Every third party rank tracker is approximating what Search Console reports directly. It is the only source with actual impression and click data for your own site, and it is free.
It is also badly served by how most people use it. Opening the performance report, reading the four headline numbers, and closing it produces almost no insight. The value is in the filters.
Performance: Queries With Impressions and Almost No Clicks
This is the single most useful view in the tool and it takes two filters to reach.
Sort your queries by impressions descending and look for high impression terms with very low click through. Each one means Google is showing your page for that query and users are not choosing it.
Three causes and they need different fixes. Your title and description do not match what the searcher wanted, which is a rewrite. Your average position is low enough that impressions come from page two, which is a ranking problem. Or the query triggers an AI Overview or a featured snippet that answers it directly, which is the pattern covered in the pipeline measurement guide and which no title rewrite will fix.
Distinguish between them by checking average position on that query. Position three with poor click through is a title problem. Position fourteen is a ranking problem.
Performance: Queries You Rank For and Never Targeted
Filter queries to those where your average position sits between roughly five and fifteen. That band contains terms you are close to ranking well for without having tried.
These are usually the cheapest wins available. A page already ranking at position eight for a term it does not explicitly address often needs a section added rather than a new page written.
Cross reference against your keyword map, per the keyword research guide. Terms appearing here that are absent from your map are demand you did not know you had.
Performance: Page Level Rather Than Site Level
Switch the performance report from queries to pages and the picture changes. You see which URLs carry your traffic, which is almost never evenly distributed.
Most sites find a small number of pages carrying the majority of organic traffic. Those pages deserve disproportionate maintenance attention, per the refresh workflow guide, and they are the natural sources for internal links to pages you want to lift.
Filtering by page and then viewing queries for that specific page tells you what a page actually ranks for versus what you intended it to rank for. Those diverge more often than people expect, and the divergence often explains underperformance.
Page Indexing: Read the Exclusion Reasons, Not the Total
The indexed versus not indexed split is a headline number. The exclusion reasons underneath are the diagnosis.
Excluded by noindex tag on pages that should rank is a straightforward configuration error, and it is common after migrations where staging settings shipped to production.
Duplicate without user-selected canonical means Google chose a different canonical than you declared. Worth investigating rather than dismissing, because it usually indicates near duplicate pages competing, which is the situation covered in the canonicalization guide.
Crawled currently not indexed is the uncomfortable one. Google found the page, read it, and chose not to index it. That is a quality assessment rather than a technical fault, and no technical fix resolves it. It usually indicates thin or duplicative content.
Discovered currently not indexed is different and more fixable. Google knows the URL exists and has not crawled it, which typically indicates crawl budget constraints on large sites or a page sitting too deep, per the site architecture guide.
Core Web Vitals: Field Data, Grouped by Template
This report uses Chrome User Experience data from real visitors rather than lab simulation, which is why it disagrees with PageSpeed Insights sometimes and why it should be trusted over lab tools.
The grouping matters more than individual URLs. Search Console clusters similar URLs, which usually means it is showing you a template rather than a page. Fixing the template fixes every page using it, which is why the report groups them.
Check mobile and desktop separately. They routinely diverge and mobile is what is indexed.
My position on prioritising this report, consistent with the technical audit guide: fix templates failing thresholds and stop optimising templates that pass. Marginal improvement above the threshold is not where remaining effort belongs.
URL Inspection: The Rendering Check
Inspecting a live URL and viewing the rendered HTML shows what Google's renderer produced, which can differ substantially from what your server returned.
Where content appears in the rendered version and not in the raw server response, Google can usually handle it and other crawlers frequently cannot, which is the split covered in the rendering guide.
Use this after any front end change touching how content loads. Framework upgrades and component library changes routinely move content from server rendered to client rendered without anyone intending it.
What Search Console Does Not Tell You
Worth being clear about the limits, because people over read this tool.
It does not distinguish clicks from AI Overviews versus classic results. Both count as organic clicks from google, which means you cannot separate them.
Query data is sampled and filtered. Low volume queries are omitted for privacy reasons, so the query list is incomplete, particularly on smaller sites.
Average position is an average across all impressions, which blends position three in one country with position thirty in another into a meaningless middle number unless you filter by country.
It reports Google only. Bing has its own tools, and no AI engine reports anything comparable, which is the gap covered throughout this blog.
The Links Report Is Thinner Than It Looks
Search Console reports external links pointing at your site and it is not a substitute for a dedicated backlink tool. Coverage is partial, the data updates slowly, and it offers little for competitive analysis since it only covers your own property.
What it is genuinely useful for: seeing which of your pages attract links organically, which tells you what kind of content earns coverage without outreach. That informs the linkable asset decisions covered in the link building guide.
The top linking sites list is also worth scanning periodically for anything unexpected. Sudden appearance of unfamiliar domains linking at volume is worth investigating, though it is rarely as alarming as it first appears.
Sitemaps Report: Submitted Versus Indexed
The gap between URLs submitted in a sitemap and URLs actually indexed is a useful health metric that most people never look at.
A large gap means either the sitemap contains URLs that should not be there, redirected pages, noindexed pages, or duplicates, or the site has an indexation problem on pages you consider important. Both are worth resolving and they have different fixes.
The sitemap guide covers structuring submissions so this report is actually diagnostic rather than a single number blending everything together.
Filters Worth Setting Up Once
Country, if you operate in multiple markets, because unfiltered position averages across markets are close to useless.
Device, because mobile and desktop performance frequently diverge and mobile is indexed.
Search appearance, which lets you isolate rich result types and see whether schema implementation is producing anything.
Page path, which turns the whole tool into a cluster level report and is how to measure the topical structures covered in the content strategy guide.
Sixteen Months Is Your Whole History
Search Console retains sixteen months of data. Anything older is gone permanently and there is no way to recover it.
Export monthly if year over year comparison matters to you, which it does for seasonal businesses and for anyone who needs to demonstrate long term trend. Setting up an automated export via the API or a connected reporting tool takes an afternoon and preserves data you will otherwise lose silently.
Where Search Console Ends
Everything above concerns Google's classic index. It reports nothing about whether AI engines are citing your content, which is increasingly a meaningful share of how buyers encounter a brand.
The NotionCue Prompt Tracker covers that adjacent surface, running your target queries across major AI engines on a consistent cadence. It is not a replacement for Search Console and it answers a question Search Console structurally cannot.
Start your free NotionCue trial and run both. Search Console tells you what Google did with your pages. Prompt tracking tells you whether anyone else is citing them.
If you do one thing in Search Console this week: filter queries to those with over a thousand impressions and under one percent click through. That list is usually short, specific, and immediately actionable through title rewrites.
Common Questions
Why does Search Console show fewer clicks than my analytics platform?
They measure different things. Search Console counts clicks from Google results. Analytics counts sessions, which includes multiple visits from one click and excludes visitors who bounce before the tracking script fires. Discrepancy is normal and consistent direction matters more than matching numbers.
How often should I check it?
Weekly for performance trends on an active site, monthly for indexing and Core Web Vitals unless something changed. Daily checking produces noise rather than insight given how the data updates.
Should I submit URLs for indexing manually?
For genuinely new important pages, occasionally. As routine practice, no. A site relying on manual submission to get pages indexed has an architecture or quality problem that manual submission masks rather than solves.