Slow 4G is around 150 ms. Fibre is under 30.
Your TTFB. Check it in PageSpeed Insights.
CSS, fonts, the hero image, blocking scripts.
Parsing, layout and paint once everything has arrived.
A planning model, not a measurement. Real loading involves caching, compression, connection reuse and priority hints. Use this to set a target, then measure the actual result with PageSpeed Insights.
Calculated entirely in your browser. Nothing you enter is sent anywhere.
How to use it
Pick a target and a connection
2.5 seconds on slow 4G is the honest starting point for most sites.
Enter your real server timings
TTFB and request count come straight from PageSpeed Insights.
Split the remaining budget
The suggested split gives roughly 45 percent to the hero image.
How the calculation works
The model works backwards. Your target is a time budget. Subtract the parts you cannot compress away — connection handshake, server response, request round trips and browser render — and what is left is transfer time.
Transfer time multiplied by connection speed gives you bytes. That is your budget, and it is usually smaller than people expect.
The request overhead matters as much as the bytes. Browsers open a limited number of parallel connections, so every group of six requests costs another round trip. Twelve blocking requests on a 150ms connection is 300ms before a single useful byte arrives.
What this calculator does not do
This is a planning model, not a measurement. Real loading involves caching, compression, connection reuse, priority hints and HTTP/2 multiplexing.
It assumes a cold cache and a single page load. Repeat visits are much faster and are not what this is for.
It cannot tell you which bytes to cut. Use it to set a target, then measure the actual result in PageSpeed Insights and WebPageTest.
Questions
What is a good LCP?
2.5 seconds or less is Google's threshold for good. Above 4 seconds is poor.
Which connection should I design for?
Slow 4G at 1.6 Mbps is the conservative choice, and it is what Lighthouse simulates by default.
My overhead already exceeds the target. Now what?
Smaller images will not fix it. Reduce blocking requests, cut server response time, or both.
How much should the hero image get?
Roughly 45 percent of the budget is a reasonable split, with 30 percent to CSS and fonts and 25 percent to blocking scripts.
Related calculators
Image File Size Estimator
Estimate output size before exporting.
LIVECSS Clamp Calculator
Fluid font sizes in one line.
LIVEAspect Ratio Calculator
Resize without distorting.
LIVESomething wrong with this calculator? Report it and it gets fixed or taken down.
