People leave
As a page goes from one second to several, the share of mobile visitors who give up climbs steeply. On a phone, on a weak signal, a heavy page is a lost enquiry — and you never find out it happened.
Speed
Every web company says their sites are fast. Here are real numbers for real sites, with the date and the method, so you can run the same test yourself.
What a visitor waits for
Measured from an ordinary home connection in Los Angeles on 29 October 2026, best of three requests per page. “First byte” is how long before the server starts answering; “document” is the whole HTML file delivered.
| Site | First byte | Document | HTML size | |
|---|---|---|---|---|
| Best City Nails & Spa bestcitynail.com |
151ms | 171ms | 21 KB | |
| Our pricing page imdgfx.com/pricing |
154ms | 244ms | 136 KB | |
| Castaic Nails & Spa castaicnailsandspa.com |
159ms | 160ms | 11 KB | |
| Our portfolio imdgfx.com/work |
165ms | 226ms | 68 KB | |
| Magic Beauty Supply magicbeautysupply.com |
165ms | 190ms | 26 KB | |
| Roy’s Plumbing roysplumbing.net |
170ms | 236ms | 48 KB | |
| Our own homepage imdgfx.com |
194ms | 303ms | 97 KB | |
| Baddy Print baddyprint.com |
309ms | 431ms | 247 KB |
One caveat worth stating: these are shared servers, and any single request can be slower than usual. Re-running this test gave first-byte figures between about 140ms and 310ms for most pages, with one outlier near 880ms. We publish the typical figure, not the luckiest one.
Why it matters
As a page goes from one second to several, the share of mobile visitors who give up climbs steeply. On a phone, on a weak signal, a heavy page is a lost enquiry — and you never find out it happened.
Google publishes its own thresholds for how quickly a page responds and settles. They are not the only ranking factor, and anyone telling you speed alone will put you first is overselling, but a slow site is a handicap you do not need.
Most slow small business sites are not slow because of the server. They are slow because of enormous uncompressed photos, a dozen tracking scripts and a page builder that ships code for features the page never uses.
What we are not claiming. These figures are time-to-first-byte and the HTML document transfer. They are not Largest Contentful Paint, which also depends on images, fonts and the visitor’s own device and connection.
We are also not publishing measurements of other companies’ websites. Picking a handful of slow competitors and putting them in a table would prove nothing — you can run our free audit on any site, including ours, and see for yourself.
The other half
Measured on the server itself, so there is no network in the way — this is pure build time, the part we control. Last run 29 October 2026. The rest of what a visitor waits for is the distance between them and the server, which no one can remove.
| Site | Server build | HTML size |
|---|---|---|
| Castaic Nails & Spa | 32ms | 11 KB |
| Best City Nails & Spa | 31ms | 21 KB |
| Magic Beauty Supply | 33ms | 26 KB |
| Roy’s Plumbing | 37ms | 48 KB |
| Our portfolio | 35ms | 68 KB |
| Our own homepage | 31ms | 97 KB |
| Our pricing page | 34ms | 136 KB |
| Baddy Print | 221ms | 247 KB |
Questions
Let’s build something great
Custom digital solutions. Real growth. Lasting results.