The Google Search Console Guide: Every Report, What It Answers, and What It Hides
Search Console has seven report groups and five limits nobody mentions. What each report actually answers, which property type to create, and where the data stops.
Search Console is Google publishing its own records about your site: what you ranked for, what it indexed, and what it refused to. It is free, it has no paid tier, and it is the only source of that data. The shape of it:
- Getting in: search.google.com/search-console, signed in with a Google account. Create a domain property and verify it with one DNS TXT record, which covers every subdomain and both protocols at once.
- Seven report groups: Performance, Indexing, Experience, Enhancements, Security and manual actions, Links, and Settings. Each answers one question, and most people only ever open two of them.
- The report you will live in: Performance. Clicks, impressions, CTR and average position, sliced by query, page, country, device and date, with separate views for Search, Discover, News and generative AI features.
- What it cannot tell you: Why a page lost. Search Console reports Google's verdict, never the cause. It also cannot see a page Googlebot never reached, which is why orphans are invisible in it.
- The five limits: 16 months of history, 1,000 rows per UI export, anonymised queries missing from the totals, one URL at a time in inspection, and Google's view only. None of them have a paid tier above them.
Verify the property today even if you never open it. Data accrues from the verification date and Google backfills nothing.
CrawlRaven connects to the Search Console property this guide teaches you to create, so this page has an obvious bias. It is written to be useful on its own: everything below works whether or not you ever connect the property to anything. Try CrawlRaven free: 1 site, no credit card →
Most Search Console guides walk the sidebar top to bottom and describe every screen. That produces a tour, not an understanding, and it is why so many people have had a verified property for three years and still only open one report.
This one is organised around the question each report answers, and it is honest about the ones that answer nothing useful. It also states the limits up front rather than letting you find them the hard way, six months into a project.
What Search Console actually is
Search Console is Google publishing its own records about your site. Not an estimate, not a crawl, not a third-party index: the actual searches Google served you for, and the actual decisions it made about your URLs.
That distinction sets the boundary of the whole product:
- It only knows what Google knows. A page Googlebot never reached does not exist in any report, which is exactly why orphan pages are invisible here and have to be found by crawling.
- It reports verdicts, not causes. "Crawled, currently not indexed" is a decision. The reason behind the decision is not in the report.
- It is free with no tier above it. Nothing in Search Console can be unlocked by paying Google, which makes its limits real constraints rather than upsells.
- It is the source other tools read from. Most paid SEO platforms connect to your property over OAuth and display its data. That is an integration, not a replacement, which is the honest answer to what can actually replace Search Console.
How to get to Search Console and add a site
The URL is search.google.com/search-console. Sign in with a Google account and you are in. There is no separate signup, no billing step, and no app.
- Open
search.google.com/search-consoleand sign in. - On the welcome screen, or from the property selector at the top left, choose Add property.
- Pick Domain and enter your bare domain, with no protocol and no www.
- Copy the TXT record Google shows you into your DNS, then click Verify.
- Wait. Reports populate over the following days, and nothing before today is ever backfilled.
That last point is the only genuinely urgent thing in this guide. A property verified in March has no February, permanently. Verify sites you are not working on yet.
The three property types, and which to pick
This is the one decision that is annoying to undo, because a new property starts its history from scratch. Get it right the first time.
The three Search Console property types
A domain property is the right answer for almost everyone. It covers every subdomain and both protocols in one place, so a stray http://www. version cannot quietly report somewhere you are not looking.
URL-prefix properties earn their place in two situations: you cannot get at DNS, or you genuinely want one section reported in isolation. A large site with a separate blog subdomain and separate owners is the usual legitimate case.
Platform properties are a different animal, added in July 2026. They report Google traffic to an Instagram, TikTok, X or YouTube account, and adding a social account to Search Console works nothing like verifying a domain.
The verification methods, ranked
- DNS TXT record. The only method that unlocks a domain property, and the most durable: it survives redesigns, CMS migrations and theme changes.
- HTML file upload. Reliable, but dies the moment someone rebuilds the site and forgets to copy the file across.
- HTML meta tag. Fine, with the same fragility. Most CMS platforms have a field for it, which is why Squarespace and Shopify setups usually take this route.
- Google Analytics or Tag Manager. Convenient, and it breaks whenever someone touches the tag setup. Treat as a fallback.
The report map: what each group answers
Search Console has seven groups of reports. Most people use two. Knowing what the other five answer is the difference between a dashboard and a diagnostic.
