Why we still hand-code marketing sites in 2025

Roughly twice a month, a prospective client asks us: "Why don\'t you just use Webflow / Wix / Squarespace / WordPress? It would be cheaper, right?"

It\'s a fair question. Page builders have improved dramatically. For a real category of project — a personal portfolio, a side project, a one-page event site — they are unambiguously the right tool. We are not page builder haters. We use them when it makes sense.

But for the kind of work our clients hire us for — a marketing site that\'s the digital front door of their business — we keep choosing to hand-code. This article is the long answer to why.

The five-year math

The pitch for page builders is that they cost less upfront. That part is true. A Webflow site might cost $4,000 to set up, versus $18,000 for the same site hand-coded. The numbers favor the builder by a wide margin.

Until you look at five years.

The question isn\'t "what does this site cost to build?" — it\'s "what does this site cost to own?"

Over five years, every marketing site we\'ve looked at incurs costs in three categories: hosting and platform fees, content updates and small redesigns, and major rebuilds when the site stops fitting the business. Page builders front-load the savings on the first category but tend to inflate the other two.

Webflow\'s "Site" plan is $14–$39 per month depending on the tier. That\'s $840–$2,340 over five years just to keep the site online. Our hand-coded sites typically run on $5/month static hosting (Netlify, Cloudflare Pages, Vercel\'s hobby tier) — $300 over the same period.

Content updates are where it gets interesting. On Webflow, a non-technical marketing team can update copy, but anything beyond text — a new section, a different layout, an image gallery treatment — requires going back into the Designer, which has a non-trivial learning curve. In practice, every Webflow client we\'ve inherited has been paying their original agency $150–$300/hour for updates that should be a 15-minute job.

A hand-coded site with a properly integrated headless CMS (we use Sanity for most projects) gives marketers a clean, focused content editor. They can update text, swap images, add new posts, and even rearrange page sections — without ever touching code. We see content teams self-serve roughly 90% of their updates after we hand off.

What "performance" actually means for business

Page builders bundle a lot of JavaScript. They have to: they\'re generic platforms supporting every possible feature, so the runtime carries the weight of features your specific site doesn\'t use. A typical Webflow site loads 600KB–1.2MB of JavaScript before the page is interactive.

A hand-coded version of the same content typically loads 0KB of JavaScript. The HTML and CSS arrive, the page is done.

This isn\'t a vanity metric. Google\'s Core Web Vitals affect search rankings. A site that fails the "Largest Contentful Paint" or "Interaction to Next Paint" thresholds will rank below an otherwise equivalent site that passes them. We\'ve had clients move from page 2 to page 1 of search results within 60 days, just by replacing a slow Wix site with a fast hand-coded one. The content didn\'t change. The performance did.

For e-commerce, the math gets even sharper. Amazon famously calculated that every 100ms of latency cost them 1% in sales. Walmart, BBC, Pinterest, and a dozen other companies have published similar numbers. If your site converts 2% of visitors and you can shave 1.5 seconds off the page load, you\'re probably looking at a 10–20% increase in conversion. That math compounds across years and outweighs almost any "savings" on the initial build.

Accessibility isn\'t optional anymore

The ADA-related lawsuit count over inaccessible websites tripled between 2018 and 2024. Most of those suits are against larger companies, but small businesses are increasingly being targeted too. WCAG 2.1 AA compliance isn\'t just good ethics — it\'s now a legal floor in much of the developed world.

Page builders make accessibility hard. They render to the DOM in ways the builder controls, so when you find a screen-reader issue with a hamburger menu or a modal dialog, you often can\'t fix it — because the builder generated the markup, not you. The official answer is usually "wait for them to fix it" or "use a third-party plugin that may or may not work."

Hand-coded sites are accessible by default if you write them properly. Semantic HTML, ARIA where needed, focus management, keyboard navigation, color contrast — all of these are controllable line-by-line. We bake them in.

"But what about WordPress?"

WordPress is different from the SaaS page builders, and it deserves its own response. WordPress powers 43% of the web, the ecosystem is enormous, and the platform itself is fine. We\'ve built WordPress sites for clients who specifically needed them, and we will again.

The problem with WordPress isn\'t WordPress — it\'s the typical implementation. A theme bought from ThemeForest, a page builder plugin like Elementor or Divi, a dozen plugins for SEO and forms and caching and analytics... each plugin is its own attack surface, its own performance cost, its own thing that might break next year when WordPress releases version X.Y. We\'ve unfucked many of these.

A well-built WordPress site — minimal plugins, custom theme, sensible hosting — is a totally legitimate choice. But by the time you\'re paying for that level of craft, the cost difference between WordPress and a hand-coded headless site is small enough that other factors matter more.

When we recommend a page builder

To be clear: we do recommend page builders in some cases, and we\'ll tell prospective clients to use one rather than hire us. The cases:

• Personal sites, portfolios, side projects — anywhere the budget is under $3,000 or the site doesn\'t need to convert. A Squarespace site is fine.
• One-page event or campaign sites with a 6-month lifespan. Performance and SEO matter less; speed-to-launch matters more.
• Internal-only sites where load times and accessibility have lower stakes.
• Founders or solopreneurs who genuinely want to maintain the site themselves and don\'t have technical help.

If you fit one of those, save your money. If you don\'t, please consider that "cheaper to build" and "cheaper to own" are not the same number.

---

Got questions about an existing site or a project you\'re scoping? We\'re happy to look at it, no obligation. Get in touch.