A marketer opens Search Console for the first time, sees the clicks line sloping down, and panics. Then she notices average position has improved and relaxes. Both reactions are wrong for the same reason: every number in Search Console is honest on its own and misleading on its own. It's the only place Google tells you directly what it did with your site, and once you can read four reports and the three ways they mislead, you'll stop guessing why traffic moved.
Set it up so the data is complete
Search Console reports on a property, and the property type decides what you see. A URL-prefix property covers one exact prefix, so https://www.example.com excludes http://, the bare domain and every subdomain. A domain property, verified with a DNS record, covers all of them at once, so add one if you only have a URL prefix.
Verify it today even if you won't read it for months. Google's Search Console help says the Performance report keeps sixteen months of history, so the year-on-year comparison is there when you need it. Add colleagues as users, and keep email notifications on so you hear when indexing breaks or a manual action lands.
The Performance report is four numbers that only make sense together
Open Performance and you get clicks, impressions, average CTR and average position for the search type you've chosen (web by default; image, video, news and Discover are reported separately).
An impression counts when a link to your site appears on the results page the searcher is looking at, whether or not they scroll to it, per Google's documentation of the report. Average position is the position of your highest-ranking result for each impression, averaged across all of them. Together they produce the first lie: average position improves when you stop appearing for queries where you ranked badly, and worsens when you start appearing for new ones where you rank low. A site that launches a batch of new pages often sees its average position fall while traffic rises. Read position with impressions beside it, never alone.
The second lie is in the query table. Google removes rare and personal queries from the list for privacy, which its documentation calls anonymised queries, so the rows never add up to the headline total. Treat the table as a sample, not a ledger.
The third is aggregation. By default the report counts per property, so two of your pages ranking for the same query is one impression at the better position; switch to the Pages tab and each page is counted on its own, so the same query can show a worse position. Don't compare numbers across the two views.
Click + New to filter by query, page, country, device or search appearance, and use the date picker's Compare tab for two like-for-like windows. Three views are worth building every month:
- Queries not containing your brand name, the last three months against the previous three. This is your non-brand organic trend, the number your SEO work actually moves.
- Pages sorted by clicks, compared against the same period last year. Pages that slipped go to Refreshing content that slipped.
- Queries on page one with a CTR well below your other page-one queries. Those are snippets doing a bad job, and the cheapest wins in SEO.
Leave the most recent few days out: data arrives late, and a "drop" that ends today is usually a report that hasn't caught up.
The Pages report tells you what Google refused to keep
Under Indexing, the Pages report splits every URL Google knows about into indexed and not indexed, with a reason for each exclusion. Some reasons are fine: Alternate page with proper canonical tag, Page with redirect and Excluded by 'noindex' tag usually describe things you did on purpose. Two others matter.
Crawled - currently not indexed means Google fetched the page and chose not to keep it, usually because it looked thin or too similar to another page. Discovered - currently not indexed means Google knows the URL exists but hasn't fetched it yet, which points to weak internal links or a site Google crawls sparingly. Scan either list for money pages and fix the cause, not the symptom; the fixes live in Sitemaps, robots.txt and indexing.
For a single URL, URL Inspection at the top of Search Console is the ground truth: whether the page is on Google, when it was last crawled, which canonical Google chose (not always the one you declared), and a Test live URL button that fetches the page as Googlebot sees it now. Request indexing queues a crawl for a page you've just fixed; it doesn't make a weak page rank.
The reports you check monthly, not daily
Sitemaps shows whether your sitemap was read and how many URLs it found; errors there mean Google may never see new pages. Core Web Vitals under Experience reports speed from real Chrome users and is covered in Core Web Vitals explained. Links lists who links to you and your most-linked internal pages, a quick test of whether internal links favour your money pages.
Manual actions and Security issues should both say no issues detected. They're the only reports that can explain a cliff-edge drop overnight, so check them first when traffic falls off a shelf.
What to take away
- Search Console is Google's own record of your site, but every number in it is only honest when read next to the others.
- Average position moves whenever your mix of queries changes, so judge rankings with impressions beside them.
- The Pages report and URL Inspection show which pages Google refused to keep and why, which is where most "why don't we rank" questions end.
- Build the three monthly views and read them with the latest few days excluded.
Next
Search Console stops at the click; GA4 for SEO picks up from the landing page and shows whether those visits do anything.