Record the review rather than score the site
A useful website audit checklist helps you record what you find and decide what to investigate next. Review the pages people need, the clarity of your offer, the proof behind it and the path to an enquiry alongside search and technical checks. Mark missing information as unknown rather than turning every unchecked item into a failure.
Start by choosing a decision and a manageable set of important pages. For a service business, that might be the home page, main service description, a relevant guide and the contact journey. Add other pages when they could change the decision; do not assume the sample describes the whole site.
The 40 checks below are a method for your own review. They do not imply that every item is included in every Alytixx order. Record the page, what you inspected and the next action, then use the worksheet to organize follow-up. A completed check records work done; it does not declare the website good.
A. Important pages and navigation
Follow the routes a prospective customer would use rather than opening every page from a saved list. Record both the useful paths and the places where a necessary answer is hard to find.
Check 1
Can visitors find your main services from the site’s main sections?
Start from the home page and main navigation and try to reach each priority service. Record the route and any missing or misleading step. If the service is intentionally secondary, ask the owner how its intended audience should find it before recommending a new menu item.
Check 2
Does each important page have a clear purpose?
Read the main heading, introduction and next action. Write the buyer question or task the page serves in one sentence. If you cannot identify a purpose, ask who needs the page and what they should understand or do; keep that question open until the role is agreed.
Check 3
Do links describe what visitors will find next?
Compare the link wording with its destination. Record vague labels or links that suggest a different action from the page they open. If the context makes a short label clear, preserve it; if readers must guess, propose a description that reflects the actual destination.
Check 4
Are important pages missing from the routes visitors are likely to follow?
Trace a service selection or enquiry journey and note where a relevant page is absent. Save the starting page, missing answer and destination that could help. If visitor routes are unknown, describe the route you tested and ask for owner or user context instead of claiming all customers use it.
Check 5
Do related pages help the reader continue the same task?
Follow the links after a guide or service explanation. Record whether they lead to a relevant example, detailed answer or appropriate action. If a link changes the subject unexpectedly, decide what the reader needs next before adding more links; link quantity alone is not the goal.
B. Search readiness and observed search data
Keep public access, eligibility and actual visibility separate. Google’s technical requirements establish eligibility to be indexed, not guaranteed indexing. Use a property owner’s inspection or suitable search records when a conclusion needs evidence beyond the public page.
Check 6
Can the important pages be reached without signing in?
Open each intended public page in a browser without an authenticated session. Record whether it loads, redirects or asks for access. If a page is private by design, mark that purpose; if it should be public, ask the responsible person to investigate the restriction rather than assuming its cause.
Check 7
Do page titles and main headings describe the actual content?
Compare the browser title and main heading with the page’s subject and offer. Save any mismatch, outdated service name or duplicate wording that obscures the difference between pages. If the correct subject is unclear, confirm the page purpose before rewriting; a title change alone does not prove improved visibility.
Check 8
Are any important pages unintentionally restricted from search?
Ask the website owner or specialist to inspect search settings and instructions for the chosen pages. Record the restriction, address and whether it is intended. If the instruction or access state is unclear, request a specific technical check; do not remove deliberate exclusions as part of a general tidy-up.
Check 9
Have you distinguished pages that are indexable from pages confirmed as indexed?
Keep eligibility checks and actual indexing evidence in separate notes. Record the source and date of an owner’s exact-page inspection when available. If you can only confirm that the page opens, write indexing unconfirmed; a browser visit is not evidence that Google has included the page in its index.
Check 10
Are claims about search visibility supported by actual query observations or owner data?
Find the record behind a statement that a page ranks or has lost visibility. Save the query, context and date, or the owner report with its filters and period. If the source is absent, leave the statement unverified and ask which observation would settle the question.
C. Offer clarity
Compare the published description with confirmed business terms. Clear wording can explain a service; the business owner still needs to establish what that service actually includes.
Check 11
Is it clear who each service is for?
Read the service page for the intended customer, situation and relevant service area. Write whom the offer appears to address and quote any conflicting description in your own notes. If the target is broad or undecided, ask the owner for the intended fit before narrowing the copy.
