Website UX Audit Checklist: 30 Checks, and How to Rank What You Find

A UX audit is only useful if someone acts on it. Many audits end as a 60-row spreadsheet of opinions, and the team fixes the easy items instead of the ones that cost the most visitors.

This checklist is built to avoid that. It gives you 30 checks you can run in about a day, grouped by the part of the experience they test, plus a simple way to rank what you find. It is written for marketing sites, landing pages and small stores. If you run a SaaS site, our SaaS conversion audit walks the signup path in more depth.

Before you start: pick the journeys and the evidence

An audit of "the whole website" turns into an audit of nothing. Pick one to three journeys that matter to the business, for example:

• Ad or search visitor lands on a page and requests a quote
• Visitor reads the homepage, checks pricing, and starts a trial
• Shopper lands on a product page and checks out

Then gather what you already have for the last 28 days: visitors to the entry pages of each journey, how many reach each step, and how many finish. Split everything by phone and desktop. Add support tickets and sales-call notes if you have them, since they often name the exact place people get confused.

Mark every check below as pass, issue or not tested. Baymard Institute recommends the same three states in its own audit framework, because "not tested" stops gaps in your review from being mistaken for passes (Baymard, UX audit checklist).

1. Clarity: can a stranger tell what this is?

1. The headline says what you offer and for whom. Read only the headline and subheading. Could a stranger name your category and your customer? A slogan is not a description.
2. The main action is on the first screen of a phone. In a 2018 eyetracking study, Nielsen Norman Group found people spent about 57% of their page-viewing time above the fold and 74% in the first two screens (NN/g, Scrolling and Attention).
3. There is one primary action per page. Other actions are visibly secondary, not five equal buttons.
4. Button labels say what happens. "Get my quote" beats "Submit." "See plans" beats "Learn more."
5. The page matches the link that brought people there. If an ad promised a price or an offer, the page shows it near the top.

2. Navigation: can people find the next step?

6. Menu labels use your customers' words, not internal team names for products or departments.
7. The pages people look for most are one click from the menu. Check your analytics for the most visited pages and confirm each one is in the main navigation.
8. Visitors can tell where they are. The current section is marked, and deep pages have a way back.
9. No dead ends. Every page in the journey has a clear next step, including thank-you and error pages.
10. Search works for the terms people use, if you have search. Try misspellings and plural forms.

3. Forms: is the main conversion easy to finish?

11. The form asks only for what you need now. Nielsen Norman Group recommends cutting fields that can be collected later or derived another way (NN/g, Website Forms Usability).
12. Every field has a visible label, not just placeholder text that disappears when someone types.
13. Errors appear next to the field, in plain words, and say how to fix the problem.
14. The form keeps what people typed after an error or when they go back a step.
15. Fields bring up the right phone keyboard (email, number, phone) and support autofill.
16. People know what happens after they submit. The confirmation says what comes next and when. Our form abandonment guide shows how to find the field where people quit. For lead forms, our guide to which fields to keep or cut goes field by field.

4. Trust: is there a reason to believe you?

17. Proof sits next to the decision it supports. A testimonial or customer logo near the button does more than a block of them at the bottom.
18. Price, or the cost of the next step, is stated before people commit. If pricing is custom, say how it works.
19. Contact details and policies are easy to find. Returns, refunds, privacy and a real way to reach you.
20. Nothing looks broken. Missing images, placeholder text, an old copyright year or a dead link can undo good copy.

5. Accessibility: can everyone use it?

21. Text contrast is at least 4.5:1 against its background for normal text, and 3:1 for large text (W3C, WCAG 2.2 Contrast (Minimum)).
22. Tap targets are big enough. WCAG 2.2 sets a minimum of 24 by 24 CSS pixels for most targets, or enough spacing around smaller ones (W3C, Target Size (Minimum)).
23. Everything works with a keyboard, and the focus outline is visible.
24. Images that carry meaning have alt text, and color is never the only way information is shown, such as a red border alone marking an error.
25. Pages work at phone width without sideways scrolling.

6. Speed and stability: does it load and stay put?

26. Largest Contentful Paint is 2.5 seconds or less on mobile, measured at the 75th percentile of page loads (web.dev, Web Vitals).
27. Interaction to Next Paint is 200 milliseconds or less, so taps and clicks respond quickly.
28. Cumulative Layout Shift is 0.1 or less, so buttons don't jump as the page loads.
29. Pop-ups don't block the first screen before people have seen what the page is about.
30. Third-party scripts are justified. Chat widgets, tag managers and old tracking pixels each add weight. Run your key pages through PageSpeed Insights and look at the mobile results first.

How to rank what you find

You will end with more issues than you can fix this month. Rank them before you change anything:

Start with high reach and strong evidence, even if the fix is dull. Then make one change at a time and compare equal periods before and after. If you don't have the traffic for A/B tests, our guide to CRO without A/B testing explains how to measure honestly.

Two habits make the ranking more reliable. First, get more than one reviewer: Nielsen Norman Group recommends that three to five people evaluate the same interface independently, because each individual misses some problems (NN/g, How to Conduct a Heuristic Evaluation). Second, watch a few real people try the journey. NN/g's long-standing advice is that testing with about 5 users finds most of the problems a qualitative study will find (NN/g, How Many Test Users).

The checklist (copy this)

• One to three key journeys picked, with 28 days of step counts split by phone and desktop
• Headline says what you offer and for whom
• Main action on the first phone screen, labeled with what happens
• One primary action per page
• Page matches the ad or link that sends traffic to it
• Menu labels in customer words; top pages one click away
• No dead ends, including thank-you and error pages
• Form trimmed to what you need now, with visible labels
• Inline errors in plain words; typed data kept
• Proof next to the decision; price or next-step cost stated
• Policies and contact details easy to find; nothing looks broken
• Contrast at least 4.5:1; targets at least 24 by 24 CSS pixels
• Keyboard and phone-width checks done
• Mobile Core Web Vitals checked on key pages
• Issues ranked by reach, severity, evidence and effort

Where Cortality fits

A checklist tells you where to look. It can't tell you which issue is costing you the most visitors on your own site.

For a quick starting point, our free page leak scan shows where visitors are likely to look on your page and what fixing it could be worth. It takes about a minute, no install needed. The scan is a prediction from our attention model, not data from your real visitors, so treat it as a place to start. You see a first-screen attention prediction right away, and the full findings arrive by email a few minutes later.

If you audit pages for clients, see Cortality for agencies.

When you want answers from your own traffic, Cortality's paid product uses a small tag on your site to measure how real visitors use every part of your page, then gives you a ranked list of fixes with the exact change for each. After you make a fix, it re-measures to show whether it worked. You can see what a finding looks like, read how the measurement works, or compare it with published CRO audit prices.

Start your free week and you see your first report before you pay. You save a card at signup, nothing is charged until 3 days after your first report, and you can cancel before then and pay nothing. Any charge can be refunded within 14 days (refund policy). Plans start at $99 a month, and the visitor limits are on the pricing page.