.

Third-party scripts. The chat widget, the heatmap tool, the three analytics tags, the A/B testing snippet, the cookie banner that loads a second cookie banner. Each one feels free. Together they are often half your load time, and they run on every page whether or not anyone uses them. We have cut a site’s load time nearly in half just by auditing the tag manager and deleting what no one had looked at in a year.

The framework tax. A marketing site built as a heavy single-page app, shipping a megabyte of JavaScript to render what could have been mostly plain HTML. Sometimes that complexity is earned. Often it is a default nobody questioned.

No caching or CDN. Serving every visitor from one server in one location, with no edge cache, so someone three time zones away waits for a round trip on every asset.

The fixes, in the order we would do them

You do not need a rebuild. You need to attack the list in the order that pays back fastest:

1. Compress and convert every image to WebP or AVIF, size them to what is actually displayed, and set explicit dimensions. Lazy-load anything below the fold.
2. Audit third-party scripts and remove anything you cannot name a current reason for. Defer the rest so they do not block the page from rendering.
3. Put a CDN in front of everything and cache aggressively. For most marketing sites this is a config change, not a project.
4. Measure on a real mid-range phone on a normal connection, not your laptop on office wifi. A developer’s MacBook is the most misleading test environment in existence.

A redesign is the expensive, slow way to get faster. Most of the time you can get 80 percent of the win from this list in a couple of weeks, and then decide whether the structure itself needs rethinking.

The honest caveat

Speed is necessary, not sufficient. A blistering-fast page selling the wrong thing to the wrong person still will not convert. We have seen perfect Core Web Vitals scores on sites that sold nothing, because the offer was the problem. Speed removes a tax. It does not write your value proposition. But of all the levers that affect conversion, it is the one that is most measurable, most often broken, and cheapest to fix relative to what it returns.

If your numbers feel soft and you cannot explain it, measure your LCP on a phone before you touch the copy. The answer is there more often than people expect. It is the kind of thing we look at first in the work we do, and the same logic holds for a store: fix the load before you reach for a redesign budget. Either way, tell us what you are seeing.

Published 29 June 2026 · Taranity