Everything in this guide happens on five properties I own: this blog, a second blog, my agency site, a small Node app on a VPS and a static site on Cloudflare Pages. None of them gets much traffic, which turns out to be a good condition for learning which parts of Search Console matter, because there is no volume to hide the mistakes behind.
I open a handful of its reports in a normal month. The sections below are the ones that have either cost me something or saved me something on my own domains. Everything else I look up when a specific thing breaks.
Pick the domain property, and put the record where the DNS actually lives
Search Console offers two kinds of property and the choice matters more than it looks. A URL-prefix property covers exactly the prefix you typed, so https://example.com/ and https://www.example.com/ are two different properties with two different sets of data. A domain property covers the whole domain: both protocols, with and without www, and every subdomain. Where I control the DNS, I now add a domain property, because one verification covers every variant and there is no second property quietly collecting the data you are looking for.
A domain property is verified with a TXT record in the DNS zone. The part that trips people up is where that zone lives, which is not necessarily where you bought the domain. On my domains whose DNS is served by Hostinger, the record goes in the hPanel zone editor. On a domain registered at DonDominio or IONOS but pointed elsewhere, the record has to go wherever the nameservers currently point, not in the registrar’s panel. For the site I host on Cloudflare Pages, that means Cloudflare’s DNS tab. If you add the record in the wrong panel, verification simply never completes and nothing tells you why. I go through registrar panels and nameservers in more detail in the guide to pointing a domain at your host without downtime.
One thing nobody mentions at setup: the record is not a one-off. Google re-checks it, so if you tidy up your DNS zone a year later and delete a TXT record you no longer recognise, you lose access to the property until you verify it again.
The sitemap box is a URL field, and it means it
This is the smallest, stupidest thing I get wrong most often. The Sitemaps screen shows your domain and an empty box, which reads like it wants a filename. It does not. Paste the complete address, https://example.com/wp-sitemap.xml, protocol and all. Type sitemap.xml on its own and you get an error rather than a helpful correction, and on a bad day you will conclude your sitemap is broken when the only broken thing was the box.
If you run WordPress you probably already have a sitemap whether you installed an SEO plugin or not: since version 5.5 core generates one at /wp-sitemap.xml. Check which file your setup is actually serving before you submit anything, because a plugin may be generating a second one at a different path. Google’s own sitemap documentation is short and worth ten minutes.
Re-submitting the same URL is also how you ask for a re-read after you have changed something substantial. It is not a magic button, but it is the only lever the interface gives you, and it costs nothing.
The two sitemaps on my own domain that described a site that no longer existed
This domain ran a static export before it ran WordPress. When I checked it properly on 19 September 2026, the old build had left two files behind that still returned 200: a /sitemap.xml index and the /sitemap-0.xml it pointed to, listing 27 URLs from the previous site. My robots.txt was pointing search engines at that index rather than at the WordPress one. So the file I was telling Google to read described a site I had replaced two months earlier.
Fixing that is two jobs, not one. In Search Console you remove the dead sitemaps from the Sitemaps report so they stop being reported on. On the server you make the files genuinely stop existing, which means a redirect or a 410 rather than leaving them returning 200 to anyone who asks. Removing the entry in the report while the file is still live fixes the dashboard and nothing else. The full account of that cleanup, including which URLs got a 301 and which got a 410, is in moving a site to WordPress without leaving ghosts behind.
Worth being clear about the limit: taking a sitemap out of the report does not remove a single URL from Google’s index. It only stops you feeding it a list you no longer stand behind.
Search Console only speaks punycode
This blog lives on a domain with an ñ in it. Browsers show the accented spelling; the DNS system and most tooling use the ASCII form, which for me is xn--reseasdehostingpro-q0b.com. Search Console works in that ASCII form. When I add the property, when I paste a URL into inspection, when I read the host name back in a report, that is the spelling I see, and the first time it happens it looks like somebody else’s domain.
