The first version of this article, published in 2018, explained how to pick a preferred domain between the www and non-www versions, how to set a target country in the international targeting tab, and how to reach index coverage by clicking "Try the new Search Console". All three are now pointless: the screens have gone. The tool has been reshaped several times since, and plenty of tutorials still in circulation describe an interface that no longer exists.
Search Console is still the first thing we open when a director calls to say "our rankings are dropping". It is free, and fed by Google's own measurements rather than a third-party estimate.
Below are the Google Search Console reports that earn their place: what each tells you, and — more importantly — what it does not.
What has been removed since the last version of this article
Before the useful reports, some housekeeping. Four features described here in 2018 are no longer available:
- The URL parameters tool, which told Google to ignore certain sorting or filtering parameters. Its retirement was announced in March 2022 on the Search Central blog.
- The mobile usability report and the mobile-friendly test tool. The old help page now redirects to the Lighthouse documentation: that is where Google points you to test how a page renders on a phone.
- International targeting. Google's documentation on localised versions of a page now covers hreflang attributes only, with no dedicated report in the interface.
- The choice between www and non-www. There is no setting to change: your redirects and your canonical tag decide, and URL inspection tells you which version Google kept.
This is not trivia. Hunting for a screen that no longer exists costs you half an hour and usually ends with you convinced your account is badly configured.
Page indexing: the screen we open first
The page indexing report sorts your URLs into two piles: indexed and not indexed. The second pile is the interesting one, because Google states a reason for each entry. The labels we meet most often are "Discovered — currently not indexed" (Google knows the URL but has not crawled it), "Crawled — currently not indexed" (it read the page and chose not to keep it), "Duplicate without user-selected canonical" and "Soft 404".
Two limits are worth knowing. The report shows a sample only, up to 1,000 entries per reason: on a large catalogue you see the nature of the problem, not the full list. And "Validate fix" takes, according to Google, roughly two weeks. It is a queue, not a magic button.
An example from our own site. In 2018, Google had filed an old 2013 article under "Crawled — currently not indexed": the page was judged too thin. We reworked it and it did return to the index, but it never generated meaningful traffic, because nobody searches for the phrases it targeted. Reindexing a page does not create demand for it. And when a large share of your pages sits in that pile, the cause is structural and belongs to a wider review of how your site gets indexed.
Performance: three readings, and three traps
The performance report opens on the last three months by default. A useful addition is the 24-hour view, where each point represents an hour and holds preliminary data that may still shift. It is the only view offering hourly granularity, and daily granularity is not available within it. In practice we use it to check that a release has not cut traffic off, not to read a trend.
Three traps come up repeatedly. First, the totals in the chart and in the table do not always agree: Google explains this by differences in aggregation, by site or by page. Second, not every query is displayed. The documentation for the Search Console connector in Looker Studio says so plainly: to protect user privacy, Search Analytics does not show all data. Rare queries are hidden, so your keyword totals will always fall short of your total clicks.
Third, the export. The date picker stops at sixteen months, which rules out any three-year comparison, and the volume you can pull out is capped.
| Method | Volume available | Worth knowing |
|---|---|---|
| Search Console interface | Three months by default, sixteen months of history | Nothing to install, but rare queries are hidden |
| Search Analytics API | 25,000 rows per response, paginated; 50,000 rows per day and per search type | Rows can be dropped when grouping page and query together |
| Bulk export to BigQuery | No equivalent row ceiling | Not retroactive, first export up to 48 hours after setup, Google Cloud costs apply |
The API figures come from Google's guidance on retrieving all your data, which notes that rows may be discarded when results are grouped by page and query. The BigQuery figures come from the help page on bulk data export, which is explicit that the first export contains nothing predating its configuration. Hence the decisive argument: if you might need your history one day, connect the export now. For routine monitoring, a dashboard wired to Search Console is enough, and sits alongside the rest of your traffic measurement.
