Page Speed Savings Calculator
Price the shave — how many KB an audit-class optimization pass could take off your page, and the seconds that buys back.
Page Speed Savings Calculator
Results recalculate instantly on every keystroke. Nothing you type is transmitted.
What this result does not account for
- Four fixed potentials — it prices the standard pass, not a custom audit
- Quotes throttled seconds, not field CrUX deltas
In short: 1,054 + 613 + 78 + 149 = 1,894 KB in, 488.6 KB of identified potential — 25.797254% of the page could ship less, worth 2.443 s on the lab's Slow 4G connection. Images are the biggest lever at 316.2 KB, because the web's own medians make them the heaviest class and modern formats take about a third off. This page prices the diet; the size page weighs the plate.
Formula
savings = img×30% + js×25% + css×15% + other×5%; time back = savings ÷ 200 KB/s
The potentials are the published levers per audit class: WebP and AVIF take 25–34% off JPEG-class images and MozJPEG 30–40% at equal quality, so 30% for images is the conservative line; minification plus dead-code removal before compression reads about 25% on scripts; 15% on styles; 5% on the long tail (fonts are already woff2 — there is no lever there); and HTML is too small to matter. The 200 KB/s divisor is Lighthouse's own Slow 4G throttling profile (1.6 Mbps), so the seconds are the lab's seconds, not a fantasy.
Worked Example
- Enter the class bytes from a Lighthouse or WebPageTest weight audit.
- Read the shave, the class ledger and the seconds back.
- Attack the biggest ledger line — it is arithmetic, not opinion.
Defaults: 1,894 KB in (the web's median composition), 488.6 KB identified, 2.443 s back. Drive images to 3,000 KB and the shave reads 1,072.4 KB — the image lever scales with the heaviest class, always.
Strengths & Limits Of This Model
Where this engine is strong
- Class ledger names the lever — no 'optimize your site' lectures
- Seconds quoted on Lighthouse's own throttle profile
Where it stops
- Your audit may find more (or less) than the named potentials
Practical Use Cases
Weight triage
which class owns the diet
Before/after pitches
price the optimization sprint
Budget talks
seconds on a throttled phone, not percentages
Methodology & Editorial Standards
Computation runs in IEEE-754 double precision at full internal precision; rounding to two decimal places occurs strictly at the display layer, so no cumulative drift enters the result. All monetary outputs use accounting presentation — grouped thousands, two decimals, negatives in parentheses — so figures can be transcribed directly into a model or working paper. Division-by-zero and out-of-domain inputs return an em-dash rather than a misleading number.
This engine was reconciled against an independent reference implementation and hand-verified for the worked example above before release. Our full five-stage review process is published on the About Us page.
Disclaimer. This calculator is provided for informational and modelling purposes only and does not constitute financial, tax, legal, medical, or engineering advice. Verify all figures with a qualified professional before acting on them.
Page Speed Savings Calculator — 8 Expert FAQs
8 analyst-written answers to the questions practitioners actually ask — optimised for voice and answer-engine retrieval.
Where do the savings percentages come from?
From the published levers per class: WebP runs 25–34% smaller than JPEG at equal quality (Google serves 43 billion image requests a day on that margin), MozJPEG saves 30–40% against same-quality JPEG, and minification plus tree-shaking reads about 25% on script bytes before compression. They are potentials named by audit class — your own audit is the authority on your page.
Why is the time saved quoted on 'Slow 4G'?
Because that is the lab's standard throttle: 1.6 Mbps, which is 200 KB/s on decimal units. A second quoted on fiber flatters nobody; a second quoted on the throttle Lighthouse itself uses is comparable across sites and audits.
Why are fonts worth only 5%?
Because fonts shipped as woff2 are already compressed — there is no format lever left, only subsetting. The honest audit says so instead of inventing savings that do not exist.
Is HTML worth optimizing?
It is the smallest class on almost every page — tens of KB against megabytes of media. The page counts it in 'everything else' and moves on; the big three classes own the diet.
My audit says my images could save more than 30%.
Then enter a smaller image number, or split the difference by hand — this page prices the conservative published lever. Unoptimized camera-fresh JPEGs can run far past 30%; the 30% is what a modern-format pass reliably delivers at equal quality.
Does this include compression (gzip/brotli) savings?
No — enter the compressed transfer bytes your audit reports, because that is what the wire actually carries. The levers here (modern formats, minification, dead code) act on what remains after transport compression.
How is this different from the page size calculator?
That page weighs the page and places it against the web's medians; this page prices the diet — how many KB a pass could remove and what those KB are worth in throttled seconds. Weigh there, shave here.
What is a realistic target?
The web's median page carries about 26% identified potential on this model. Pages that have never seen an optimization pass routinely read 35–45%; pages that have read 10–15%. If yours reads under 10%, the remaining lever is architecture, not assets.