The Search Console report map
| Group | Reports inside it | The question it answers | Freshness |
|---|---|---|---|
| Performance | Search results, Discover, Google News, Generative AI features | What did people search before they saw me, and did they click? | Daily, 16-month window |
| Indexing | Pages, Video pages, Sitemaps, Removals | Which of my URLs are in the index, and why are the rest not? | Recrawled on Google's schedule |
| Experience | Core Web Vitals, HTTPS | Are real visitors hitting slow or unstable pages? | 28-day rolling field data |
| Enhancements | One report per structured-data type Google detects | Is my markup valid, and is it eligible for a rich result? | Appears only once Google finds the markup |
| Security and manual actions | Manual actions, Security issues | Has a human reviewer or a malware scan flagged my site? | Event-driven, emailed when it fires |
| Links | External links, Internal links | Who links to me, and which of my pages do I link to least? | Refreshed periodically, counts are approximate |
| Settings | Crawl stats, Users and permissions, Ownership verification | How is Googlebot spending its time, and who else can see this? | Crawl stats: trailing 90 days |
The groups are not equally useful, and pretending otherwise is why tours of the sidebar fail. Performance and Indexing carry the work. Security and manual actions matters enormously on the rare day it fires and not at all otherwise.
Performance: the report you will live in
Four metrics, and every serious question you ask of Search Console is some combination of them:
- Clicks: someone clicked through to your site from a Google result.
- Impressions: your URL appeared in results a searcher saw.
- CTR: clicks divided by impressions, which is a listing problem when it is low and rankings are fine.
- Average position: the mean of your best ranking per query, which is an average of averages and should never be reported on its own.
Those four slice by query, page, country, device, search appearance and date. The search-type toggle switches between Search results, Discover, Google News and the generative AI features report Google added in June 2026.
Filters are where the report becomes useful. Regex filters on query or page turn it into a real analysis tool, and the free GSC regex generator writes the RE2 pattern for you when the syntax fights back.
Three filter patterns worth learning on day one:
- Split brand from non-brand. A query regex that excludes your brand name separates demand you earned from demand you already had. The mechanics are in branded versus non-branded queries.
- Filter to positions 11 to 20. These are pages one good edit from page one, which is the striking distance play and usually the highest-return hour in the account.
- Group by section. A page regex per template or directory turns a site-wide average into content groups, so a growing section stops being hidden by a dying one.
Average position is the metric I would remove from the interface if I could. It is an average of averages, it moves when a page you do not care about enters or leaves the index, and it gets pasted into client reports as though it were a rank.
Clicks and impressions are counts of things that happened. Position is a derived statistic doing a poor impression of one, and the number of arguments I have had about a 0.4 shift in it is not a good use of anyone's quarter.
Indexing: Pages, Sitemaps, Removals
The Indexing group answers a single question with three tools: which of my URLs are in Google, and what happened to the rest.
- Pages. Every URL Google knows about, split into indexed and not indexed, with a reason per URL. The reasons are jargon-heavy and each means something specific, which is why every page indexing status has its own explainer.
- Sitemaps. Submit an XML sitemap URL, then read the discovered count against your real page count. A large gap is a finding on its own, and the free sitemap validator checks the file before you submit it.
- Removals. Temporarily hides a URL from results for about six months. It is not deletion, and removing a site from Search Console properly means understanding which of four different things you actually want.
URL Inspection sits above these in the sidebar and is the single most useful button in the product. Paste a URL and Google tells you whether it is indexed, when it last crawled it, which canonical it picked, and what the rendered HTML looked like.
Its limit is that it works on one URL at a time. When a template breaks across four hundred pages, inspecting them individually is not a plan; the free page indexability checker covers the bulk case.
Links, Experience, and the reports that find you
Three groups that most people never open, two of which are worth ten minutes a month.
- Links. External links, internal links, top linking sites and top linking text. The counts are approximate and the sample is partial, so use it to spot patterns rather than to measure link growth. Its most underused screen is internal links, where your least linked-to pages are a to-do list.
- Experience. Core Web Vitals built from real visitor data over a trailing 28 days, grouped by URL pattern. Field data changes slowly, so do not expect a fix to show up this week.
- Security and manual actions. Almost always empty, and the one report you genuinely want an email from. A manual action means a human reviewer acted against your site.
