What a Fast Website Is Actually Worth
Everyone agrees fast is better. Almost nobody can say what a second is worth to their business. Here is how to work it out, and what to fix first once you have.
"Fast" is the least controversial recommendation in web development and one of the least acted on. Everyone nods. Almost nobody funds it, because nobody has put a number on it.
Here is how to put a number on it, and what to do once you have.
The arithmetic
Start with what you know:
- Monthly visitors to your key page
- Conversion rate on that page
- Value of a conversion
Multiply them and you have the page's monthly revenue. Now the only unknown is how much conversion rate moves with load time.
The published research clusters in a consistent range: somewhere between 3% and 7% relative conversion loss per additional second of load time, worse on mobile, worse on commerce, worse the slower you already are. Take the conservative end of that.
Say a services page gets 2,000 visits a month, converts at 3% into inquiries, and one in four inquiries becomes a $4,000 project. That is 60 inquiries, 15 projects, $60,000 a month.
Cut a second off the load time at a conservative 4% relative improvement, and that is roughly 2.4 extra inquiries and about $2,400 a month. Nearly $29,000 a year, for one second.
Now the question "is it worth two days of engineering time to remove a second" answers itself.
Run the numbers with your own inputs before you believe mine. The point is not the specific figure — it is that the figure is knowable, and once it is on paper the conversation stops being about taste.
Where the seconds actually are
Once you are motivated, the next question is where the time goes. Almost always in the same four places.
Images. Still the biggest offender on most sites. An unoptimised hero image can be larger than every other asset on the page combined. Serve modern formats, size them to their display size, and give them explicit dimensions so the layout does not shift when they arrive.
JavaScript that is not needed yet. A chart library on a page with no chart above the fold. An animation library for an effect that fires on scroll. A 3D scene behind a headline. Every one of these can be loaded on demand instead of up front. This is usually the single largest win available on a modern framework site.
Fonts. A custom font that blocks rendering costs you the entire visible page. Use a display strategy that shows fallback text immediately, preconnect to the font host, and subset to the characters you use.
Third-party scripts. Analytics, chat widgets, tag managers, heat maps. Each one is a request to a server you do not control, running code you did not write, on your main thread. Audit them yearly. Most sites are carrying at least one script nobody remembers adding and nobody looks at.
Measure the right thing
Two distinctions matter here, and getting them wrong wastes effort.
Lab data versus field data. A synthetic test on a fast connection tells you about your code. Field data from real visitors tells you about your business. They disagree constantly, and the field data is the one that counts — it includes the phone on a bad connection, which is where your conversions are being lost.
Load time versus responsiveness. These are separate failures. A page can paint quickly and then be unresponsive for two seconds while JavaScript hydrates. The visitor taps a button and nothing happens. They tap again. From their side that is a broken site, and it does not show up in a load time number at all.
Track the paint metric and the interaction metric separately, from real users.
Fix in this order
Assuming a typical site, this is the order with the best return per hour spent:
- Compress and correctly size images. Highest impact, lowest skill required, almost always available.
- Defer the JavaScript that is not needed for first paint. Anything decorative, anything below the fold, anything triggered by an interaction.
- Fix font loading. A display strategy and a preconnect. Fifteen minutes.
- Remove third-party scripts nobody uses. Free performance, plus a smaller privacy surface.
- Reserve space for anything that loads late. Explicit dimensions on images, embeds, and ad slots. Layout shift is the cheapest metric to fix and one of the most annoying to experience.
- Only then look at framework-level optimisation. Server rendering strategy, bundle splitting, caching headers. Real gains, but not before the first five.
Set a budget and hold it
The reason performance work does not last is that nobody guards it. The site gets fast, and then over six months a chat widget, two tracking pixels, and a hero video arrive, and it is slow again.
A performance budget is a number you agree not to exceed — total JavaScript on the homepage, or a target for the paint metric — checked automatically on every deploy. When a change would break it, that is a conversation before the merge rather than a mystery two quarters later.
The budget does not have to be aggressive. It has to exist.
The part that is not about numbers
There is a second argument for speed that does not show up in conversion arithmetic, and for a small business it might matter more.
A fast site feels competent. A slow one feels neglected, and visitors generalise from it. If your site takes four seconds to show anything, a prospect forms an impression about how you handle detail before they have read a single word about you.
You cannot put that in a spreadsheet. But it is the reason the fastest site in a given local market often is not the biggest company — it is the one that cared.