A UX Audit Before a Redesign:
What to Look For
Before you redesign, audit the experience you already have: the tasks people come to do, where they struggle, who is blocked and which fixes matter most.
A UX audit before a redesign shows what is wrong with the current experience, where it goes wrong and how much each problem matters. Without one, a redesign can replace the parts that work and carry the real problems into the new version. With one, the redesign starts from a ranked list of findings, each with its evidence.
The short answer: a UX audit checks the tasks people come to do against four kinds of evidence: an expert review against usability heuristics, analytics that show where people drop off, basic accessibility checks against WCAG and, where useful, a short test with real users. Each finding gets a severity and an effort estimate, so the redesign starts with the problems that matter most.
Audit or redesign: which problem do you have?
A redesign is one possible answer. The audit finds the question. The usual reasons for considering a redesign, such as falling enquiries, complaints about finding things, a dated look or a new offer, often call for different responses.
| What it may point to | What the audit checks |
|---|---|
| Signal: People leave at one step of a form or checkout | |
| A local problem in that step | Labels, error messages, required fields, step order |
| Signal: Visitors cannot find pages or products that exist | |
| Navigation and labels that do not match the visitor's words | Menu structure, search, labels, page titles |
| Signal: Traffic holds steady while enquiries fall | |
| An unclear offer or a weak path to contact | Calls to action, page content, the enquiry form |
| Signal: The product looks dated but works | |
| Visual design rather than usability | Whether a visual refresh can keep the structure that works |
| Signal: Problems appear across most pages and tasks | |
| Structure, content model or platform | Whether fixing pages one by one is still realistic |
People leave at one step of a form or checkout
What it may point to
A local problem in that step
What the audit checks
Labels, error messages, required fields, step order
Visitors cannot find pages or products that exist
What it may point to
Navigation and labels that do not match the visitor's words
What the audit checks
Menu structure, search, labels, page titles
Traffic holds steady while enquiries fall
What it may point to
An unclear offer or a weak path to contact
What the audit checks
Calls to action, page content, the enquiry form
The product looks dated but works
What it may point to
Visual design rather than usability
What the audit checks
Whether a visual refresh can keep the structure that works
Problems appear across most pages and tasks
What it may point to
Structure, content model or platform
What the audit checks
Whether fixing pages one by one is still realistic
Define the tasks to audit
Start with what people come to do. Nielsen Norman Group describes task analysis as learning how users work, and which tasks they perform, to reach their goals. For an audit, that means a short list of the tasks that matter most to your users and your business: find a product and buy it, request a quote, book a call, invite a colleague, export a report.
For each task, write down who performs it, where it usually starts, what counts as success and which data already exists about it. The task list keeps the audit focused. Every finding should point to a task it affects.
How to do a UX audit: the expert review
The core of the review is a heuristic evaluation: experienced reviewers walk through each task and compare the interface with established usability principles. A common reference is Jakob Nielsen's 10 usability heuristics, which he describes as broad rules of thumb rather than specific guidelines. These questions form the core of a UX audit checklist for every task:
- Visibility of system status: After a click, a save or a payment, does the person know what happened?
- Match with the real world: Do labels use the customer's words or internal jargon?
- User control and freedom: Can people go back, undo or leave a flow without losing their work?
- Consistency and standards: Do the same actions look and behave the same way on every page?
- Error prevention and recovery: Does the form prevent mistakes and explain the ones it catches in plain language?
- Recognition rather than recall: Are options and earlier choices visible, or must people remember them?
NN/g's guide to running a heuristic evaluation recommends three to five evaluators who review independently before comparing notes, so that one person's view does not shape the others. It also warns that heuristic evaluation cannot replace research with users. The review finds likely problems quickly; the other evidence checks them and finds what the review misses.
Five inputs, one ranked list.
- Key tasks What must work for users and the business?
- Heuristic review Where does the interface break the heuristics?
- Analytics Where do people drop off?
- Accessibility checks Who is blocked, and by which WCAG criteria?
- Test with five users Why do people struggle?
Findings rated by severity and effort
Each one names the task it affects
Read the analytics for where people drop off
Analytics show where people struggle, though rarely why. Look for these signals on each key task:
- Drop-off between steps: a funnel exploration in Google Analytics shows how many people complete each step of a task and where they leave.
- Search behaviour: searches that return nothing, or are followed by an exit, show what people expect to find.
- Form behaviour: fields that trigger errors or where people abandon the form.
- Devices: a step that works on desktop and fails on phones.
- Speed and stability: Google's Core Web Vitals set good thresholds of 2.5 seconds for Largest Contentful Paint, 200 milliseconds for Interaction to Next Paint and 0.1 for Cumulative Layout Shift, assessed at the 75th percentile separately for mobile and desktop.
- Support questions: recurring questions to support or sales often point to missing or unclear content.
Practical note: Check the tracking before you trust the numbers. A missing event or a step counted twice can make a healthy page look broken, or hide a broken one.
