Search checks

Every page audit, and every journey on every page it reaches, also checks for search, the work usually called SEO: whether Google can reach and index each page, how it appears in Google’s results, and whether it answers what people search for. Your customers often start their journey on Google, so a page they cannot find, or that does not answer their search, loses them before they arrive. The checks are UX Robot’s own rules, written from Google Search Central, Google’s own documentation for website owners, with Schema.org for structured data. Each rule links to the Google page it comes from.

• Crawl and index: the page answers with a 200 status, robots.txt and robots tags let Google in, one clear canonical, HTTPS, a sitemap, a readable address, and the content in the HTML the server sends, not only added by JavaScript.
• How it appears in Google: a title that says what the page offers, a meta description that sums it up, a main heading, the page’s language and a favicon.
• Content and intent: the page answers the searches that bring people to it, made for people with substance, clear who is behind it, headings that describe the content and nothing that looks like spam.
• Links, images and structured data: links Google can follow with words that say where they go, images with meaningful alt text, and valid markup that matches the page.
• Page experience, languages and regions: set up for phones, no pop-ups in the way, the main content first, and correct language versions when the website has them.
• In Google: with Search Console connected, whether Google has indexed the page, whether it uses your canonical, and whether its rich results have errors.

Search checks are part of every page audit on every plan, and every journey on Enterprise, at no extra cost.

• UX Robot reads the page as a crawler would. It leaves out a link’s tracking tags, such as utm_source or gclid, because they never change the page and Google indexes it without them, so a tagged link never reads as a messy address, and a canonical that drops your own campaign tag counts as the page itself. It fetches the page, follows each redirect, and reads the headers, the HTML and your website’s robots.txt. These checks need nothing from you.
• Google’s URL Inspection. With Search Console connected, UX Robot asks Google what it knows about the page: whether it is indexed, when it was last crawled, the canonical Google chose and any rich results it found.
• Your agent judges the rest in the browser. A step after accessibility and usability checks what only a person can judge, such as whether the title and content answer the searches that bring people to the page. Your agent is given what UX Robot already read, so it does not repeat it.
• In a journey, every page the journey reached. Once your agent has tested the journey, UX Robot reads each page it reached, up to twenty, the same pages the accessibility and usability checks cover. This takes about a minute and your agent waits for it. Your agent then checks every page in one step, against each page’s top searches from Search Console and the searches a visitor with the journey’s goal makes.
• Signed-out and signed-in pages. When your team completes a step such as signing up or signing in, the pages before it, including the sign-up or sign-in page, are checked before your team does the step, still signed out, as a new visitor and Google see them. Once the journey is done, UX Robot opens each page after the step signed out, as Google does. A page anyone can open, such as a public menu or profile, is checked for search like every other page, and your agent judges it in a separate signed-out window, so the page is checked as Google sees it. A private screen, such as a dashboard or an editor that sends signed-out visitors to sign in, is one Google cannot see, so search checks do not apply to it: the check table shows them as not applicable there and says which pages are private screens.
• Your own data. With GA4 and Search Console connected, your agent sees the page’s traffic and top searches, checks the page answers each search, and weighs every problem by how many people meet it.

• Fix list: every search problem sits in the same fix list as the accessibility and usability issues, labelled Search, with its severity, the exact markup that is wrong and the change to make. A problem that fails both an accessibility criterion and a search rule, such as images without alt text, is one issue: your agent is told what UX Robot’s own checks found, so when it finds the same problem it records it under the same id and your team gets one ticket for it.
• What we checked: the table shows every search rule with its result, under areas such as Search: crawl and index.
• How Google sees this page: whether it is indexed, when Google last crawled it, the canonical Google uses, the title in the HTML, the structured data found and whether JavaScript builds the page, or, when the website sent UX Robot’s server elsewhere, where it went and what your agent found on the page itself.
• In a journey, the same search problem on several pages, such as a missing meta description, is one issue listing every page it is on, and the table shows each rule’s result on each page.
• On Pro and Enterprise, every search problem becomes a tracked issue you can send to your tracker. When an audit of the same page, or the next run of the journey, no longer finds a problem UX Robot checks itself on any page it was on, it is marked fixed. For a problem your agent judges, such as a vague title, your agent says whether it is gone.

The free check on our home page also shows Lighthouse’s automated search score with the checks it failed. That is the start: the page audit adds every rule above.

Google changes its search documentation often. Every Monday UX Robot reads Google’s search documentation changelog. When Google has changed something, UX Robot reads the pages it changed and updates the rules that need it, citing the Google page and changelog entry behind each change. Most weeks nothing needs to change.

Each audit records the version of the search checks it used, and a rule’s id never changes, so an issue keeps its history from one version to the next. There is nothing for you to do.

• They check the page you audit, or the pages a journey reaches, signed-in pages included, not a crawl of the whole website.
• They describe what helps Google and searchers understand the page. Nobody can promise a ranking, and the report never does.
• Google’s URL Inspection needs Search Console connected, with the page inside the connected property, and Search Console allows a set number of inspections a day.
• If your website shows search engines a different version than visitors, the checks see what visitors get.
• Search demand, Google’s results page and competitors’ rankings come from DataForSEO in Enterprise journeys, not in page audits. See Journey analyses.