Start with the decision, rather than a score
This sample uses Demo Website, a fictional business. It contains four teaching findings, not an audit of a live customer website. Read it to understand how an observation can lead to a useful task. There are no measured scores or business results.
Choose the finding closest to your question. If the offer is unclear, start with the enquiry-path example. If an important page should be found in search, start with search access. You do not need to work through every example before using the structure.
What needs attention?
Identify the specific page and the condition that was found. A report is easier to use when you can locate the issue.
What should we do next?
Look for the suggested change or investigation, its responsible role and the check that will confirm the work.
Follow one finding from observation to action
Take the enquiry-path example below. The fictional service page has a short promise and a form, but does not explain the included work or the response a reader should expect. Follow those facts through the recommendation before deciding how you would handle a similar question on your own site.
1. Locate the page
The example concerns /service/. That identifies where the information is missing. It does not say every page on the website has the same issue.
2. Read the observation
The page does not explain the included work or what happens after the form. This is a description of visible content, not a claim about visitor behaviour.
3. Check what supports it
The imagined page contains a short promise and a form. The example explicitly says that the scope and response expectations are absent.
4. Keep the open question visible
No form analytics or visitor research was measured. The example cannot tell you how many people left because of the missing explanation.
5. Choose the action and role
The business owner confirms the terms; the editor explains them on the page. Do not ask a designer to invent the deliverable or response process.
6. Check the changed page
A reader should be able to describe what they are enquiring about and what response to expect. Behaviour and business results need separate evidence.
Search: can the intended page be considered?
The fictional search finding concerns an instruction that tells a search engine not to index a page. Before changing it, the owner needs to confirm that the page is intended for search. A private or temporary page may have that restriction deliberately.
The next check is two-part: confirm that the deployed instruction matches the intended use, then obtain the search engine’s actual indexing state separately. Correcting an instruction and confirming indexing are different tasks.
Illustrative example. Demo Website and every finding below are fictional. No live website was measured; scores are not measured.
The imagined page contains a noindex instruction.
A fictional public service page is intended for search discovery. Example page: /service/
- Evidence
- The example response would contain a robots meta tag with noindex.
- Limitation
- This is not a retrieved response. Removing noindex alone does not establish indexing or ranking.
- Action
- Confirm the intended indexation. If the page should appear in search, remove the unintended instruction and check other response controls.
- Priority
- Confirm before changing
- Owner role
- Website developer
- Acceptance check
- The intended public response has no conflicting indexation instruction; the owner verifies indexing separately in Search Console.
- Measurement
- Not measured. No score is assigned.
AI readiness: are the business facts consistent?
The fictional offer pages describe support differently: one includes it, while another says it must be purchased separately. The useful task is to confirm the real terms and explain any legitimate exception.
That correction is valuable even without an AI visibility measurement. It does not establish that the contradiction caused a missing citation or an inaccurate answer. Any claim about an AI answer needs its own recorded observation.
Illustrative example. Demo Website and every finding below are fictional. No live website was measured; scores are not measured.
The pages give different explanations of what is included.
A fictional offer is described on two pages. Example page: /service/ and /terms/
- Evidence
- In this example, one page says support is included; another says it must be purchased separately.
- Limitation
- An inconsistency is not evidence that an AI system has cited or recommended the website incorrectly.
- Action
- Agree on the actual offer terms and align the relevant pages. Identify any legitimate exceptions explicitly.
- Priority
- Clarify the offer
- Owner role
- Business owner and editor
- Acceptance check
- Both pages describe the same approved terms or explain the different contexts. No AI visibility claim is inferred.
- Measurement
- Not measured. No score is assigned.
Content and conversion: does the next step make sense?
The fictional service page asks for an enquiry without explaining what is included or what happens after the form. The recommendation addresses that missing information using confirmed business terms.
The completion check concerns understanding: can a reader explain the offer and response before submitting? It does not turn a copy change into a measured increase in leads.
Illustrative example. Demo Website and every finding below are fictional. No live website was measured; scores are not measured.
The example page does not explain what is included or what happens after the form.
