The Five-Round Card
Five rounds, no weighting, one decimal. This is the method behind every score on the site — what each round measures, where the evidence comes from, and how a decision gets written.
Here is how we score website builders, in full, so you can check the working. Every fighter on this site is judged over five rounds — Ease, Design, Value, Power and Support — each marked from 0 to 10. The overall is the mean of those five, carried to one decimal place. Nothing is weighted and nothing is adjusted after the fact. The card is identical on a solo profile and inside a head-to-head, which is the only reason the numbers are comparable at all.
A method is only worth publishing if it can embarrass us. So this page also states where the evidence comes from, what would force a score to move, and what the card deliberately refuses to measure.
The five rounds
Round one — Ease
How far a competent non-developer gets on their own. Not "is there a drag-and-drop editor" — everything has one. The question is how much of the platform a person can reach without reading documentation, how quickly they get from signup to a page they would show a client, and how badly the thing punishes a mistake. Ease is the widest-spread round on this site, running from 4.5 to 9.0, and it decides more bouts than Design does. A high Ease score is not a compliment about simplicity; it is a measurement of how much capability survives contact with a beginner.
Round two — Design
Both ends of the range: the floor and the ceiling. The floor is what you get for free — template craft, typography, art direction, how much of the visual work is already done before you touch anything. The ceiling is how far a capable designer can push the platform before it stops them. A builder with a thousand mediocre templates does not beat one with two hundred excellent ones, and a builder with total CSS control still needs somewhere decent to start. Design is where the top of this card compresses hardest: three platforms sit on 9.5 for very different reasons.
Round three — Value
Not the cheapest headline price. What the plan a real buyer actually needs costs, at the billing cycle they will really use, over the life of the site. That means renewal rates count — a $2.99 introductory price that becomes roughly four times that at renewal is scored as what it becomes. Seats count. Workspace fees count. Ecommerce add-ons count. Transaction fees count, because a 7 percent cut of digital sales is a price whatever the vendor calls it. Value is the round where marketing pages and reality separate most reliably.
Round four — Power
The ceiling on what the site can become. CMS depth and item limits, structured content and reference fields, ecommerce capability, technical SEO control including redirects, canonicals, schema and localization, integrations and APIs, staging and versioning, multi-site and team tooling. Power is graded against published plan limits rather than feature-page adjectives, and against the tier a buyer would have to reach to actually use the feature. A capability that only exists on an enterprise contract is scored as an enterprise capability.
Round five — Support
Channels, hours, and whether a human is reachable on the day it breaks. Phone availability matters more than vendors think it does, and the absence of it is scored as an absence, however good the written documentation is. Hours matter too: weekday-only chat is a materially different product from 24/7 cover. Self-serve resources count in this round — a genuinely excellent free video curriculum lifts a score — but they do not fully replace a person, because the situations that generate support tickets are usually the ones documentation did not anticipate.
| Round | What it measures | How evidence is gathered |
|---|---|---|
| Ease | How far a competent non-developer gets unaided, and how fast they get a publishable page | Editor model, onboarding flow and the documented learning path, weighed against the consensus of 2026 reviewers |
| Design | The floor and the ceiling of the output: template craft, art direction, and how far a designer can push it | Template libraries counted and inspected, plus the editor's real capability — grid, free-form, or CSS-level control |
| Value | What the plan a real buyer needs costs over the life of the site | Live vendor pricing on the date shown, annual and monthly, plus renewal terms, seats, add-ons and transaction fees |
| Power | CMS depth, ecommerce, technical SEO control, integrations, team and multi-site tooling | Vendor-published plan limits and help-centre documentation, checked at the tier where the feature actually unlocks |
| Support | Channels, hours, and whether a human is reachable when the site is down | Official support policies and stated hours, cross-checked against what customers report through 2026 |
How the overall is calculated
Add the five round scores, divide by five, round to one decimal. That is the entire formula. Squarespace scores 8.5, 9.5, 6.5, 7.0 and 7.0, which sums to 38.5 and gives an overall of 7.7. You can do that arithmetic on any fighter on the site and it will come out to the number printed on the page — the tale-of-the-tape component recalculates it from the axes so the two can never disagree.
No weighting is applied, and that is a deliberate choice rather than laziness. Weighting encodes an assumption about who is reading, and the reader who needs a 24-hour phone line is not the reader who needs 20,000 CMS items. An unweighted mean lets you apply your own weighting: ignore the overall, find the round that decides your project, and read that column instead.
How a bout is decided
Matchups are not settled by comparing two overalls. They are scored round by round. Each round goes to the higher number, the fighter who takes more rounds takes the decision, and the overall is reported alongside as context rather than as the verdict. This is why a builder can lose a bout while carrying the higher overall in a different one, and why the round-by-round on each bout page is the part actually worth reading. Every published decision is on the full website builder comparison card.
Every bout also carries a use-case breakdown, because a decision on points is not a recommendation for everybody. A builder can take three rounds out of five and still be the wrong choice for your specific job, and the bout page has to say so plainly rather than hiding behind the scoreline.
What a drawn round means
A drawn round is an identical score, and it counts for neither corner. It is not a hedge. When Squarespace and Webflow both take 9.5 for Design, that is a finding: two platforms reaching the same standard from opposite directions, one by handing you finished art direction and one by handing you the tools to make your own. Reading that as indecision misses the useful part.
A whole bout can be drawn too. If the rounds split evenly and the two overalls sit within 0.2 of each other, the decision goes in the book as a draw. Two of the five bouts on the current card ended that way, and in both cases forcing a winner would have meant inventing a preference the evidence does not support.
Why the losing corner's strengths get published
Every bout page lists what the losing side does well, at the same length and in the same format as the winner's. Partly because it is true — a builder that takes 4.8 overall still posts an 8.5 for Ease, and pretending otherwise would be dishonest. Mostly because the loser wins some readers. If the round you care about is the one the loser took, you need that stated clearly, not buried under a verdict written for somebody else.
It is also a check on the scoring. A one-sided page is usually a badly researched one. If a fighter's strengths column is hard to fill, that is a signal to go back to the evidence rather than to publish the beating.
When a bout gets rescored
Four triggers force a rescore. A price change, including a renewal rate change or a new transaction fee. A plan restructure that moves what a tier includes. A capability change — a feature shipped, a feature retired, an app pulled. Or a support change, such as a channel closing or hours being cut.
2026 has produced all four. Squarespace raised prices in July by up to 26 percent, applied to existing subscribers at renewal. Webflow collapsed five site plans into three in May, which is why a great many pricing tables published elsewhere still describe a ladder that no longer exists. Square announced a 67-country wind-down of Weebly in June. Framer refreshed its limits and cut seat prices in May. Each of those moved at least one round, and every affected bout was rescored and redated rather than quietly patched.
Where sources genuinely conflict, the page says so instead of picking a side. A disputed price printed confidently is worse than no price at all, and readers can tell the difference.