What a tenth of a second is worth
A 0.1 second improvement in mobile load time raised lead-form completions by 21.6% across 37 brands and 30 million sessions. Not seconds: a tenth of one.
Published August 20, 2026
Everyone accepts that a slow website is bad. What almost nobody has a feel for is the scale: whether we are talking about a rounding error or a real number, and how much speed you need to gain before anything changes.
There is a good answer to that, and it is more dramatic than most people expect.
The study worth knowing about
Google commissioned research carried out by 55 and Deloitte, published as Milliseconds Make Millions. They instrumented 37 leading European and American brand sites, audited each site's analytics setup, and collected data on more than 30 million user sessions, monitoring mobile load times hour by hour for 30 days.
Crucially, no site went through a UX redesign during the study. That matters enormously: it means the measured differences are attributable to speed rather than to someone also improving the page.
They then measured what a 0.1 second improvement in mobile speed did to behaviour. Not one second. One tenth of one.
- Retail: conversion rate up 8.4%, and consumers spent 9.2% more.
- Travel: conversion rate up 10.1%.
- Luxury: 40.1% more progression from product detail to add-to-basket.
- Lead generation: 21.6% more users reached the form submission page.
Why the lead generation number is the one to look at
If you run a service business (a contractor, a firm, a clinic, an agency), you are the lead generation column, and it carries the largest effect in the study outside luxury retail.
That fits how a form actually feels to fill in. Buying a specific item you already decided on is desire; you will tolerate some friction to get it. Filling in an enquiry form is hesitation. The visitor is weighing whether this is worth the effort and whether they want to be contacted at all, and at that moment every extra beat of waiting is an invitation to reconsider.
What makes a small business site slow
The causes are boringly consistent, and in most cases none of them require a rebuild.
Images, almost always
The single most common cause is a photo uploaded at the size it came off the camera, served at full resolution to a phone that will display it 400 pixels wide. A 4MB hero image is not unusual on a site nobody has audited. Correctly sized, compressed, and served in a modern format, the same image is often under 100KB and looks identical.
Third-party scripts you forgot you added
Every chat widget, heatmap tool, pixel, review embed and font service is code from someone else's server that has to be fetched and run before your page settles. Each one seemed reasonable on its own. Together they are frequently the majority of a page's load, and a good number of them are for tools nobody at the business has opened in a year.
Hosting that is cheap for a reason
Shared hosting at a few dollars a month puts your site on a machine with hundreds of others. Server response time is the floor everything else is built on: if the first byte takes 800 milliseconds to arrive, no amount of front-end work gets you back under budget.
Layout that moves while it loads
Not strictly speed, but it registers as slowness to the person using it. If content jumps as images and fonts arrive, the page feels unfinished and unstable, and taps land on the wrong thing. This is what Cumulative Layout Shift measures: see Core Web Vitals in plain English for how it is scored.
How to find out where you stand
- 01Run your site through PageSpeed Insights on mobile, and read the field data section rather than the lab score: that is real visits from real people, which is what Google uses.
- 02Sort your page's resources by size. If the top of that list is images, you have found your afternoon's work and most of your available gain.
- 03List every third-party script and ask who at the business uses the tool it belongs to. Remove the ones with no answer.
- 04Check server response time separately from everything else. If it is slow, fix hosting before touching anything on the page.
One caveat worth stating plainly: the study measured large brands with high traffic, where a percentage point is a lot of money. The percentages should transfer to a smaller site, but the absolute numbers obviously will not, and anyone quoting you a revenue figure from it for your specific business is guessing.
What does transfer is the direction and the scale of effort. This is one of the few improvements available to a small business where the work is measured in hours, the cost is near zero, and the effect is measurable in enquiries.
Every site we build is tuned against these metrics before launch and monitored after, because a fast site is the cheapest conversion work available. If you want to know where yours currently sits, a free audit will tell you.
Frequently asked questions
In the largest controlled study available, a 0.1 second improvement in mobile load time raised retail conversions 8.4%, travel conversions 10.1%, and lead-generation form completions 21.6%. Google commissioned the research; 55 and Deloitte carried it out across 37 brand sites and more than 30 million sessions.
Yes, and that is the surprising part. The study deliberately measured a 0.1 second delta rather than whole seconds, and found measurable movement across the entire purchase funnel at that scale. No redesigns took place during the study, so the change is attributable to speed rather than to layout.
The biggest measured effect (21.6%) was on users reaching the form submission page. A form is a moment of hesitation rather than desire: the visitor is deciding whether the effort is worth it, and friction weighs heavier at that moment than it does for someone already committed to buying a specific item.
Images almost always. Oversized, uncompressed images in modern formats are the single most common cause of a slow small-business site, and fixing them requires no rebuild. After that: how many third-party scripts you load, and whether your hosting is genuinely fast.