New: UX Signal for Figma
Sign up
Home
The checklist

All 56 heuristics, in the open.

Every rule we score a page against, across 21 weighted categories. The categories and weights render from the same configuration the audit runs on, so this page cannot drift from what actually scores your site.

The rules are our own. Each one names the published principle it stands on and links to it for further reading. Those sources are independent of UX Signal and do not endorse it.

Where the 100 points live

Each column is one experience type's rubric; each cell is the category's point weight there. Darker means more of the score rides on it.

Marketing siteE-commerceMarketplaceProductClarity & Messaging151515—Navigation15151515Visual Hierarchy / First Impression101010—Lead Generation Friction10———Content & Copy Quality10———Trust & Credibility101010—Accessibility10101010Mobile Experience10101010SEO-UX Signals5———Performance5555Checkout Friction—10——First Purchase Experience—10——Form & Error UX—5—5Listing & Conversion Friction——10—Search & Discovery——10—Content Quality——5—Clarity & Microcopy———12Scanability & Hierarchy———10Task & Workflow Efficiency———13Onboarding & Empty States———10System Feedback & Recovery———10

Weights in each column sum to 100 — the whole score, allocated before the audit starts. 56 heuristics feed these 21 categories; the full checklist follows.

Clarity & Messaging

15% of score

01Offer stated above the fold — The page says what it is and who it serves before any scroll.Based on NN/g: Inverted pyramid (opens in a new tab)

02Audience named — Copy identifies who this is for, not only what it does.Based on Laws of UX: Selective attention (opens in a new tab)

03CTAs promise outcomes — Buttons say what happens next — never bare “Submit” or “Learn more”.Based on NN/g: “Learn More” links (opens in a new tab)

A visitor decides whether to keep reading in a few seconds. If the page cannot say what it is and who it is for in that window, everything below it is wasted. Typical failure: A hero that names the company but never the offer.

Applies to Marketing site · E-commerce · Marketplace

Visual Hierarchy / First Impression

10% of score

07One dominant focal point — Above the fold, the eye is told where to start — by an image, a composed hero, or a deliberate type contrast, never a wall of text.Based on Laws of UX: Von Restorff effect (opens in a new tab)

08Emphasis follows a scale — Headings, body, and captions sit on distinct sizes, and size ranks importance.Based on NN/g: Visual hierarchy (opens in a new tab)

09The page looks made — A colour system, styled controls, and differentiated sections — not browser defaults on white.Based on Laws of UX: Aesthetic-usability effect (opens in a new tab)

Hierarchy tells the eye where to go first. Without it a page is a list, and the most important thing on it is whatever happens to be biggest. Typical failure: Three competing elements at the same visual weight above the fold.

Applies to Marketing site · E-commerce · Marketplace

Lead Generation Friction

10% of score

10The CTA keeps its promise — Not an application dressed as a demo.Based on NN/g: Information scent (opens in a new tab)

11What happens next is stated — Response time and process after submitting.Based on Nielsen #1: Visibility of system status (opens in a new tab)

12A path for the not-ready — Pricing, docs, or examples before a sales call.Based on NN/g: The reciprocity principle (opens in a new tab)

A call to action is a promise about what happens next. Breaking it costs more than a weak CTA ever would. Typical failure: A CTA that leads to an application rather than the promised action.

Applies to Marketing site

Content & Copy Quality

10% of score

13Written outside-in — The reader’s problem before the product’s vocabulary.Based on Laws of UX: Mental model (opens in a new tab)

14Scannable structure — Headings and lists carry the argument on a skim.Based on NN/g: How users read on the web (opens in a new tab)

Copy written from the inside describes the product. Copy written from the outside describes the reader's problem. Typical failure: Feature names invented by the company, unexplained.

Applies to Marketing site

Trust & Credibility

10% of score

15Social proof is attributed — Testimonials carry a name, role, or company.Based on Stanford Web Credibility #4: Trustworthy people behind the site (opens in a new tab)

16Contact details are findable — A real way to reach a real organization.Based on Stanford Web Credibility #5: Easy to contact (opens in a new tab)

17Claims come with specifics — Numbers and evidence where adjectives would be cheaper.Based on Stanford Web Credibility #1: Easy to verify (opens in a new tab)

Trust is the tax every unknown brand pays. Specific proof lowers it; generic reassurance does not. Typical failure: Testimonials with no name, company, or face attached.

Applies to Marketing site · E-commerce · Marketplace

Accessibility

10% of score

18Body text meets 4.5:1 contrast — Measured, not eyeballed.Based on WCAG 1.4.3: Contrast (Minimum) (opens in a new tab)

19Every control has an accessible name — Icon-only buttons included.Based on WCAG 4.1.2: Name, Role, Value (opens in a new tab)

20Images carry meaningful alt text — Describing content, not repeating filenames.Based on WCAG 1.1.1: Non-text Content (opens in a new tab)

Accessibility failures are the only findings that are also legal exposure, and they usually indicate the same carelessness elsewhere. Typical failure: Body copy under 4.5:1 contrast, or an icon button with no label.

