# Site Guard Page Guard, the checks it runs, and how fix suggestions improve your titles and meta descriptions. Page Guard reviews the pages you already have. It reads the parts of your site you selected, up to 1000 pages, and checks every page for three things: a click-through rate below what its search position should earn, a title outside the recommended length, and a meta description outside the recommended length. Pages that fail a check appear on the Site Guard page, where you may generate a suggested fix for each one. Every day, RankYak re-checks the pages it has already read against your latest Search Console data. So a finding appears the day the data shows it, and a page you fixed resolves without waiting for another crawl. The crawl itself, which reads your pages again, runs on its own about once a week, and the audit page shows the date of the last one. Generating fixes needs **Site Guard**, the $99-a-month add-on covering both Guards on one site. You may add it on the Site Guard page itself; see [billing](/help/billing) for how the charge works. Without it, a site still gets its weekly crawl, its daily re-check, and the count of open issues. What it does not get is the detail behind those counts or the fixes. The audit lives under **Site Guard** in the left-hand menu. With Site Guard on the site, the page carries two tabs: **Page Guard**, the one this article covers, and **Cannibalization Guard**, a separate audit for pages competing on the same searches, covered by the [cannibalization](/help/cannibalization) article. Everything on the Page Guard tab, including the page selection and the title branding list, applies to Page Guard only. Both Guards are built on your site's search performance data, so they require a [Search Console connection](/help/google-search-console). You should connect it before your first run; without the connection, an audit cannot start. ## Selecting pages Before the first audit runs, you should choose which parts of your site to include. Click **Select pages** to open the site tree, which is built from your site's sitemap. RankYak reads the sitemap once a day, so a new page appears in the tree within a day of being added to your sitemap. If a page you expect is missing, your sitemap most likely does not list it. Every section starts checked; you may uncheck a directory to skip it entirely, or open a directory to include or exclude individual pages. A more specific choice always wins, so you may exclude a directory and still keep one page inside it. The homepage sits at the top of the tree with its own checkbox, so you may leave it out, which is common when its title is deliberately brand-shaped, while the rest of the site stays selected. Once a selection is saved and Search Console is connected, RankYak crawls it on its own: the first crawl runs within a week, and the selection is crawled again about once a week after that. Between crawls, the daily re-check keeps your existing findings current against your latest search data. ## The checks Each audited page is fetched from your site and checked. - **Click-through rate.** Using your Google Search Console data from the past 30 days, the page's click-through rate is compared to what a page at its average position would typically earn. A page falling well short is flagged, since a weak title or description is usually what loses the click. - **Title length.** Titles shorter than 25 or longer than 65 characters are flagged: Google truncates longer titles, and very short ones leave the search snippet mostly unused. - **Meta description length.** Descriptions shorter than 70 or longer than 165 characters are flagged for the same reason. A page written mostly in Chinese, Japanese or Korean has its own ranges, because each of these characters takes about twice the width of a Latin letter, and Google cuts a search result by width. On such a page, titles shorter than 13 or longer than 33 characters, and descriptions shorter than 32 or longer than 75 characters, are flagged. Fix suggestions for such a page fit the same ranges. A finding stays open until a later check no longer detects it. When you apply a suggested fix in RankYak, the length findings it cures resolve at once. However, a title or meta description you change by hand shows at the next crawl of your site, which happens about once a week. A click-through rate finding resolves more slowly. The check reads the past 30 days of Search Console data, so the finding stays open until those days show the page earning the clicks its position should. The **Fixed** count on the overview includes a page as soon as you apply or dismiss its fix. A title or description finding that comes back after you apply a fix keeps the page on the **Handled** tab. A click-through rate finding is the exception. When a full 30 days of Search Console data from after the day you applied the fix show a low click-through rate, the page returns to the **Open** tab, marked **Low CTR after fix**, and its details show the click-through rate it had when you applied the fix next to the current one. Search Console publishes its data about two days late, so this happens about 32 days after the fix at the earliest. From there, you may generate a new fix or click **Mute issues**. If the click-through rate recovers later, the page moves back to the **Handled** tab on its own. ## Title prefixes and suffixes Most sites add a brand name to every page title, for example "Pricing | RankYak" or "RankYak | Robots.txt Validator", which would push almost every title past the recommended length through no fault of the page. The audit accounts for this: your first audit detects the branding your site adds, before or after the title, and records it on the audit page, and from then on titles are measured without it. Fix suggestions leave room for it too, so a suggested title still fits once your site adds the brand. Each entry carries its separator on the side that touches the title, so "| RankYak" is ignored at the end of a title and "RankYak |" at the start. You may also set the list yourself, before any audit has run or afterwards, by clicking it on the audit page. Once you save a list, it is yours: detection never overwrites it. Changing the list re-checks your existing findings immediately against the pages already audited, so a correction updates the audit without waiting for the next crawl. ## Fixing a page Clicking a flagged page opens its details, where **Generate fix** writes a suggestion covering what the findings implicate: a flagged description gets a new description, a flagged title a new title, and a low click-through rate both, since either half of the snippet may be losing the click. Whatever the findings did not implicate is left as it is. The suggestion is based on the page's content and on the search queries the page already earns clicks for, so the query users actually click on leads the copy. Nothing changes on your site when a suggestion is generated. You review the suggestion next to the current title and description first, and then decide. - **Pages RankYak published.** When the page is an article RankYak wrote for you, **Apply & republish** updates the suggested fields on the article and pushes the change to your site through your integration, the same way [editing an article](/help/editing-articles) does. - **Other pages.** For pages outside RankYak's control, you may copy the suggested title and description into your own CMS and click **Mark as resolved**. The next crawl verifies the change landed. If you would rather leave a page as it is, you may dismiss its fix; the page moves out of the list until you return to it. ## Cannibalization Guard RankYak also watches your site for cannibalization: several pages competing for the same searches and splitting the clicks between them. This is separate from Page Guard and needs no page selection; it lives on the **Cannibalization Guard** tab of the audit page and produces a consolidation plan and a watch list. The [cannibalization](/help/cannibalization) article covers how it works and how to resolve an entry.