Hand Coded vs WordPress vs Wix: Performance Data From 130 Builds

By Joseph W. Anady · Published 2026-04-21 · Last reviewed 2026-04-21 · 8 min read

Core Web Vitals moved from nice to have to primary ranking signal in March 2024 and tightened in the March 2026 core update. The LCP threshold dropped from two point five seconds to two point zero seconds. INP replaced FID as the primary interaction metric with a two hundred millisecond good threshold.

Under the new thresholds, thirty to forty percent of WordPress sites in the portfolio would fail without active optimization work. The same percentage of Wix sites fail because the platform caches are not always tuned for the newer thresholds. Hand coded sites pass because every line of code is included for a reason and nothing is loaded that will not be used.

The compounding effect matters. Every second of additional load time is worth roughly seven percent of conversions. A hand coded site loading in zero point seven seconds against a WordPress competitor loading in three point five seconds has a structural advantage of roughly twenty percent in conversion rate before any other factor is considered.

The WordPress plugin ecosystem is simultaneously its greatest strength and its greatest weakness. Estimates place the number of actively exploited WordPress plugin vulnerabilities at several hundred at any given time. Patch cadence varies by plugin maintainer. Abandoned plugins remain installed on millions of sites.

Wix transfers security responsibility to the platform, which is better than managing plugins individually but creates a different risk. Platform lock in means a security incident at Wix affects every customer simultaneously. Platform deprecation of features has no recourse for the customer.

Hand coded sites on Bubbles ship with HTTP Strict Transport Security with preload, X Content Type Options nosniff, Referrer Policy strict origin when cross origin, Permissions Policy restricting camera, microphone, and geolocation access, and X Frame Options same origin. All five headers come from the server config. No plugin dependencies.

The Yoast versus RankMath debate inside the WordPress community is a conversation about which plugin best approximates what a hand coded site ships with by default. Both plugins add schema, meta tags, and sitemap generation. Neither reliably produces the structural answer capsule plus speakable pattern that AI engines reward.

Wix produces passable SEO output for most small business scenarios. It does not produce llms.txt or llms-full.txt. It does not allow direct control of the schema graph at the @id level. It does not support Nginx SSI based schema injection because you do not control the server.

Hand coded sites in the portfolio pass all one hundred twenty six checks in the framework verification gate. The same checks running against a recent WordPress site with Yoast installed typically pass eighty to ninety percent, failing on the AI specific structural checks that Yoast does not yet implement.

The visible three year cost is only part of the calculation. A Wix Business Premium site at $2,124 total sounds cheap until you factor in the conversion rate delta against a hand coded site. On a service business booking fifty leads per month at two thousand dollars average client value, a twenty percent conversion rate disadvantage costs two hundred forty thousand dollars over three years.

WordPress plugin sprawl adds hidden costs beyond the visible subscriptions. Plugin conflicts consume developer time. Security incidents from plugin vulnerabilities cause emergency work. Performance plugins attempt to clean up the mess created by feature plugins, adding their own overhead.

The hand coded site with the monthly retainer at $250 or $397 includes ongoing optimization work. The retainer is a forward facing investment in ranking, AI citations, and conversion rate. It is not a maintenance fee masquerading as progress.

The content editor argument is real but often misapplied. A small business publishing one blog post per month does not need WordPress for editor ergonomics. A hand coded site with a simple markdown pipeline through a headless CMS solves the same problem with a fraction of the platform overhead.

The plugin argument is real when the plugin is load bearing. WooCommerce for a complex e-commerce operation, LearnDash for a course platform, or MemberPress for a membership site can justify WordPress. For a standard service business, the entire plugin dependency tree is replaceable with a few hundred lines of hand written code.

Wix wins for the client who genuinely cannot afford ongoing work and who values a simple drag and drop interface over structural optimization. This client exists. For every other client, the three year math favors the hand coded approach.

The migration mechanics are straightforward when the source data is clean. Page content, URL structure, and inbound links transfer cleanly. The rebuild adds the answer capsule, speakable, schema, and AI specific files that the original WordPress site lacks. Redirects from old URLs to new URLs preserve ranking signal.

The disruption window is two weeks of build plus one week of DNS propagation and monitoring. During that window, the original site stays live. Cut over happens in a maintenance window usually under one hour. Organic traffic typically returns to baseline within seven days and exceeds baseline within thirty days of re-crawl.

The ROI calculation is specific to each business. A free audit grades the existing site across the fourteen tier framework and projects the traffic and conversion lift that a rebuild would deliver. The document is the decision support tool for the migration or stay choice, not a sales pitch.