Check 12
Does the page explain what the customer receives?
Look for a concrete deliverable, service or result of the work. Record the explanation or the place where only a broad promise appears. If the outcome varies, ask what remains consistent and what is agreed later; do not turn a desired customer result into a guaranteed deliverable.
Check 13
Can visitors understand what is included and excluded?
Compare the main description, terms and enquiry text. Record omissions or contradictions that affect the buying decision. If a likely expectation is not addressed, confirm the real scope with the owner and explain it near the offer rather than inventing an inclusion or exclusion.
Check 14
Is the pricing approach explained where it is relevant to the decision?
Check whether the page explains the actual fixed price, quotation process or factors used to establish scope where relevant. Record the visible explanation and unresolved questions. If terms are not confirmed, request the current facts; do not supply a typical market price or copy an obsolete amount.
Check 15
Are prerequisites, conditions and the next step clear?
Identify what a customer must provide, which conditions affect suitability and what happens when they act. Record the information available before the form or purchase step. If requirements emerge only after contact, ask which can be explained earlier without pretending every request follows the same process.
D. Proof and trust
Match proof to the claim it is meant to support. A genuine example with relevant context can be more useful than a collection of impressive-looking statements about a different service.
Check 16
Does each major claim have appropriate supporting evidence?
Pick the claims that could influence a purchase and locate their support. Record the source, what it establishes and any missing context. If a claim cannot be substantiated, ask for the relevant record or propose wording that matches the evidence; unsupported certainty should not become the selling point.
Check 17
Do examples relate to the service being described?
Check each example’s service, starting situation and deliverable against the current page. Record the relationship and any important difference. If an example concerns another offer, explain that context or replace it with relevant support rather than assuming its success transfers to the current service.
Check 18
Are testimonials genuine, attributable and relevant?
Ask the responsible owner to confirm the source and permission for published testimonials. Record what each review concerns and how it relates to this decision. If attribution or approval is missing, request verification; do not rewrite anonymous praise into a named customer endorsement.
Check 19
Is it clear who provides the service and how to contact them?
Look for an accurate provider identity and a working contact route. Record the visible information and where it can be found. If details conflict, confirm them with the business and consider whether a deliberate tracking or routing setup explains the difference before normalizing the page.
Check 20
Are fictional examples clearly distinguished from customer results?
Read the label and surrounding claims for every teaching scenario or sample report. Record whether a reader could mistake it for measured work. If the status is unclear, ask for its origin and label the example accurately; an invented walkthrough must not imply a real customer outcome.
E. Enquiry journey
Follow the whole action, including receipt. Use an agreed test so the people handling enquiries can recognize it and distinguish it from a real customer request.
Check 21
Does the page offer an appropriate next action?
Compare the page’s purpose with its main action. Record whether the reader can request the relevant service, inspect an example or find the missing detail. If readiness is uncertain, ask what decision comes next; immediately demanding a purchase may not suit an explanatory guide.
Check 22
Is the result of clicking the main call to action clear?
Read the button and nearby explanation, then open its destination. Record whether the result matches what was promised. If a general label opens a detailed application, make that commitment understandable before the click rather than judging the label only by how short it is.
Check 23
Are form questions understandable and relevant to the request?
Read each field label and identify why its answer is needed at this stage. Record ambiguous questions, unexplained requirements and missing guidance. If the purpose is unclear, ask the person handling enquiries before removing the field; some details may be necessary for qualification or routing.
Check 24
Can a permitted test enquiry be completed and received?
Agree a recognizable test and follow it through submission, confirmation and receipt. Record the address, time, device and recipient confirmation. If the request arrives nowhere or delivery access is unavailable, leave delivery unresolved and assign the owner check; a success message alone does not settle it.
Check 25
Does the confirmation explain what happens next without inventing a response time?
Read the confirmation after the agreed test. Record whether it tells the visitor what the request means and how the business continues. If a response deadline appears, confirm it is current and supportable; otherwise explain the next step without adding an unverified promise.
F. Usability and technical observations
Make a problem reproducible. Record the device, browser, page and steps, then describe what you expected and what happened. One successful test does not cover every user or condition.