Enhancements reports appear only when Google finds structured data on your site, one report per type. They validate markup; they do not promise a rich result, and Google has retired several rich result types while leaving the reports in place.
Settings: crawl stats, users, and who else can see this
Settings is where the property administration lives, and two screens in it matter more than their placement suggests.
- Crawl stats. Ninety days of Googlebot requests, split by response code, file type, purpose and Googlebot type. This is where a crawl budget problem becomes visible: a large site spending its crawl on redirects and 404s is a chart, not a theory.
- Users and permissions. Who can see this property and what they can do. Owners can add and remove people, full users can act on everything except user management, and restricted users can mostly only look.
- Ownership verification. Which method is currently proving you own the property. Worth a glance before a site migration, because a redesign that drops the verification file silently unverifies you.
For agencies, the permissions screen is the one to get right at the start. Adding a client contact as an owner is easy and reversing it later is a conversation.
The five limits everyone eventually hits
These are not bugs and none of them can be paid away. Knowing them early changes how you set the account up.
The five limits everyone eventually hits
| Limit | Value | Who it bites | The way around it |
|---|---|---|---|
| Performance history | 16 months | Year-on-year comparison beyond one cycle, and any migration older than that | Export it yourself, starting today |
| Rows per UI export | 1,000 | Any site whose long tail is the point | The API returns far more per request; BigQuery returns all of it |
| Queries shown | Anonymised below a threshold | Rare and personal searches vanish, so your totals never reconcile | Read totals from the chart, not the sum of the query table |
| URL inspection | One URL at a time | Diagnosing a template failure across 400 pages | Crawl the site and check indexability in bulk |
| What it can see | Google's view only | Anything Googlebot never reached is invisible, orphans included | A crawl of your own site is the missing half |
The 16-month window is the one that costs people real work. Every morning the oldest day drops off, and keeping Search Console data past 16 months only ever preserves history from the day you start.
The row cap is the one that misleads. A 1,000-row export from a site with 40,000 queries is not a sample of your search demand, it is the top of it, and the BigQuery bulk export is the route to the rest.
The anonymised queries are the one that starts arguments. Google withholds rare searches to protect privacy, so the query table never sums to the chart total, and anonymous queries explains how large the gap usually is.
The question Search Console cannot answer
Everything above is Google telling you what it decided. Not one report tells you why it decided it, and that is a structural fact rather than a gap Google is going to close.
Three examples of the shape of the problem:
- A page drops from position 4 to 19. Search Console shows the drop with a date. It cannot show you the internal link that disappeared in the redesign three weeks earlier.
- Forty URLs land in "Crawled, currently not indexed". The status is the verdict. That they all share one template with near-identical copy is not in the report.
- Clicks fall while impressions hold. A CTR problem is visible. Whether the title got truncated, or a competitor added review stars, is not.
Each of those needs a second data set: your site as it actually is, which means a crawl. That is also why GSC and GA4 numbers never match matters more than it sounds, since the join is where diagnosis starts.
Where to go deeper on each report
This guide is the map. Each of these covers one part of it properly.
- Indexing: every page indexing status explained, and removals versus deletion versus noindex.
- Performance analysis: striking distance keywords, branded versus non-branded, content groups, and content pruning.
- Getting the data out: the Search Console API, the BigQuery bulk export, keeping data past 16 months, and reading Search Console from Claude over MCP.
- Joining it to other sources: why GSC and GA4 disagree, Ahrefs versus GA4 versus GSC, and finding orphan pages across all three.
- Measuring change: SEO A/B testing in GSC, tracking a site migration, and SEO annotations.
- The 2026 additions: the generative AI features report, platform properties, and everything else Google shipped this year.
- Setup on hosted platforms: Squarespace, Shopify, and Weebly and Wix.
Key Takeaways
- →What it is: Google publishing its own records about your site: the queries it served you for, the URLs it indexed, and the ones it refused. Free, with no paid tier above it.
- →Where to start: search.google.com/search-console, a domain property, verified with one DNS TXT record. It covers every subdomain and both protocols so nothing reports somewhere you are not looking.
- →The two reports that carry the work: Performance for what you ranked for and whether anyone clicked. Pages for what got indexed and why the rest did not. The other five groups are situational.
- →Learn filters before anything else: Regex filters on query and page turn Performance from a dashboard into an analysis tool: brand versus non-brand, positions 11 to 20, and per-section content groups.
- →The limits are real: 16 months of history, 1,000 rows per UI export, anonymised queries missing from the query table, one URL per inspection, and Google's view only. None can be paid away.
