What Should a Professional Website Audit Include?

By Dave TanguayPublished Editorial draft

Quick Answer

A useful website audit explains which important tasks work, where visitors or administrators encounter problems, and what to fix first. Ask for a defined scope, representative pages and journeys, reproducible evidence, prioritized recommendations and a retest plan. A list of automated warnings alone cannot tell you whether your website serves your business.

Before commissioning an audit, name the decision it should support. You might need to investigate missing inquiries, prepare a redesign, assess an inherited WordPress site or find why customers abandon a purchasing step. Those questions call for different access, expertise and tests.

Define the scope before testing

Agree on the website, subdomains, environments and third-party systems included. List the page types and visitor tasks to sample. A brochure site might need its main services, contact form and mobile navigation reviewed. A store also needs representative products, search, variants, cart behavior and an approved checkout test.

Record the browsers, devices and account states being tested. A logged-out visitor, returning customer and administrator can encounter different behavior. A sample-based review should identify its coverage and exclusions clearly.

Access also sets limits. Public pages reveal some problems; confirming form delivery, redirects, permissions, backup arrangements or integration failures may require authorized administrative access. Agree on whether the audit includes recommendations, implementation or both. Our website audit service is a starting point for discussing that scope.

What should the audit examine?

Visitor tasks and content

Can someone understand the offer, find the relevant service or product, and complete the next step? Review navigation labels, page hierarchy, mobile controls, form instructions and completion messages. Check content against the actual business offer: outdated service details and missing delivery information can matter more than a cosmetic inconsistency.

Search access and page information

Review representative response codes, indexability settings, canonical URLs, internal links, page titles and content organization. Ask the reviewer to distinguish an observed technical issue from a proposed editorial improvement. Google’s SEO Starter Guide provides a useful baseline; an audit should apply it to your site rather than promise a particular ranking.

Performance and functionality

Investigate important page types and interactions. State whether findings come from real visitor measurements or a controlled test, and record relevant test conditions. A slow product selector and an oversized homepage image need different investigations. Our Core Web Vitals guide explains that measurement distinction.

Accessibility

Combine appropriate automated checks with manual evaluation of representative tasks. W3C explains that no tool alone can determine whether a site meets accessibility standards. The report should state its evaluation method, pages and interactions covered, and known limitations. A quick scan is a different deliverable from a scoped conformance evaluation.

Administration, security and privacy implementation

Where included and authorized, review software maintenance, account ownership, permissions, recovery arrangements, data flows and consent implementation. Ask which checks were observational and which were actually tested. Seeing a backup setting does not establish that a restore works. A website audit also needs a separate, explicit scope before anyone performs intrusive security testing or evaluates legal obligations.

What does a useful finding look like?

Consider this illustrative finding, not a result from a client audit:

  • Task: Submit an inquiry using the mobile contact form.
  • Observation to investigate: The page displays a success message, but the designated recipient reports no message.
  • Evidence to collect: Test time, page URL, device, submission reference and authorized delivery-log review. Keep personal data out of shared screenshots.
  • Impact: A prospective customer may believe the business received an inquiry that never reached its destination.
  • Next action: Trace the submission through form storage, sending service and receiving mailbox before assigning a cause.
  • Acceptance check: Repeat an approved test and verify the expected record and delivery, with a named owner confirming receipt.

This record gives a developer something to investigate and an owner something to verify. “Improve the contact form” does neither.

Turn the report into an ordered work plan

Prioritize blocked tasks, exposure and business impact alongside the strength of the evidence. Separate a confirmed failure from a suspected cause. Note when one repair unlocks several others, such as correcting a shared navigation component.

Each accepted recommendation needs an owner, an implementation boundary and a way to check completion. If a rebuild is proposed, ask which findings require replacement and which can be repaired. Findings can then feed a defined development work plan instead of an open-ended redesign.

Key Takeaways

  • Start with the business decision and visitor tasks the audit must explain.
  • Require explicit coverage, access requirements and exclusions.
  • Useful findings connect evidence, impact, next action and acceptance checks.
  • Budget for resolving and retesting the important findings.

Frequently Asked Questions

Can a free website scanner replace an audit?

It can help identify possible issues within its coverage. It cannot establish every task outcome, content problem or accessibility barrier. Use its results as inputs for review rather than a complete diagnosis.

Does an audit include fixing the website?

Only when that work is in the agreed scope. Ask whether the deliverable includes investigation, recommendations, implementation, retesting and any follow-up review. Keep those responsibilities explicit.

Should every page be tested?

The appropriate coverage depends on the site and decision. A representative sample can reveal shared-template problems, but unique workflows, unusual content and high-impact tasks may need separate testing. The report should name what was covered.

Archives