By the SI Systems team
A lightweight HTML website for a small business is a page that ships only the bytes a visitor needs: clean markup, a small stylesheet, a few optimized images and almost no JavaScript. The median page on the web now weighs more than 2.3 MB. A plain, well-built HTML page can come in at a small fraction of that, and the public data shows why that gap matters to customers, search visibility and your own maintenance bill.
If you edit HTML by hand in a browser editor, you already have an advantage. This guide uses the HTTP Archive's 2024 Web Almanac and Google's Core Web Vitals guidance to show what "heavy" looks like, what a lean small-business page should aim for, and where automation and AI agents can take over the repetitive parts.
The Web Almanac measured real pages in October 2024. The numbers below come straight from the Page Weight chapter.
| Measure (2024) | Desktop | Mobile | Source |
|---|---|---|---|
| Median page weight | 2,652 KB | 2,311 KB | Web Almanac 2024, Page Weight |
| Median page weight, 2022 | 2,312 KB | 2,037 KB | Web Almanac 2024, Page Weight |
| Change in 10 years (Oct 2014 to Oct 2024) | +120% | +357% | Web Almanac 2024, Page Weight |
| Median requests per page | 71 | 66 | Web Almanac 2024, Page Weight |
| Median homepage HTML | 18 KB | 18 KB | Web Almanac 2024, Page Weight |
| Median homepage CSS | 78 KB | 73 KB | Web Almanac 2024, Page Weight |
| Median homepage JavaScript | 613 KB | 558 KB | Web Almanac 2024, Page Weight |
| Median homepage images | 1,054 KB | 900 KB | Web Almanac 2024, Page Weight |
Source: HTTP Archive, "Page Weight," The 2024 Web Almanac, https://almanac.httparchive.org/en/2024/page-weight
Notice the middle of the table. The HTML itself, the thing you actually write, is about 18 KB at the median. Almost everything else is images, scripts and fonts layered on top. That is the practical lesson: on most pages the markup is not the problem.
The Almanac links weight to performance directly. Using Lighthouse lab tests from June 2024, it found that the heavier the page, the slower the Largest Contentful Paint, and that the effect is steeper on mobile.
| Percentile (June 2024 tests) | Desktop page weight | Mobile page weight |
|---|---|---|
| 10th | 549 KB | 471 KB |
| 25th | 1,138 KB | 995 KB |
| 50th | 2,157 KB | 1,938 KB |
| 75th | 4,169 KB | 3,766 KB |
| 90th | 8,375 KB | 7,680 KB |
Source: Web Almanac 2024, Page Weight, Figure 16.37.
On interactivity, the gap is larger. The Almanac reports Total Blocking Time, a lab proxy for INP, at a median of 53 ms on desktop but 1,235 ms on mobile, and 648 ms versus 5,786 ms at the 90th percentile. Google's target is 200 ms or less. Most of that blocking comes from JavaScript, which a small static page barely uses.
Google's web.dev documentation defines the targets and how to read them.
| Metric | Good | What it measures | Source |
|---|---|---|---|
| LCP | 2.5 s or less | Loading | web.dev, "Web Vitals" |
| INP | 200 ms or less | Responsiveness | web.dev, "Web Vitals" |
| CLS | 0.1 or less | Visual stability | web.dev, "Web Vitals" |
Source: https://web.dev/articles/vitals. Google recommends judging each metric at the 75th percentile of page loads, split by mobile and desktop. A page passes only when all three are met.
The Almanac cites performance researcher Alex Russell's 2024 analysis: a budget of about 365 KB for a page to load in under 3 seconds at the 75th percentile on typical mobile conditions, and about 650 KB for a 5-second load if the page is JavaScript-heavy. Treat these as planning guides, not guarantees. Against the 2.3 MB median, they show how much room a lean page has.
A reasonable personal budget for a service-business homepage:
The Almanac found that only 70% of desktop homepages correctly compress text resources, 62% correctly minify CSS and 58% correctly minify JavaScript. All three are one-time setup tasks. If your host offers Brotli or gzip, turn it on and check it.
The Almanac notes that layout shift grows with page weight and points at images defined without dimensions as a likely cause. Always write width and height attributes on every img tag.
Videos, chat widgets and social embeds load large scripts. The Almanac notes that 70% of desktop homepages had embeds that could be replaced with a facade, a static preview that loads the real embed only on click.
Keep the structure boring. A title, a meta description, one h1, a short list of services, proof, and one clear contact action:
That structure also happens to be easy for search engines and AI assistants to read, because the answers are in plain text.
A static page is not the answer to everything. Add real software when you need:
This is where an AI agent or small automation earns its place: it handles the repetitive replies and data entry behind a fast front end, instead of bolting heavy plugins onto the page. If you want that built for you, the SI Systems team builds websites, AI agents and apps; you can see the services and how the process works. The current offer, "Launch at Zero Cost," covers a website built, launched and maintained for 6 weeks with payment only at the end. Details are on the SI Systems homepage.
What is a lightweight HTML website? A site built from plain HTML and a small stylesheet, with minimal scripts and compressed images, so each page transfers only a few hundred kilobytes.
How big should a small-business homepage be? There is no legal limit. The 2024 Web Almanac median is 2,311 KB on mobile. Planning budgets cited in the same chapter are around 365 KB for a quick load, so aim well under 1 MB if you can.
Does page weight affect Google rankings? Google surfaces Core Web Vitals in its tools and says they reflect real user experience. They are one signal among many. Content relevance still matters most, but a slow, jumpy page loses visitors before they read it.
Is static HTML bad for SEO? No. Static pages are easy to crawl and usually fast. You still need good titles, headings, clear copy and a sitemap.
When should I move beyond static HTML? When you need bookings, payments, logins, or automation such as lead replies. Keep the public pages lean and add those features as separate, well-scoped pieces.
Do I need JavaScript at all? Often not for a brochure site. 97.8% of mobile homepages use it, and median JavaScript is 558 KB on mobile, but a contact link and a phone number work without any.
How do I measure Core Web Vitals? Use PageSpeed Insights or Search Console for real-user data from Chrome's field dataset, and Lighthouse in DevTools for lab tests. Google recommends field data as the final judge.
Written by the SI Systems team. Start a conversation at https://www.sisystems.site/#contact.