The 40-point standard: how we score website builders

Why a fixed rubric

This website builder testing methodology exists so that every score on this site can be traced back to a rule written before the score was awarded. Six criteria, fixed weights, one total out of 40. The alternative — assessing each platform on whichever qualities happen to suit it — produces rankings that cannot be checked, cannot be compared across revisions, and cannot be argued with on any basis except taste.

A fixed rubric has one visible cost, and it is worth stating plainly. It will sometimes rank a platform below one you would have chosen, because it refuses to reweight itself around a favourite. GoDaddy takes a full 8 out of 8 on ease of use, the only builder to do so, and finishes last overall. Webflow takes the only full mark on SEO and performance, 7 out of 7, and finishes third. Those results are features of a fixed rubric, not defects in it. A reader who weights design at 20 points instead of 8 is welcome to recalculate; the working is published precisely so that recalculation is possible.

Scores are assessments of documented capability and published pricing, checked against vendor documentation and, where available, third-party measurement. Where sources conflict — Hostinger's ecommerce product cap is reported as 500 by some and 1,000 by others — the range is stated rather than resolved by preference. Where a figure could not be confirmed, it is described as an estimate at the time of checking, and it is not used to move a score.

The six criteria and their weights

The weights are not equal, because the risks they cover are not equal. Two criteria carry 8 points, two carry 7, and two carry 5. They sum to 40.

Ease of use — 8 points

Measured: how quickly someone without design or development training gets from an empty account to a published site that looks intentional and works on a phone. Onboarding, editor model, mobile handling, and how easily the editor lets you make a mess are all in scope. This carries the joint-highest weight because a builder's capability is worthless if the build never finishes, and the most common failure mode is abandonment rather than a bad result. The spread on this criterion is the widest in the standard, level with SEO: GoDaddy takes 8, Webflow takes 3.

Design & templates — 8 points

Measured: the quality and range of the stock templates, and how much control remains after you pick one. A large library of dated templates does not score the same as a small library of good ones. Reversibility counts here too — Wix and Hostinger cannot switch template after publishing, which means a redesign is a rebuild. Equal weight to ease of use, because for most sites the visual output is the product: it is what a visitor uses to decide whether the organisation behind it is credible.

SEO & performance — 7 points

Measured: the technical controls a platform actually exposes — editable meta, custom slugs, 301 redirects, canonical tags, structured data, sitemap and robots control, head-code injection — plus published third-party performance evidence. The evidence base is the HTTP Archive Core Web Vitals Technology Report of November 2025, in which 74.86% of Wix sites and 70.39% of Squarespace sites passed, behind Duda at 84.87%. Shopify, Webflow and GoDaddy were not named in that report cut, so no figure is invented for them and their scores rest on documented capability alone.

Seven points rather than five, because the ceilings here are structural. A missing canonical control or an unavailable redirect manager cannot be added later by trying harder, and a forced URL structure — Shopify's /products/ and /collections/ paths — cannot be flattened at any price. Seven rather than eight, because a large share of sites never need to compete in organic search, and a rubric that weighted SEO above ease of use would misprice those sites.

Value for money — 7 points

Measured: the total cost of a comparable small-business site across two years, not one — a mid-tier plan, one domain and one business mailbox, priced for year one and again for the renewal year. Transaction fees, mandatory add-ons and term-length conditions are part of the assessment. Seven points, equal to SEO, because the gap between the advertised price and the real one is the single largest source of error in this category. In the standard two-year scenario Hostinger rises 513% between year one and the renewal year and GoDaddy rises 136%, while Wix, Squarespace, Shopify and Webflow move by 5% or less. A criterion weighted on headline prices alone would have inverted that result.

Ecommerce & extensibility — 5 points

Measured: selling capability — catalogue limits, payment options, platform fees — together with the depth of the app and integration ecosystem, which is what determines whether a platform can be extended past what it ships with. Five points, not more, because this criterion is decisive for a minority of sites and irrelevant to the rest. Weighting it at 8 would have pushed a commerce platform to the top of a general website builder ranking, which would be a category error. Anyone whose site exists to sell should read this column instead of the total.

Support & lock-in — 5 points

Measured: two things that matter at the same moment. Support coverage — channels, hours, whether a phone line exists at all — and exit rights, meaning whether you can take the site with you. These are combined because both are tested when something has gone wrong. Webflow scores 4 here despite what appears to be the thinnest support coverage of the six, because it is the only platform offering genuine code export. Five points, not more, because most site owners contact support rarely and migrate once or never; five points, not fewer, because when either becomes relevant it is expensive.

How scores convert

The six criterion scores are added. The total out of 40 is divided by four and given to one decimal, rounding half up. There is no weighting applied after the fact, no bonus, and no adjustment for how well a platform is regarded. Squarespace scores 31, and 31 ÷ 4 = 7.75, which rounds to 7.8. Webflow scores 29, and 29 ÷ 4 = 7.25, which rounds to 7.3. Wix scores 32, which divides exactly to 8.0.

The total is the authoritative figure and the headline score is a convenience. Where they appear to disagree, the total governs: Shopify and Hostinger both convert to 7.0, but they are not equally placed, because they hold different criterion scores underneath an identical total. A worked example of the full arithmetic, criterion by criterion, is set out in how we score: the rubric explained.

The documented tiebreak

Where two builders finish on the same total, the higher SEO and performance score takes the higher rank. The rule was fixed before any score was awarded, and it has no discretionary component. SEO is the tiebreak criterion because it is the most heavily weighted of the non-cosmetic criteria and the one whose limits are hardest to work around after the fact.

It applies once in revision 2026.8. Shopify and Hostinger both total 28 of 40, a headline score of 7.0 each. Shopify holds 5 out of 7 on SEO and performance; Hostinger holds 4. Shopify therefore ranks fourth and Hostinger fifth in the best website builders standings. The two are otherwise very different — Hostinger takes the only 7/7 on value, Shopify the only 5/5 on ecommerce — which is the case for reading the criterion scores rather than the position.

What we do not score

Brand reputation is not a criterion. Neither is market share: on the W3Techs survey of 21 August 2026, Wix runs 4.2% of all websites and Webflow 0.8%, while Duda holds 0.7% and leads every platform measured on Core Web Vitals at 84.87%. Popularity and measured performance are separate axes, and only one of them is testable. Note also that 30.9% of all websites use none of the systems W3Techs monitors, which is the honest denominator for any share figure.

Affiliate rates are not a criterion either, and this is the disclosure that matters most. Commission structures across these six platforms differ by an order of magnitude in both rate and duration. A ranking optimised on expected revenue would not resemble this one, and the fact that the highest-paying programs do not occupy the top positions is the observable check on that claim. The commercial terms are set out in the affiliate disclosure. Also excluded: unverified uptime SLAs, hands-on load-time measurements we have not made, and user sentiment, none of which are reproducible from published sources.

Revisions and corrections

The standard is re-checked monthly and each pass carries a revision number. The current one is 2026.8. Criteria and weights are held constant between revisions so that totals remain comparable; only the scores move. Changing a weight would be a change to the standard rather than to a score, and would be published as an amendment with the reasoning attached.

Two things trigger a re-score between revisions: a pricing or plan change, and a capability shipping or being withdrawn. Webflow's 13 May 2026 restructure, which merged the old CMS and Business site plans into a single Premium tier, is an example of the first; native canonical tags and noindex arriving in June 2025, replacing a custom-code workaround, is an example of the second. Corrections are recorded rather than silently applied, and a corrected figure that changes a total also changes the revision number.