Check accessibility basics against WCAG
WCAG 2.2, a W3C Recommendation, sets the success criteria for accessible web content. For an audit, start with W3C's Easy Checks, which include page titles, text alternatives for images, headings, colour contrast, visible keyboard focus, zoom, form labels and captions. Three criteria from WCAG that audits often test:
- 1.4.3 Contrast (Minimum): normal text needs a contrast ratio of at least 4.5:1 against its background.
- 2.4.7 Focus Visible: people using a keyboard can see which element has focus.
- 3.3.2 Labels or Instructions: fields that need input have labels or instructions.
W3C is clear that these quick checks are not a complete evaluation: a page can pass them and still have significant barriers. Many accessibility findings are also usability findings. A form field without a label fails a WCAG criterion and makes the form harder for everyone.
Watch five people try the key tasks
The expert review and the analytics tell you where to look. A short test with real users shows why people struggle. Jakob Nielsen's analysis of testing with five users argues that a handful of participants reveals most usability problems, and that several small rounds, with fixes in between, improve a design more than one large study. If your site serves clearly different groups, such as buyers and suppliers, include people from each group: Nielsen suggests three or four per group when there are two.
Give each participant the key tasks, ask them to think aloud and avoid helping. Note where they hesitate, what they expected to happen and what they say while doing it.
Prioritize findings by impact and effort
Rate every finding the same way. Nielsen's severity ratings combine three factors: how often the problem occurs, how hard it is for people to overcome and whether it bothers them once or repeatedly. His scale runs from 0, not a usability problem, to 4, a usability catastrophe that must be fixed before release.
Then estimate the effort to fix each finding: a content change, a design change, development work or a change to the platform itself. Plot both on a prioritization matrix, the two-criteria chart NN/g describes for basing decisions on objective criteria instead of opinion.
Ranking findings: impact against effort
Impact
Severity and how many key tasks it affects
High impact · Low effort:
Fix first
Severe problems in key tasks, quick to change.
For example: a form field with no label
High impact · High effort:
Plan into the redesign
Structural problems that repeat across the key tasks.
For example: navigation by internal department
Low impact · Low effort:
Batch with other fixes
Small issues, cheap to fix, done together in one pass.
For example: inconsistent button wording
Low impact · High effort:
Park and re-check
Revisit when new evidence raises the impact.
For example: rebuilding a rarely used page
Effort
Content edit → Design change → Development → Platform change
Ranked this way, the redesign brief follows the evidence, and every item on it names the task it affects.
When the findings point to a redesign
Fix in place when the problems are local: one step, one form, unclear labels, low contrast. Consider a redesign when severe findings repeat across most key tasks and share a cause in the structure, such as the navigation, the content model, the page templates or a platform that makes every fix expensive.
If you redesign, the findings become your website redesign checklist, and they also tell the redesign what to keep. Tasks that already work well should come through the website redesign unchanged, or improved on purpose. For software products, the same findings become the starting point for product design work on the affected flows.
What a UX audit report should include
- Scope: The tasks, user groups, devices and pages reviewed.
- Method: Who evaluated, which heuristics, which analytics period and how many test participants.
- Findings: Each with the task, the location, the evidence, the heuristic or WCAG criterion, a severity and an effort estimate.
- What works: Tasks and patterns the redesign should keep.
- Priorities: The matrix and the first fixes.
- Recommendation: Fix in place, redesign some parts or redesign the whole, with the reasons.
- Measures: The baseline numbers to compare after the changes.
Invite your team
Good work starts with the right people.
Something went wrong.1
Email address
alex@gmail.com
Role
Member
CancelSend invitation
The task stops. The screen gives no way forward.
The finding
An error message.
No useful answer.
- Visible in the interface The message does not explain why the invite failed or how to correct it.
- A hypothesis to test People may retry the same address or ask support. The screen alone cannot confirm this.
- How to check Review invite errors and support tickets. Observe people completing this task.
A UX audit report with all seven parts can serve as the brief for the work that follows.
Sources
- Nielsen Norman Group — 10 Usability Heuristics for User Interface Design
- Nielsen Norman Group — How to Conduct a Heuristic Evaluation
- Nielsen Norman Group — Task Analysis: Support Users in Achieving Their Goals
- Nielsen Norman Group — Why You Only Need to Test with 5 Users
- Nielsen Norman Group — Severity Ratings for Usability Problems
- Nielsen Norman Group — Using Prioritization Matrices to Inform UX Decisions
- Google Analytics Help — Funnel exploration
- web.dev — Web Vitals
- W3C — Web Content Accessibility Guidelines (WCAG) 2.2
- W3C WAI — Easy Checks: A First Review of Web Accessibility
Frequently asked questions
Yes. Use several evaluators who review independently, work task by task with the heuristics and add a short test with real users. Include at least one reviewer who did not build the product, because the people who built it already know where everything is.
No. A UX audit includes basic accessibility checks. A full accessibility evaluation tests the site against all applicable WCAG success criteria, and W3C notes that its Easy Checks are only a first review.
An audit finds and ranks problems. Conversion rate optimization tests changes on the live site and measures whether they improve the result. The audit's findings often become the first changes worth testing.
Start your redesign with evidence
Tell us which site or product you plan to redesign and what prompted the decision. We can help you choose the tasks to audit, gather the evidence and decide what to fix first.
