Landing Page Load Speed: Impact on Conversions
What each second costs in measured conversion, and the four fixes that account for most of the available gain.
On this page
The short version
- The relationship between speed and conversion is real and it is not linear — the damage between two and four seconds is far larger than between six and eight.
- Four fixes account for most of the available gain, and three of them are about images and third-party scripts rather than anything structural.
- Measure on a mid-range phone on a mobile connection. A desktop test on office broadband measures a condition almost none of your visitors are in.
Speed advice tends to arrive as a threshold — under three seconds, under two — which suggests a cliff that does not exist. What the evidence supports is a curve, steepest at the start: the difference between one second and three costs far more than the difference between five and seven.
That matters for prioritisation. A page at 2.5 seconds has more to gain from a second of improvement than a page at seven does, which is the opposite of where the effort usually goes.
Relative conversion by load time
Indexed to a one-second load. The shape is what matters — the loss is front-loaded, so the first second saved is worth more than the third.
Indicative of the shape rather than measured figures for your page. The point is the curvature: measure your own and prioritise the steep part.
The four fixes
In order of how much they typically return.
What to do, in order
Remove third-party scripts you are not using
Landing pages accumulate tags — an old analytics tool, a heatmap trial, a chat widget nobody answers, two tag managers. Each one is a request to somebody else's server, and a slow third party blocks your page.
This is usually the single largest win and it costs nothing but the audit.
Size and format the images properly
A hero image exported at 3000px and displayed at 600 is the most common single-file problem. Serve at twice the display width, in a modern format with a fallback, and set width and height so the layout does not shift as it loads.
Stop the fonts blocking the text
Web fonts that block rendering leave the page blank while they download. Setting the font display behaviour so text renders immediately in a fallback is one line, and it converts a blank screen into a readable one.
Defer everything that is not the first screen
Images below the fold, scripts for later sections, anything the visitor cannot yet see. The first screen is what the measurement is about and what the visitor is waiting for.
Layout shift costs conversions too
A page that loads quickly and then jumps as images and fonts arrive is experienced as broken, and the specific failure is a visitor tapping a button that moves out from under them.
The fixes are mechanical: width and height attributes on every image, reserved space for anything injected later, and no content inserted above existing content once the page has rendered. Cookie banners and promotional bars are the usual culprits.
It is worth treating as part of speed rather than as a separate concern, because from the visitor's side both are the same complaint: the page was not ready when it looked ready.
Speed check
- Third-party scripts audited, unused ones removed
- Hero image sized for its display width
- Fonts do not block text rendering
- Below-the-fold images and scripts deferred
- Every image has width and height attributes
- Nothing is inserted above content after render
- Measured on a throttled mid-range phone
- Re-measured after each redesign, not once
Frequently asked questions
How fast is fast enough?
Faster is better with diminishing returns, so the useful target is relative: get the first screen usable in about two seconds on a mid-range phone. Chasing a specific score past that point returns less than the effort costs.
Does a page builder make this impossible?
Harder, not impossible. Most of the weight from builders is unused scripts and oversized images, both of which can be addressed without leaving the tool.
Is a separate lightweight page worth building?
Occasionally, for high-volume paid traffic, where a small conversion gain pays for the duplication. For most pages the effort is better spent making the existing one lighter.