How the scoring works.

Most comparison sites work backwards: pick the winner, usually whoever pays the highest commission, then justify it. SiteVersus has no affiliate links, so the only way to prove that is to show the machinery. If you disagree with a verdict, you can see exactly which inputs produced it, and tell us which input is wrong.

Website platforms a real team might build on in 2026: hosted builders (Wix, Squarespace), hosted CMSs (Webflow, Shopify), self-hosted CMSs (WordPress, Drupal, Joomla), and headless systems (Contentful, Sanity, Strapi). Each platform’s type is listed prominently: comparing Wix to Strapi head-to-head is a category error, and the tool says so instead of pretending one wins.

• 124 attributes per platform across 15 categories: pricing, editor, design, hosting, performance, analytics and insight, SEO controls, content, ecommerce, ownership and lock-in, extensibility, AI and AEO, collaboration, support, and multilingual.
• Every factual claim carries a source URL and a verified date. Official vendor documentation and pricing pages are the preferred source; hands-on testing fills the gaps. We never cite other comparison or affiliate sites.
• A handful of attributes are explicitly editorial (learning curve, design flexibility, lock-in risk, documentation quality) and are labeled as judgments with reasoning, never presented as facts.
• Unverified is shown as unverified. We never guess silently.
• The whole dataset is open. Download it at siteversus.com/data or browse it on GitHub, licensed CC BY 4.0. The public commit history is the proof that the rubric predates any self-interested change.

Each scored category rolls up to a 0–5 from its attributes; editorial attributes never contribute more than 20% of a category score. There is no global “best platform” number, because a single overall score is how affiliate sites launder bias. Verdicts are use-case-weighted: category scores multiplied by the weight profile for what you are actually building.

Every attribute becomes a 0 to 5 by a fixed rule, published here so a score is never a matter of opinion about the number, only about the underlying fact (which you can dispute).

• Yes / no attributes: yes is 5, no is 0. A few are inverted, where the good answer is no (for example, renewal pricing that jumps after the intro term).
• Graded enums map on a fixed scale: for example custom code as yes / plan-gated / no scores 5 / 2.5 / 0; content export as full / partial / none scores 5 / 2.5 / 0; lock-in risk as low / moderate / high scores 5 / 2.5 / 0.
• Core Web Vitals: the share of real sites passing, divided by 20, so 100% maps to 5 and 60% maps to 3.
• Entry price bands: 5 at $5 or under per month, dropping a point roughly every $10 to $15, reaching 0 past $60.
• Category type, launch year, and open-source status are shown as facts but never scored, so no platform is penalized for being, say, headless rather than a builder.
• Unverified cells are excluded from the math entirely and the category weight renormalizes, so a missing value never silently counts as a zero.

• Priority order: official pricing and docs pages → hands-on testing → vendor changelogs → independent measurement such as HTTP Archive’s Core Web Vitals data. Never other comparison sites.
• Prices are USD, US market, annual-billing monthly equivalents. Intro-versus-renewal pricing games get their own attribute.
• Rolling corrections as vendors ship; a full re-verification sweep every quarter. Every page shows its data’s verified dates.

• No affiliate links today: none, not disclosed-affiliate, not “partner” links. If that ever changes, it will be announced on the changelog, disclosed on every link, and scores will keep coming from this public rubric, which is version-controlled, so you can check.
• No sponsored placements, no pay-for-listing, no vendor-supplied scores.
• SiteVersus is built by K Squared Consulting, an agency that builds client sites on many of these platforms. That is the point: we live with these tools’ consequences.
• We don’t edit Wikipedia or seed citations ourselves.

K Squared builds a CMS: BrightSite. It is brand new and early in its development, still building out core features. That is exactly why it is deliberately not listed at launch: scoring it honestly today would mostly document what it does not do yet. If and when it is mature enough to be added:

• It will be labeled “our product” on every page where it appears.
• It will be scored against this rubric as published before its listing; the version-controlled dataset proves which came first.
• It gets no weight changes, no new categories, and no new attribute definitions in the release that adds it.
• Its data gets the same per-claim sourcing, and disputes about its cells go through the same corrections process as everyone else’s.

Vendors and readers can dispute any value. Confirmed errors are fixed within 72 hours and logged publicly.

Report incorrect data See the changelog