We can say
A link returns 404. A control has no visible effect. Several pages share the same title. A public-page attribute has a measured value.
Evidence before certainty
How an independent website review reaches a useful judgement.
We read the public website as a visitor can see it. A GreatInternet reviewer then checks the evidence, challenges possible capture artefacts, and prepares one clear report: what matters, what is working, what to leave alone and the next sensible move.
The evidence boundary
What we cannot see is part of the answer.
A public website review cannot see source code, deployment settings, private analytics, customer research or internal systems. We do not fill those gaps with confidence. Observable behaviour is reported as evidence; an inferred cause is labelled as a hypothesis.
We will not say
The hidden code is poor, a particular refactor is required, a result is guaranteed, or the website is certified or compliant.
From pages to priorities
A consistent sweep, followed by selective judgement.
01
Choose the useful pages
We focus on the public routes that explain the offer, build trust and let a visitor act.
02
Collect what can be checked
Visible copy, links, screenshots, metadata, response behaviour and inspection limits are brought together.
03
Challenge the instrument
Before calling something broken, we check that our own capture can scroll, focus, load and activate comparable controls.
04
Read the pages as a set
We follow important journeys and look for cross-page patterns that disappear when each page is considered alone.
05
Group symptoms
Several small problems may share one cause. We rank the underlying issues by their consequence for visitors.
06
Protect the good work
The strongest page or pattern becomes a benchmark to preserve and repeat, so the review also reduces unnecessary change.
Visible confidence
Confidence describes the evidence, not our mood.
If we cannot distinguish the website’s behaviour from our own capture, the candidate finding is withheld rather than published.
High
The mechanism is visible or measured in the captured evidence.
Medium
The pattern is consistent and repeatable, but its cause is inferred.
Low
The evidence is limited, so the report says what would strengthen it.
The consistent sweep
What every review looks across.
Contact and conversion routes; cross-page clarity; titles, descriptions, sharing metadata and canonicals; structured data; links and downloads; cookies and consent; missing-page behaviour; JavaScript and reduced-motion resilience; page weight and timing. The aim is not a long checklist. It is the few observations that materially change a decision.
Independent by design
A direction, not manufactured work.
The review prioritises and recommends. It does not turn every observation into a rebuild we hope to sell. Strengths and what to leave alone are first-class parts of the report.
Why we do not lead with an automated score →Start with the website your visitors can see.
A free outside view shows what stands out and the first useful next moves. Go deeper only when it is useful.