Auditing a site you inherited
An audit that returns forty findings is not a to-do list. Three of them matter and the rest are noise until those three are done.
Web Analyzer runs SEO analysis with concrete recommendations, performance scoring, security vulnerability scanning, WCAG accessibility checks, mobile responsiveness validation, and writes an AI summary of what it found.
Run it once on a site you did not build and you get forty findings. That number is the problem: forty findings feel like a project, and projects get postponed.
Fix in this order
The findings are not equally weighted, whatever the score says.
- Security first. An exposed admin path or an out-of-date library is the only category that gets worse on its own while you decide.
- Then anything blocking indexing. A stray
noindexor a robots rule beats every other SEO finding combined — the page is invisible, so nothing else about it matters. - Then performance, but only the largest number. One oversized hero image usually accounts for more than the next nine findings together.
- Then the rest, when there is time.
The accessibility section is not optional in the way people assume
WCAG findings get filed under "nice to have" and they should not be. Contrast failures and missing labels are the two that appear on almost every audit, both take minutes, and both affect real visitors immediately rather than theoretically.
They also overlap with SEO more than people expect: an image with no alt text is a finding in two sections at once.
Read the AI report last
The generated summary is good at telling you what a set of findings adds up to — "this site is fast but invisible to search" is more useful than forty rows. It is not good at prioritising against your business, because it does not know your business.
So use it to understand the shape, then apply the order above. And re-run after the first three fixes rather than after all forty; the second audit is where you find out whether the fix worked.
Questions people actually ask
Where should I start with forty findings?
Security, then anything blocking indexing, then the single biggest performance number. Those three usually account for most of the real damage; the remaining findings are worth doing but not worth blocking on.
Can I audit a site I do not own?
Yes — the scan works against any public URL, which makes it useful for looking at a competitor or at a site a client is asking you to take over.
How often should I re-run it?
After each round of fixes rather than on a schedule. A re-run right after changes tells you whether they worked; a monthly re-run on an unchanged site tells you nothing new.