Check 26
Can people read the content comfortably on a phone?
Open important pages on a phone and inspect headings, paragraphs, tables and overlays. Record clipped content, obstructed text or a need to scroll sideways during the main task. If the issue depends on the screen size, retain that context and ask for a check on the affected range.
Check 27
Are forms, menus and links usable in the intended journeys?
Complete the chosen route with a pointer and keyboard where relevant, including opening and closing navigation. Record the control and steps for any failure. If a specialist assessment is needed, ask for it; a quick journey test cannot certify accessibility for all users.
Check 28
Do errors and empty states explain what the user can do?
Trigger a permitted validation error or inspect a genuinely empty result. Record whether the message identifies the problem and offers a way forward. If you cannot safely reproduce the state, request an example from the responsible team instead of assuming the hidden behaviour is correct.
Check 29
Are important images or diagrams accompanied by meaningful explanations?
Read the text around each image that carries an important fact or instruction. Record what a reader would lose if the image were unavailable. If meaning depends on the visual alone, ask for an appropriate text explanation; decorative images do not need to repeat nearby prose.
Check 30
Are performance concerns based on an actual observation or measurement rather than a guess?
Record the page, device, connection context and observed loading or interaction problem. Keep any tool result with its test conditions. If you only have one run, describe that run and investigate its relevance; it cannot establish the experience of every visitor.
G. Competitor context and AI-related content
Compare relevant decisions and keep your own facts consistent. Include AI-answer observations only when they matter to the review, and describe exactly what was seen.
Check 31
Are the competitors and page types genuinely comparable?
Check the audience, service, market and page purpose for each comparison. Record why the pages belong in the same review and any differences. If the business models differ, compare a specific shared question rather than treating one website as an all-purpose benchmark.
Check 32
Does the comparison cover offer, proof and next step rather than design alone?
Use the same buyer question on each page: what is provided, what supports the claim and how to proceed. Record the visible answers and unknowns. If you only prefer a layout, call that a design preference; it is not evidence of better sales or service.
Check 33
Are important product facts consistent across your own pages?
Compare service names, deliverables, conditions and exclusions on the home, service and enquiry pages. Record conflicting statements with their locations. If versions differ, ask the owner for the current facts and update the relevant descriptions together rather than choosing the most persuasive version.
Check 34
Do your pages answer the questions a prospective buyer needs resolved?
Use real enquiry or sales questions where available and find their answers on the page. Record the unanswered question and why it matters to the decision. If demand is unknown, label the question as a candidate to test with buyers, not a proven popular search.
Check 35
Are mentions, citations and recommendations described separately when you review AI answers?
Save the question, AI service or Google feature, time and answer. Record separately whether the brand is named, a page is linked or the service is recommended. If the answer cannot be retrieved, mark the observation unavailable; do not turn an absent record into zero visibility.
H. Decisions and follow-up
Select actions from the recorded findings, not the number of boxes checked. A confirmed broken enquiry and an unanswered question need different next steps; colour or severity labels alone cannot decide their business importance.
Check 36
Is each proposed change connected to a recorded observation?
Trace the task to a page condition, owner fact or specific user finding. Record that connection with the proposed change. If the task is only a preference, identify it as such; if no basis can be found, investigate before treating the change as necessary.
Check 37
Have confirmed failures been separated from hypotheses?
Read the conclusion alongside its source. Record whether the failure was reproduced or the explanation is still a possibility. If the language moves from unclear scope to lost sales without supporting data, narrow the conclusion and define the check that could test the suspected effect.
Check 38
Is it clear which page or part of the journey needs attention?
Make the task specific enough for the responsible person to locate the issue. Record the address, section or shared element and any affected journey. If coverage is unknown, add an investigation of similar pages rather than describing one observed problem as a sitewide failure.
Check 39
Does each chosen action have a practical way to check completion?
Write the expected result and how someone will verify it after publication. Record an approved wording comparison, route check or permitted enquiry test as appropriate. If the check is improve user experience, make it specific enough to know whether the task has been completed.
Check 40
Will you distinguish a corrected page issue from a measured business improvement?