Applies to Marketing site · E-commerce · Marketplace · Product

Mobile Experience

10% of score

21Tap targets at least 44px — Fingers, not cursors. Measured: clear misses and small targets crowded together cost points.Based on WCAG 2.5.5: Target Size (opens in a new tab)

22No horizontal scroll on phones — The layout fits the viewport it was opened in.Based on WCAG 1.4.10: Reflow (opens in a new tab)

23Content parity — Mobile keeps the page’s core content and actions.Based on Google Search: Mobile-first indexing (opens in a new tab)

Most first impressions happen on a phone. A desktop-first layout that merely survives on mobile is not the same as one designed for it. Typical failure: Controls well under 44px, small targets crowded together, or a layout that scrolls sideways.

Applies to Marketing site · E-commerce · Marketplace · Product

Performance

5% of score

26Layout holds still while loading — Nothing shifts as assets arrive. Measured on a real load.Based on Core Web Vitals: Cumulative Layout Shift (opens in a new tab)

27Hero media arrives with the headline — Not seconds after it.Based on Core Web Vitals: Largest Contentful Paint (opens in a new tab)

Speed is a feature people feel before they can name it. Layout that shifts while loading reads as broken even when nothing is. Typical failure: A hero image that arrives after the headline has already rendered.

Applies to Marketing site · E-commerce · Marketplace · Product

Checkout Friction

10% of score

28Full cost early — Shipping and taxes before the payment step.Based on Baymard checkout research: Extra costs (opens in a new tab)

29Guest checkout exists — An account is an invitation, not a toll.Based on Baymard checkout research: Forced account creation (opens in a new tab)

30Reassurance at payment — Security cues and accepted methods where the card number goes.Based on Baymard checkout research: Payment trust (opens in a new tab)

Surprise at checkout is the single largest cause of abandoned carts. Cost known early converts better than cost revealed late. Typical failure: Shipping cost first disclosed at the payment step.

Applies to E-commerce

First Purchase Experience

10% of score

31No account wall before purchase — The first sale happens before the relationship is assumed.Based on NN/g: Login walls (opens in a new tab)

32Returns findable before buying — Guarantees where the decision is made.Based on Baymard checkout research: Returns policy (opens in a new tab)

A first purchase is the only one made without trust already established, so every extra requirement is charged against a balance that is still empty. Typical failure: An account wall between the cart and the purchase.

Applies to E-commerce

Form & Error UX

5% of score

33Errors name the fix — Next to the field that needs it.Based on WCAG 3.3.3: Error Suggestion (opens in a new tab)

34Validation waits for intent — No red text on the first keystroke.Based on Baymard: Inline form validation (opens in a new tab)

Forms are where intent turns into revenue. Every avoidable moment of friction there is measured directly in abandonment. Typical failure: Validation that fires on every keystroke, or errors with no recovery path.

Applies to E-commerce · Product

Clarity & Microcopy

12% of score

42Controls name their action — Buttons say what happens — never bare “Submit” or the system’s internal term.Based on Nielsen #2: Match between the system and the real world (opens in a new tab)

43Terminology is consistent — One name per concept, everywhere it appears.Based on Nielsen #4: Consistency and standards (opens in a new tab)

44Microcopy answers the next question — Helper text sits where the doubt happens, not in a docs link.Based on Nielsen #10: Help and documentation (opens in a new tab)

Inside a product, words are the interface. Every label a user has to decode is a small tax paid on every single use. Typical failure: A primary button that just says “Submit”, or a feature named in internal jargon.

Applies to Product

Scanability & Hierarchy

10% of score

45The key number reads first — On any dashboard, the eye lands on what matters most.Based on NN/g: Preattentive processing in dashboards (opens in a new tab)

46Tables are built for scanning — Aligned numerals, ranked columns, no decoration where data should be.Based on NN/g: Data tables (opens in a new tab)

47Density has a rhythm — Working screens pack information without turning into noise.Based on Laws of UX: Chunking (opens in a new tab)

Working screens are read hundreds of times. Hierarchy decides whether each read takes a glance or a search. Typical failure: A dashboard where every metric renders at the same visual weight.

Applies to Product

Task & Workflow Efficiency

13% of score

48Core jobs take few steps — The everyday task is measured in clicks, and it’s a small number.Based on Nielsen #7: Flexibility and efficiency of use (opens in a new tab)

49Defaults do the remembering — Known context is pre-filled, never asked twice.Based on WCAG 3.3.7: Redundant Entry (opens in a new tab)

50Frequent actions sit shallow — Daily actions are never buried three menus deep.Based on Laws of UX: Pareto principle (opens in a new tab)

A signed-in user has already converted — the product now competes on how fast it gets them through their job. Friction here compounds daily. Typical failure: Re-entering context the product already knows, on every run.

Applies to Product

System Feedback & Recovery

10% of score

54Every action answers back — Saves, loads, and long waits announce themselves.Based on Nielsen #1: Visibility of system status (opens in a new tab)

55Destructive actions are guarded — Delete confirms, or undo exists — ideally both.Based on Nielsen #5: Error prevention (opens in a new tab)

56Errors keep work safe — A failure explains itself and never eats what the user typed.Based on Nielsen #9: Recognize, diagnose, and recover from errors (opens in a new tab)

In-app trust isn’t testimonials — it’s whether the interface answers back. Silence after a click reads as breakage even when nothing broke. Typical failure: A destructive action with no confirmation and no undo.

Applies to Product