URL inspection: the answer page by page
The URL inspection tool answers the question "why does this page not show up". Hold on to one distinction: by default it shows the last indexed version, not the version currently online. The "Test live URL" button fetches the page in real time. The two can diverge, and the explanation you want often sits in that gap.
What we read first: the canonical URL Google selected — not always the one you declared — the date of the last crawl, the crawler used, and the sitemap that led to discovery. You can request indexing from here, but Google limits daily requests, states that submitting does not guarantee indexing, and warns that it can take up to two weeks. Sending the same URL ten times does not move it up the queue.
Core Web Vitals, manual actions, sitemaps and removals
The Core Web Vitals report rests on three metrics with precise thresholds: to be rated "Good", LCP under 2.5 s, INP under 200 ms, CLS under 0.1. INP replaced FID on 12 March 2024. The critical point is where the numbers come from: the CrUX report, meaning real visits over a rolling 28-day window. A fix deployed today will not show for several weeks, and a low-traffic site may have no data at all. It is why speed work has to be judged over a quarter, not a day.
The manual actions report is the one you hope stays empty. It means a human reviewer has judged that part of the site breaks the spam policies: unnatural links, thin content, cloaking, hidden text, site reputation abuse. Once fixed, you request a review; Google says this takes a few days to several weeks, longer for link issues.
The sitemaps report remains useful for confirming your file is actually being read. The official limits are 50,000 URLs and 50 MB uncompressed per file, beyond which you split it and use a sitemap index. Google itself notes that a small site of around 500 pages, where every page is reachable through internal links, probably does not need a sitemap at all. A sitemap does not rescue a shaky structure, and a technical review usually finds the navigation problem underneath.
Finally, the removals tool is widely misunderstood. It hides a URL from results for roughly six months. It does not take the page out of the index, and it does not stop Google crawling it. It is an emergency dressing, useful the day a confidential page turns up in search. Lasting removal needs deleted content, a 404 or 410 status, or a noindex tag:
<meta name="robots" content="noindex">How often to look at all this
In its getting-started guide, Google recommends a monthly check rather than a daily one, plus a look after any significant change. That is our practice: a monthly review of performance and indexing, a systematic check after a redesign or a migration, and the 24-hour view on release day. The rest of the time, the alert emails are enough.
Used this way, the tool detects symptoms rather than delivering a diagnosis. It tells you pages are not indexed, clicks have fallen, a template is slow. It does not tell you why, and it will not rank the fixes for you.
Common questions
Is Search Console free, and do I need Google Analytics as well?
Search Console is free and needs nothing more than a Google account. It works independently of Google Analytics: you can link the two to cross-reference data, but that is not a condition of use. The only requirement is proving you own the site, through a file on the server, a tag in the code or a DNS record.
Why does my total click count not match the sum of my keywords?
Google hides rare queries to protect users' privacy: they count towards the total but do not appear in the table. The totals in the chart and in the table can also differ depending on whether aggregation happens by site or by page. The gap is normal and does not indicate a fault.
How long does Search Console keep my data?
The performance report holds a rolling sixteen months, which makes multi-year comparison impossible from the interface. To keep more, extract the data through the API or switch on the bulk export to BigQuery. That export contains nothing predating its configuration, so enable it before you need it.
Does requesting indexing for a URL genuinely speed things up?
The URL inspection tool lets you submit a page, but Google limits daily requests and states that submitting does not guarantee indexing. Processing can take up to two weeks, and sending the same address several times gives it no extra priority.
Does the mobile-friendly test still exist in Search Console?
No. The mobile usability report and the mobile-friendly test tool have both been withdrawn, and the old help page now redirects to the Lighthouse documentation. To check how a page renders on mobile, use Lighthouse or PageSpeed Insights, plus the Core Web Vitals report for field data.
Does a temporary removal delete a page from Google?
No, it hides the URL from search results for around six months. The page stays in the index and Google keeps crawling it if it remains reachable. For permanent removal, delete the content and return a 404 or 410 status, or add a noindex tag.