State what the immediate verification can show and what needs later business data. Record a repaired form separately from a change in qualified enquiries. If results are not measured, report the correction accurately without adding revenue, ranking or conversion claims.
Turn notes into a useful record
Keep the observation, interpretation and next task readable together. The rows below are fictional examples of a completed record. An investigation can be a valid next action when it resolves something necessary for the decision.
| Page | Observation | Why it may matter | Next action | What to verify | Status |
|---|---|---|---|---|---|
| Service page | The deliverable is not described. | A buyer cannot determine what the service supplies. | Confirm the scope and add the deliverable. | Published wording matches current terms; readers can explain it. | Needs follow-up |
| Contact form | A permitted test shows a repeated validation error. | The requested journey cannot be completed in the recorded test. | Reproduce the failure and repair it. | The agreed test completes and is received. | Problem confirmed; fix outstanding |
| Useful guide | Search performance records are unavailable. | Removal could discard a useful route, but its actual search use is unknown. | Ask the property owner for relevant records before deciding. | The decision cites available evidence or states the remaining uncertainty. | Owner check needed |
Choose a few next actions from the longer list
First address a confirmed failure that blocks an important task, then settle a question that could change a larger commitment. A broken submission path can deserve attention before a cosmetic change. Unknown indexing may deserve investigation before deleting a useful guide.
Group repeated observations when they share a cause. Several service pages with inconsistent terms may need one owner decision followed by coordinated copy changes. Keep dependent tasks in order so the team does not design around facts that are still undecided.
For each chosen action, write the page, responsible role and acceptance check. Leave the other findings recorded rather than turning them all into immediate work. You do not need a mathematical website score to justify a sensible sequence.
Use the worksheet to organize follow-up
The worksheet below summarizes this article into nine broader review stages. Its progress count tracks those nine stages, not the 40 individual questions above. For each stage, use the note field to reference the relevant checks, pages and outstanding tasks.
Not checked means you have not performed the review. Checked means you did it; a checked stage may still contain a problem. Needs follow-up identifies an outstanding investigation or action. Not applicable means the stage does not fit the defined review, and you should note why.
For example: Enquiry journey; checks 21–25 reviewed; contact test fails after correcting a required field; developer to reproduce, repair and repeat agreed delivery test. Keep the note limited to the information needed for follow-up.
The worksheet does not analyze your website, send a request or start a complete audit. Your marks and notes stay only in the open page and reset on refresh. Print before leaving if you need to retain them. Its progress count is not a quality score.
Questions about using the checklist
Which pages should I start with?
Choose the pages connected to your decision: main offer, supporting explanation or example, and the route to an enquiry. Add a guide or other entry page when its role matters. Record the selection so unreviewed pages are not silently included.
What can I check myself?
You can read the offer, follow navigation, compare related terms and inspect relevant proof. You can also perform an agreed enquiry test and record its result. Ask for help when a conclusion requires technical interpretation, private records or specialist user research.
What if I have no analytics?
Continue the public-page and journey checks. Describe what you can observe and leave conversion rate, traffic composition and business effects unconfirmed. Ask the owner which records exist before deciding that the information is permanently unavailable.
When should I involve a specialist?
When the issue cannot be reproduced or interpreted reliably, account evidence is needed, or a proposed change risks important journeys. Give the specialist the page, steps, evidence and unresolved question so the next check is specific.
When should I repeat the review?
Recheck changed areas after release and revisit the wider journey when services, audiences, navigation or enquiry processes change. Choose the timing by purpose and risk. Keep previous coverage and definitions if you plan a meaningful comparison.
Bring the observations together
Use the checklist to collect observations. Use a complete audit when you need those findings reviewed together and organized into a clearer plan. The sample report shows how a finding can become a specific task with a practical completion check.
Sources
Official guidance behind the search checks on this page.
- Google: Search technical requirements — checked
- Google: Canonical URLs — checked
- Google: Link best practices — checked
Turn the checklist into your next decisions
Request a complete website audit when you need the search, offer, proof and enquiry observations considered together, so your team can choose what to investigate or change next.
Request a complete website audit →