Why isn't my own Google search a fair check?
Because Google doesn't show everyone the same page. Typing the keyword into Google is most people's answer to how to check my Google ranking, and it's the least reliable one. When an owner tells me their site dropped because it wasn't on page one when they searched this morning, the first thing I do is open the site's Search Console data and read what Google actually served, since a single search from one laptop can't tell you much.
Google's own help pages list the reasons results differ between people. Time is one: new pages get added all day, and ranking changes roll out to data centres gradually. Location, language and device change the page too, and some results are personalised from your Google Account, though Google says not every result is.
Private browsing doesn't fix this. Google says it works out your location from your IP address and your device, and neither of those changes when you open an incognito window. So an incognito search in Edmonton is still an Edmonton search on your own connection. It's a real result, it just isn't the result, and it's one searcher out of thousands.
There's a second problem with checking by hand. It only tells you about the words you thought to type. The queries that actually bring people to a site are often ones the owner has never searched, and you'll only find those in Search Console.
How I check website ranking on Google in Search Console
Plenty of owners have never had a reason to use Google Search Console beyond confirming it's connected. This is where it earns its place. Here's the routine I run on a client's property: after a couple of goes it's about ten minutes' work, and it costs nothing beyond a verified property.
- Step 01
Open the Performance report for Search results
The default view covers the past three months of web search. That's a sensible window for a first look, because it's long enough to smooth out a bad week and short enough to reflect the site as it is now.
- Step 02
Switch on all four metrics
Four metric headings sit above the chart, and clicking one toggles it on or off. I make sure CTR and Average position are on alongside clicks and impressions, because a position with no impressions beside it doesn't mean much, and I want every row to carry all four.
- Step 03
Filter to the query you care about
This is how to check keyword ranking for one search rather than the whole site. I add a query filter with Queries containing and a word from the search, then read the Queries tab. For a family of searches, Custom (regex) takes a list like
(seo audit|technical seo), and Google says the match is partial, so it catches longer phrasings too. - Step 04
Open the Pages tab with the filter still on
This shows which URL Google is ranking for that query. If it's not the page you'd expect, that's often the real finding, and when two of your own pages keep trading places, that's the job keyword cannibalization fixes for WordPress exist for.
- Step 05
Add a country filter
A Canadian business that also gets impressions from other countries can have its average dragged by searches it'll never serve. I filter to Canada, or to the one country the business sells into, before I read a single position.
- Step 06
Compare mobile and desktop
The Devices tab splits the same data by device. Rankings and click rates can differ a lot between the two, and how to improve organic CTR depends on which one is lagging, so I read them separately.
- Step 07
Compare two date ranges
The date filter has a Compare option, and the table then adds a Difference column for every row. I compare the last 28 days with the 28 before, and that comparison is the rank data I carry into a monthly report; what to include in an SEO report covers the rest.
Search Console data usually arrives two to three days late, according to Google. If yesterday looks empty, it isn't a drop. It's just not counted yet.
How can I check my website ranking on Google for free?
Five free ways, and they answer different questions. Search Console is the record of what Google showed for your own site; the rest are samples of one results page. I read what each tool says about itself on its own page, on 28 September 2026, and here's how they compare.
| Free check | What it tells you | Where it falls short |
|---|---|---|
| Search Console Performance report | The positions Google actually showed your pages, by query, page, country and device | Your own verified site only, and the data arrives two to three days late |
| Ad Preview and Diagnosis in Google Ads | The results page for one search term in a location and language you choose | One search at a time, from inside a Google Ads account |
| Ahrefs' free Keyword Rank Checker | Where a site ranks for a keyword, across hundreds of locations | Ahrefs says its full ranking data comes from a keyword database updated monthly, not a live count of searchers |
| SEO.ai's Google website rank checker | Your domain's position for one keyword within the top 100 results | One keyword, one check, one moment |
| The Google rank checker extension for Chrome | Your site's position flagged in the results as you search | Its listing says it fetches the top 100 with a cookie-free request, like incognito, and that its rank may differ from regular results |
I use a checker for the one thing Search Console can't do, which is seeing where a competitor sits, and I read its number as one sample. For your own site, a position with impressions beside it in Search Console beats any of them.
Rather have me do this? One person, whole job.
Book a call on UpworkWhy can average position mislead?
Because it's an average of the topmost spot your site held, across every query and every impression in the view. Google's example is simple: a site at position 2 for one query and position 3 for another averages 2.5. Add a few hundred queries where the site sits on page four, and that one number stops describing any page anyone actually saw.
It shows up on real sites all the time. One practice's site averaged 12.5 across a quarter while its brand sat at 1.1, and several town searches ran between 2.4 and 9.6. That 12.5 wasn't a ranking anybody searched. It was a blend of strong local positions and a long tail of informational reach. The same case shows the device point from the steps above: desktop got far more impressions but clicked at 0.56%, while mobile clicked at 1.94%.
The chart and the table don't aggregate the same way either. Google says data grouped by queries, countries, devices or dates is counted by property, while data grouped by pages is counted by page. So the site's average in the chart and a page's average in the table can both be right and still not match.
| What you see | What it actually is | How I read it |
|---|---|---|
| Average position in the chart | The topmost result from the whole site, averaged across all queries | A trend line only, never a ranking |
| Average position on a query row | The average for that one query over the date range | The closest thing to where you rank, if impressions are high |
| Totals above the table | Every impression, including rare queries | They'll exceed the rows, because tables drop rare queries and stop at 1,000 rows |
| A query with 5 impressions | A handful of searches | Too few to trust the position either way |
One more reason a ranking can look like it's bouncing: two of your own pages taking turns for the same query. If the Pages tab shows two URLs splitting one query's impressions, the free checker I built reads that export for you, and it's a quicker answer than watching the average jump.
How do I check rankings for one city?
Search Console's location dimension stops at country. There's no Calgary filter and no Red Deer filter, so a search like SEO in Red Deer only shows up as a query that contains the town's name. When I do SEO for Alberta businesses, remotely, I use a few free checks together, because none of them is perfect alone.
- i.
Filter queries by the town name
A query filter containing
calgary, or a regex like(calgary|airdrie|cochrane), pulls out every search where someone typed the place. Those positions are from real searches, and they're usually the most commercial queries on a local site. - ii.
Use Google's own preview tool
The Ad Preview and Diagnosis tool lives inside Google Ads. You enter a search term plus a location and language, and Google says you'll see results similar to a normal search without piling ad impressions onto the account's stats. That's how I look at a Lethbridge results page from anywhere.
- iii.
Treat the map pack as a separate ranking
The map results and the ordinary results aren't the same list. A site can sit fourth in the web results and not appear in the map at all, so I check both and report them separately.
- iv.
Read the town in context
When I measured demand for my Alberta town pages, several smaller towns returned no measurable volume for the service at all. So I don't judge a town by one position on a query that nobody types.
When is a paid rank tracker worth it?
For most small sites, it isn't worth it yet. Search Console already tells you the positions your pages held in real searches, which is the thing a tracker estimates. I'd only add one when a question comes up that Search Console can't answer.
The first is competitors: Search Console only shows your own site. The second is a query you've never appeared for, since Search Console has no row for a search without a single impression. The third is scale: dozens of towns checked daily at a fixed location is tedious by hand and cheap for a tracker.
Even then, I treat a tracker's number as a sample and Search Console as the record. When they disagree, Search Console's the one that counted real searchers.
If you've looked at the Performance report and still can't tell whether a drop is real, send me the site. I'll read the query and page data, tell you what moved and why, and say plainly if nothing's wrong.
Book a call on Upwork
