What Is a Website Audit? A Plain-English Guide

The definition, without the jargon

An audit compares what your website actually does against what it should do, then documents the gap. The standard you compare against depends on the audit type: a technical audit checks against engineering best practice, an SEO audit checks against what search engines reward, a conversion audit checks against what a real customer needs in order to buy.

Two things separate an audit from casual clicking around. First, it is systematic: a defined scope, so a clean result actually means something. Second, it produces evidence. A finding like "your checkout is confusing" is an opinion. A finding like "the shipping cost only appears on step three, after the visitor has entered payment details" is an audit result — you can verify it, prioritize it, and fix it.

One thing an audit is not: an analytics review. Analytics tells you what happened (traffic dropped, carts got abandoned). An audit tells you why, by inspecting the site itself. You usually need both, but they answer different questions.

The five main types of website audit

Most audits fall into one of five categories, and most paid audits blend two or three of them. Knowing the categories keeps you from buying the wrong one.

The boundaries blur in practice — slow pages hurt SEO and conversion alike. But when someone offers you "a website audit," ask which of these it actually covers; the answer changes what you should pay and what you get back.

• Conversion and UX audit: evaluates whether visitors can and want to do the thing your site exists for. Covers messaging clarity, trust signals, calls to action, form and checkout flows, and the mobile experience. This is the type most directly tied to revenue.
• Technical audit: checks the machinery — page speed, broken links, redirect chains, SSL, crawl errors, rendering problems. Invisible when healthy, expensive when not.
• SEO audit: examines how search engines see your site — indexability, metadata, site structure, internal linking, and whether your content matches what people actually search for.
• Accessibility audit: measures conformance with standards like WCAG so people using screen readers, keyboards, or assistive tech can operate the site. Also a legal-exposure question in many jurisdictions.
• Content audit: an inventory of every page with a verdict — keep, update, merge, or delete. Most useful for older sites that have accumulated hundreds of pages nobody remembers publishing.

What a good audit report contains

The output matters more than the process. A useful report, whatever its type, has a recognizable shape — and once you know it, you can spot a weak audit in a minute.

The red flags are just as consistent: raw tool output pasted into a PDF, findings with no evidence, forty items all marked critical, and advice generic enough to apply to any website. "Improve your value proposition" is not a finding. "Your headline says the company name and nothing else — here is a rewrite that says what you sell" is.

To calibrate your expectations, there is a real audit report published at /sample-report — its structure (scores, evidence, rewrites, prioritized plan) is the shape to look for whatever tool or firm you use.

• Scores or benchmarks, scoped per area, so you know where you stand and can measure change after fixes
• Evidence for every finding: a screenshot, a URL, or a reproduction step — never a bare assertion
• Prioritization by impact, so you know which three fixes matter this week and which twenty can wait
• Specific fixes, not directions: rewritten copy you can paste in, not "strengthen your messaging"
• A format you can hand to a developer, writer, or agency without translating it first

DIY, automated tools, or an agency?

Doing it yourself with a checklist is free and better than nothing, but it has a structural flaw: you know too much. You already know what your company does and what the jargon means, so you cannot experience the site as a first-time visitor. DIY audits reliably catch broken links and typos and reliably miss the confusing headline that costs the most money.

Automated grader tools — the category built on Lighthouse-style metrics — are fast, cheap or free, and genuinely good at what they measure: load times, missing tags, technical hygiene. But they only measure what a machine can score without understanding the page. A site can post excellent technical numbers while its homepage fails to say what the business sells. They are a fine first pass, not an evaluation of whether the site persuades anyone.

Agencies and consultants sit at the other end: deep, contextual, tailored to your market — and priced in the thousands, delivered in weeks. For a large site with real budget, that depth is worth it. For a small business that mainly needs to know why visitors are not converting, it is often more firepower than the problem requires.

A newer middle option is AI-driven audits that behave like a customer rather than a scanner. ExperienceGrade works this way: an AI customer opens your site in a real browser on desktop and mobile, reads up to five key pages, and attempts your primary customer journey — actually starting a checkout or a quote request — then scores six areas from clarity to technical health, for $29.99 one-time. It answers the question graders cannot: what happens when a stranger tries to buy.

The test most audits skip: the customer journey

Here is the gap in most audit reports, regardless of price: they evaluate pages in isolation, and nobody ever tries to complete the journey. The homepage gets scored, the product page gets scored, and the path between them — the part where money changes hands — goes untested.

Journeys fail in ways page-level checks structurally cannot see. A form that errors only after submission. A mobile menu that covers the buy button. A checkout that demands account creation with no explanation. Each page can look fine on its own while the sequence quietly fails.

Whatever audit you choose, apply one test: does it answer the question "can a motivated stranger figure out what this site sells and complete a purchase or inquiry, start to finish, on a phone?" If the report cannot answer that, it audited your pages but not your website. Ask for journey testing explicitly, or run the walkthrough yourself — logged out, on a phone, following only what the screen tells you.

How to choose the right audit for your situation

Match the audit to the symptom, not to whatever a vendor happens to sell. A few common situations and the honest answer for each:

Whatever you pick, insist on the report shape described above — evidence, priorities, specific fixes. A cheap audit with evidence beats an expensive one without it.

• Traffic is fine but nobody converts: you need a conversion and UX audit with real journey testing. Technical scans will tell you the site is fast while it fails to sell.
• Traffic itself has dropped: start with an SEO audit plus a technical crawl — indexing problems and slow or broken pages are the usual suspects.
• You just launched or redesigned: run a broad audit across all areas before spending on traffic. Sending paid visitors to an untested site is the most expensive way to find bugs.
• You sell to government, enterprise, or the public sector: an accessibility audit is not optional; conformance gaps show up in procurement reviews and legal complaints.
• The site is old and sprawling: begin with a content audit to shrink the surface area, then audit what survives.

How often should you audit, and what happens after?

An audit is a snapshot, not a subscription requirement. For most small and mid-sized sites: a full audit after any launch or redesign, a check after major changes to pricing or checkout, and otherwise once or twice a year. Auditing monthly without shipping fixes in between is just paying to be told the same things.

The step people skip is the retest. Fixes fail, regress, or get half-implemented, and you will not know unless something verifies them. Build verification into the plan from the start. ExperienceGrade's $49.99 option includes three retests over 90 days for exactly this reason — the fix-verify loop is where audits pay off or quietly evaporate.

The pattern that actually moves numbers is boring: audit, fix the top three items, verify, then fix the next three. Teams that do that outperform teams that commission an impressive report and file it.