Docket for your situation
The same 99 checks run on every site, but which findings matter most changes a lot by who you are. These pages cover what to look at first.
That is not a marketing framing. Docket scores every lane on every site, and the same finding genuinely carries different weight depending on how the business gets found.
Why the same finding lands differently
Missing LocalBusiness schema is close to fatal for a business that lives on
"near me" searches and completely irrelevant to a SaaS company selling to another country. A
slow product page costs an ecommerce site money on every session and costs a reference site
almost nothing. Neither of those is a judgement about how good the site is; they are facts
about which door customers come through.
The place this matters most is the middle of the list. Critical findings are critical for everyone — a page nobody can index cannot help anybody. It is the twenty MEDIUM findings underneath where knowing your own situation turns a report into a plan, and where a tool that only knows how to sort by severity leaves you to guess.
Two things are worth reading whoever you are, because they now apply to every kind of site: whether AI answer engines can reach and cite you, and which of your pages an AI answer replaces outright. Those were niche two years ago.
For SEO agencies
Unlimited client audits with no per-seat or per-crawl cost, and a client-ready PDF that does not need rebuilding in a deck.
For developers
The CLI as a deploy gate — exit codes, a working GitHub Actions job, and what it costs to run on a macOS runner.
For online shops
We audited ten large retailers before writing this one, and the result contradicted the pitch: none had broken product markup. What did show up, and where Screaming Frog is the better tool for a catalogue.
For local businesses
Why you are not in the map pack — LocalBusiness schema, NAP consistency, and the geo signals that decide "near me" results.
For SaaS companies
0 of the 20 SaaS sites in the Index block an answer engine, so the blocking scare is not the story. The story is a marketing page that is empty until JavaScript runs, and a docs subdomain competing with it.
One more coming
An in-house marketing page will appear when it has something specific to say. Near-duplicate pages are the real flag risk in a programmatic set, not thin ones, so a page gets written when there is a measurement behind it rather than to fill a gap in a list.
That standard is why the ecommerce page took as long as it did: the honest version could only be written after auditing real shops, and what came back disagreed with what the page was going to say. The SaaS page went the same way — the pitch it was going to make is the one its own data contradicts. If the last one never produces a finding worth publishing, it will not be written, and this paragraph will say so rather than promising it indefinitely.