It is not a fault, and nothing is broken, but it does mean that a copy-pasted accented URL sometimes will not match and you get a puzzling “URL is not on this property” message. If you are considering a domain with accents or non-Latin characters, I wrote up everything that got more complicated in buying a domain: the trap nobody warns you about.
URL inspection, and what requesting indexing really does
If I had to keep one screen it would be this one. Paste a URL and it tells you whether Google has it, when it last crawled it, which canonical Google chose, and whether the live page differs from what it has stored. That canonical field is the single most useful thing in the product, because the canonical Google picks is not always the one your theme declares, and you will never discover that by looking at your own site.
Then there is the button. Requesting indexing puts one URL into a queue, with a daily quota, and gets it looked at sooner than it might otherwise have been. Google’s own page on asking for a recrawl is careful about this and so am I: it is a request, not an instruction, and being crawled is not the same as being indexed, which in turn is not the same as ranking for anything.
An older post on this site used to promise you could compress the timeline for a new domain with the right tactics. I have deleted it and redirected it here, because I cannot support that claim with anything I have measured. What I can tell you is the narrow, true version: requesting indexing is how you get a specific page looked at after you publish or change it, and it is worth doing for the handful of URLs that matter. It is not a growth strategy.
Five clicks in three months, and what the performance report is still good for
Here is the number I would rather not publish. When I audited this blog on 19 September 2026, its performance report showed five clicks across the previous three months. Five. That is not a trend, a baseline or a benchmark; it is noise with a chart drawn through it, and any post that used those numbers to teach you about click-through rates would be lying to you with real data.
So what do I use it for on a site that small? One thing: the query list. Even with almost no clicks, the impressions show you which words Google has decided your pages are about, and that is frequently not what you thought you wrote. The gap between the two is the most useful signal a small site gets. Average position, on numbers like these, moves for reasons that have nothing to do with you, and I try not to build anything on it. I go through which free tools earn their place at this scale in the free SEO tools I actually open.
Users and permissions: the screen nobody opens
Settings, then Users and permissions. It lists everyone with access to the property and what kind: owners, full users, restricted users. A verified owner is not simply an entry on a list. Ownership is tied to a verification method, so an owner who has their own token in your DNS, their own file on your server or their own tag in your template stays verified even if you delete their row. To actually remove them you have to remove the verification method too, and the interface does not make that obvious.
I check that list on every property I touch now, and I check the verification methods underneath it. The one time this genuinely mattered to me was on a site that is not mine to write about, so there is no war story here, only the habit that came out of it. If you have ever handed a site to an agency, a developer or a plugin that offered to “connect” your Search Console for you, spend the minute. The rest of the access review I run on my own installs is in hardening WordPress after finding my own site wide open.
Four things it will not do, whatever the tutorials say
- It is not a ranking factor. Verifying a property does not make Google like your site; it makes Google show you what it already thinks.
- It does not give you a crawler you control. You can ask; the queue is not yours.
- It shows you nothing about anyone else. No competitor queries, no competitor backlinks, no market data. That gap is exactly what the paid suites sell.
- It cannot manufacture data that does not exist. On a site with no traffic, most of the reports are empty, and an empty report is not a problem to be fixed.
A new property, start to finish
- Add a domain property and verify it with a TXT record in whichever panel serves your DNS today.
- Find out which sitemap your site actually serves, open it in a browser, and check it lists the pages you expect.
- Submit that sitemap with its full https address, and delete any older sitemaps left over from a previous build.
- Make sure
robots.txtpoints at the same file and is not blocking anything you want crawled. - Inspect three URLs by hand: the homepage, one post, and one page you have just changed. Read the canonical Google reports back to you.
- Open Users and permissions once, read the list, and close it again knowing who is in there.
That is an hour of work, most of it waiting for DNS. After that, the honest cadence on a site my size is a short look once a week: look at the query list, inspect anything I have changed, and resist the urge to interpret five clicks. Search Console is free, it is the only tool that reports what Google actually did rather than what a third-party crawler guesses, and it is at its most useful when you stop expecting it to be a growth dashboard.