A fictional customer is deciding whether to make an enquiry. Example page: /service/
- Evidence
- The imagined page shows a short promise and a form, without scope or response expectations.
- Limitation
- No form analytics, visitor research or conversion rate was measured. The effect on enquiries is a hypothesis.
- Action
- Explain the included work, relevant conditions and next step using approved information.
- Priority
- Improve clarity
- Owner role
- Business owner and editor
- Acceptance check
- A reader can identify what they are enquiring about and what response to expect before submitting. Measure behaviour separately if available.
- Measurement
- Not measured. No score is assigned.
Competitors: which buying question is left unanswered?
The comparison asks the same question of three fictional alternatives: which situations are outside the offer? Alternative A explains exclusions; Demo Website does not. This is a prompt to confirm and explain relevant exclusions on your own page.
It is not a ranking of the businesses. The example does not measure sales, service quality or search position. The useful difference is the information a buyer can find before making the next decision.
Illustrative example. Demo Website and every finding below are fictional. No live website was measured; scores are not measured.
Alternative A explains exclusions; Demo Website does not.
Three fictional alternatives address the same customer need. Example page: /service/
- Evidence
- The fictional comparison uses the same question: which situations are outside the offer?
- Limitation
- These are invented alternatives, not evidence that another business performs better or ranks higher.
- Action
- Confirm which exclusions matter and describe them on the relevant offer page, using your own wording.
- Priority
- Check relevance
- Owner role
- Business owner
- Acceptance check
- The comparison records the same question and comparable offer context; your page accurately explains relevant exclusions.
- Measurement
- Not measured. No score is assigned.
Read the reason for a priority
A useful priority gives you a reason to act. Clarifying an offer can be important because the team first needs one accurate explanation. A technical instruction may require confirmation before a developer changes it. A suspected behaviour problem may need investigation before money is spent on a redesign.
Treat priorities as decisions to discuss in your own context. The sample does not disclose a scoring formula or estimate the revenue effect of each task. Ask which customer route is affected, how certain the finding is and whether another task must happen first.
A repair
The condition can be reproduced and the intended behaviour is known. Define the change and the check that will confirm it.
An investigation
The explanation is still uncertain. Define the missing information and the person who can obtain it before choosing a larger change.
Use the report with your team
Choose one important finding and turn it into work someone can finish. Add the page, confirmed business facts and responsible role to your task list. Separate the person who knows the terms from the person who publishes the page.
Agree the completion check before work starts. After the change, revisit the same page and question. If you want to understand later sales or search results, keep the relevant measurement and context separate from the direct check of the repair.
| Task detail | What to write |
|---|---|
| Page | /service/ — the fictional page shown in the finding. |
| Owner input | Confirm the included work, conditions and actual next response. |
| Page change | Explain those facts before the form in language a new buyer can understand. |
| Responsible role | Business owner for the terms; editor for the explanation. |
| Completion check | Ask a reader to describe the offer and response without help. |
| Later question | Investigate behaviour with suitable data; do not infer uplift from publishing the copy. |
What this sample cannot tell you
The sample contains no customer analytics, live search results, observed AI answers or real competitor measurements. Its examples explain the structure of a finding. They do not promise that every order has the same issues, fields or number of recommendations.
For your own report, page selection and available sources affect what can be concluded. A public page review can identify missing information. Account-only evidence, customer behaviour or delivery checks may require additional access and a separate agreed check.
- Read the pages and sources covered by the finding.
- Keep missing information as an open question, rather than treating it as zero.
- Distinguish a suggested action from a confirmed business outcome.
How to judge the report for your own decision
Ask whether you can follow the explanation: what was seen, what supports it, what remains unknown and what would settle the next question. A longer list of issues is not necessarily a better basis for spending money.
When you request your own audit, bring the website and decision you need to make. Confirm the relevant pages, available evidence and deliverables through the existing enquiry process. Use this sample to discuss the kind of explanation that would help your team act.
Want to apply this structure to your own website?
Bring the website and the decision you need to make. The complete audit can connect the relevant findings into a practical improvement plan, with the actual coverage and deliverables agreed for your enquiry.
Request a complete website audit →