Screaming Frog SEO Spider
Every address fetched individually — not extrapolated.
Screaming Frog is a crawler for your own machine: it calls up every address on a website individually and records what comes back – status code, title, description, headings, redirects, image text, canonical references. No sampling procedure, no extrapolation.
How we assess it
Conexa Digital works with Screaming Frog, and does so at precisely the point where most inspections stop too early. We do not measure what is in the source code, but what arrives on the wire: every address from the sitemap, fetched individually, with a random parameter against the cache. At our last complete inspection that came to 275 addresses on conexadigital.eu and 68 on a client site – zero failures. We carry out this kind of inspection before and after every release.
Relevant to these services
What you get out of it
The most expensive faults on websites are the silent ones: a redirect chain that bleeds ranking strength, a canonical reference pointing at the wrong address, a page nobody links to any more. None of it is visible at a glance, and no visitor reports it. A complete inspection finds it before it costs months.
In practice
After a changeover, a site reported clean on every sample – until the complete inspection showed that a single overly broad redirect rule was intercepting every address beginning with a particular letter. The pages affected were ones that appeared in no sample. That is exactly the difference between "we checked it" and "we fetched every address individually": the fault was not where anyone would have suspected it, and an extrapolation would never have found it.
When we do not use it
On a website with five pages a crawler is excessive – everything there can be checked by hand. And a crawler answers no question of substance: it tells you that a page is reachable and has a title, not whether the text is any good. Where the actual problem lies in the content, a technical inspection is no help.
Frequently asked questions
What does an inspection like that typically find?
Internal references pointing at redirects instead of the destination, duplicate or missing page titles, images without alternative text, orphan pages linked from nowhere, and canonical references pointing at the wrong version.
Doesn't Google Search Console do that as well?
In part, but with a delay and only for what Google has already visited. Your own crawl shows the current state immediately and in full – including the pages Google has never retrieved.
How often should an inspection be run?
Before and after every larger release, and regularly after that. Most faults do not arise during creation, but at the tenth change.