- →What it will never tell you: Why. Every report is a verdict without a cause, and the cause almost always lives in your own site rather than in Google's record of it.
- →Do this today: Verify every property you might ever want data for, including sites nobody is working on yet. History accrues from the verification date and Google backfills nothing.
Primary sources
Frequently asked questions
What is Google Search Console?
Google Search Console is a free Google service that reports how your site performs in Google Search. It shows the queries you appeared for with clicks, impressions, CTR and average position; which URLs Google indexed and why it skipped the rest; Core Web Vitals from real visitors; structured-data validity; manual actions and security issues; and the sites linking to you. It is Google publishing its own records about your site, which is why no other tool can replace it.
How do I get to Google Search Console?
Go to search.google.com/search-console and sign in with a Google account. There is no separate signup, no billing step and no app to install. If you have never added a site, the welcome screen asks you to create a property; if you have, the property selector at the top left switches between the sites you own or have been granted access to.
What is Google Search Console used for in SEO?
Four jobs, in practice. Finding the queries you already rank for so you can improve pages that are close rather than write new ones. Diagnosing why pages are not indexed. Catching manual actions, security issues and indexing failures before traffic drops. And measuring whether a change worked, since it is the only source of Google's own before-and-after numbers for your pages.
Is Google Search Console free?
Yes, entirely, with no paid tier above it and no feature gated behind a subscription. The limits it does have, 16 months of performance history and 1,000 rows per interface export, apply to every account equally and cannot be raised by paying Google. The BigQuery bulk export gets you past the row cap, though Google Cloud charges for the storage and queries.
Which Search Console property type should I create?
A domain property, in almost every case. It covers every subdomain and both protocols in one place, so http, https, www, non-www and any subdomain report together and nothing falls through a gap. It requires a DNS TXT record. Choose a URL-prefix property only when you cannot reach DNS, or when you deliberately want one section reported separately.
How long does Search Console keep my data?
Performance data runs on a rolling 16-month window, and every day the oldest day drops off permanently. There is no setting to extend it and no way to recover what has already gone. If you want longer history, you have to capture it yourself through the BigQuery bulk export, scheduled API pulls, or a tool that stores it, and each only preserves data from its own setup date onward.
Why do Search Console's numbers not match Google Analytics?
Because they count different events. Search Console counts a click on a Google result; GA4 counts a session that started on your site. A click that bounced before the tag fired is a Search Console click and not a GA4 session, and one session can follow several clicks. Expect a persistent gap, and treat the two as answers to different questions rather than as one number to reconcile.
Does Search Console show every query I ranked for?
No. Google withholds queries made by very few users to protect searcher privacy, so rare and personal searches never appear in the query table. That is why the query rows never sum to the total shown on the chart. The gap is normal and often large on small sites; read totals from the chart and use the query table for relative comparison.
Can Search Console tell me why my rankings dropped?
It can tell you what happened and when, never why. You get the clicks, impressions and average position for the affected pages and queries, plus any manual action or indexing change. The cause, whether a competitor improved, a template broke, an internal link vanished or a core update landed, is not in any Search Console report. Diagnosis needs a crawl of your own site alongside the data.
Do I still need Search Console if I use a paid SEO tool?
Yes, and most paid tools assume you do. They connect to your verified property over OAuth and display its data alongside their own, which means the property is the source they read from rather than something they replace. Keep it verified even if you never open it directly, because manual actions and security issues arrive there first.
How long does it take for data to appear in Search Console?
Performance data typically starts appearing within a couple of days of verification, and indexing reports fill in as Google recrawls. Nothing is backfilled: a property verified today has no history from yesterday and never will. That is the practical argument for verifying a site the day it launches, even a site nobody is working on yet.
Can I add my Instagram or TikTok account to Search Console?
Yes, through platform properties, which Google launched on 7 July 2026 for Instagram, TikTok, X and YouTube. You verify the account by signing in to it rather than by owning a domain, and you get a Performance report covering the traffic Google Search, Discover and News sent to those posts. It reports nothing about in-app reach and gives you no technical controls.
15+ years of growing SaaS websites through SEO | Author, 200-Point Audit Checklist
Aditi has spent 15+ years helping SaaS companies scale organic traffic through technical SEO and content strategy. She is the author of the CrawlRaven 200-Point Audit checklist used by agencies and in-house teams to systematically improve search performance.