How to check Google results from another country now that country domains redirect

What happened to google.com.mx, google.com.co and the other country domains

If you landed here it is because you tried to check Google results from another country the way you always have — typing the national domain into the address bar — and the browser left you on google.com. It is not your mistake and it is not your ISP blocking anything. In 2025 Google started redirecting the country domains for search to google.com, and by now that behavior is generalized: type google.com.mx, google.com.co, google.com.ar or google.cl and you end up in the same place.

It is worth being precise with the vocabulary, because imprecision is what causes the panic. Plenty of people search for this as google.com.mx removed, and that word is misleading: the domain was not removed. It is still registered, it still resolves in DNS and it still answers requests. What it does is return a redirect to google.com. You can verify that in ten seconds from your terminal or from the browser inspector: the request goes out, gets a redirect response, and lands on the global domain. Nothing died; it simply stopped taking you to a different search.

And there is the key to this entire article: what you lost was not a website, it was a measurement signal. For years, the SEO industry used the country domain as a mental switch. «Let me see how this looks in Mexico» meant «let me open google.com.mx». That gesture gave you a sense of control: I pick the country, Google shows it to me. When the gesture disappears, the sense of control disappears with it, and a question surfaces that most people had never asked themselves seriously: how do I actually know what position I hold in another country?

That question is not news, it is method. Which is why this post is not written like a news item. The news runs out in two paragraphs. The measurement problem lasts years, and it is the one that affects you if you sell in more than one market, if you run clients from several countries out of one agency, or if you simply want to know whether your Spanish-language content competes as well in Bogotá as it does in Mexico City. Everything you can monitor continuously is summarized in the Rankiamos feature set; what follows is the reasoning for deciding what to monitor and why.

Why typing the country domain stopped working (in fact, back in 2017)

This is the uncomfortable part. The trick did not break in 2025: it broke in 2017. That year Google decoupled the search domain from the country of the results. From then on, the country version you get served is not decided by the extension you type but by your location. Whether you typed google.com or google.com.mx, if you were sitting in Bogotá you got the Colombian version. The domain extension became cosmetic.

In other words: between 2017 and 2025 a lot of people were performing a gesture that changed nothing and drawing conclusions from it. They typed the Mexican domain from an office in Medellín, saw the Colombian SERP and wrote down «this is how it looks in Mexico». It is a silent failure, the worst kind: it produces no warning message, it produces a number that looks fine. What Google did in 2025 was take down the façade. By redirecting the country domain, it made visible a disconnect that had already been there for eight years.

You can verify it yourself right now without installing anything. Open a new window, type the country domain of a market you are not in, and search for any term with local intent — something like «restaurants near me» or «employment lawyer». You will land on google.com and you will see businesses from your own city. That is the whole experiment. If your workflow depended on that gesture, you now know how solid it was.

What actually decides the results you see today

With the domain out of the equation, three factors are left in charge, and ranking them helps explain why the manual methods fail:

Notice that none of the three is the domain. And notice that all three live in your browser, not in the query. That is the structural problem with manual checking: you are not measuring a country's SERP, you are measuring a country's SERP filtered through your own session. How much that distorts the position you see is a separate topic — the one about being signed in and personalization — and it deserves its own analysis. Here we stay at country level.

Google's gl parameter by hand: what it actually does and why it is not enough

When word got around about the redirect, the first replacement doing the rounds in forums and on LinkedIn was pasting Google's gl parameter onto the end of the search URL: something like google.com/search?q=shoes&gl=mx&hl=es. It is worth explaining what each piece does, because using those two parameters correctly really is the foundation of any serious country-level measurement. What is not correct is the context you are pasting them into.

The gl parameter (for geolocation) declares the country the query runs against, using the two-letter code: mx, co, ar, cl, pe, es. The hl parameter (for host language) declares the language. They are the pair that Google's search data APIs and professional SERP providers use to say «give me Mexican search in Spanish». In that context they work, and they are the industry standard.

In your browser it is a different story. Web search never documented gl as a public user-facing feature, and Google layers your real IP, your account region and your saved preferences on top of it. The practical result is half-working behavior: sometimes you see a clear change, sometimes you see the same SERP you always get, and sometimes two tabs open on the same URL return two different things. That inconsistency is not a bug you can route around; it is the symptom of using a parameter outside the place where it has a contract.

There is also a background trend worth keeping in mind: Google has spent years closing off manual parameter shortcuts. The most talked-about case was the parameter that forced a hundred results per page, which stopped responding consistently in September 2025 and threw the entire SEO tooling industry into disarray for weeks. Building your measurement process on an undocumented parameter means building on something that can vanish on a Tuesday with no notice.

The bigger problem is not whether gl works

Let us be generous and assume the parameter worked perfectly in your browser. You would still have four problems no parameter solves, and they are the ones that actually matter when the goal is to measure rather than to browse:

• It is not reproducible. Today's measurement and tomorrow's come out of different sessions, with different cookies and possibly from a different IP. Two numbers that do not share conditions do not form a series.
• Nothing is stored. Close the tab and the data is gone. When the client asks a month from now «what position were we in Mexico in August?», you will have nothing to answer with.
• It does not scale. One query at a time is fine for three keywords. With forty keywords across three markets that is a hundred and twenty manual searches every time you want a data point, and nobody sustains that for two weeks.
• It is not defensible. A screenshot from your browser proves nothing. Anyone can argue you were signed in, had a different region configured, or carried a history that flattered you.

The conclusion is not that gl is wrong. It is that gl belongs on the server side, not the browser side. A rank tracker runs the query against a SERP data provider declaring country and language explicitly, with no session, no cookies and none of your history on top. That is what turns a glance into a measurement.

Searching Google from another country with a VPN: what it fixes and what it breaks

The second answer going around is the VPN, and credit where it is due: it is the only manual option that attacks the right factor. If country is decided by your inferred location, changing the IP is exactly the correct lever. Connect to a node in Mexico, run a search, and Google infers you are in Mexico and serves you something fairly close to that SERP. To search Google from another country once, qualitatively, a VPN is the best thing available without paying for tools.

The problem shows up when you try to turn that into a process. Here is the honest balance sheet, without selling anything yet:

The exit city deserves its own paragraph because it is the point most people skip. A «Mexico» node may physically sit in Mexico City, in Querétaro or in Monterrey, and that difference is not cosmetic: for queries with local intent, results change from one city to another inside the same country. If your measurement depends on which server you happened to exit through today, your historical series is noise. City-level geolocation has its own rules and its own impact on quota, and we cover those in the guide on tracking rankings across multiple cities. What matters here is simply that a VPN takes control of that variable away from you.

The second blind spot is the session. If you run the search while signed in, what you see carries your personalization, and the VPN does not touch that. It is a topic that deserves its own article and we will not develop it here; the operational warning is enough: changing the IP is not the same as seeing a clean SERP.

So, the practical verdict: use a VPN when the question is qualitative and one-off — «what does this SERP look like in Chile?», «is this competitor showing up over there?». Do not use it when the question is «did we go up or down in Mexico this month?», because answering that needs two numbers taken under the same conditions, and a VPN, by design, cannot guarantee them.

Measuring Google rankings by country seriously means accepting one simple idea: your browser should not be the thing running the query. It should be run by a service that declares country and language explicitly, that carries none of your session, that runs at the same time every day and that stores the result. That is the entire mindset shift. Everything else is implementation.

Before the implementation, here is the comparison of the five methods available today, judged on the criteria that matter when someone is going to hold you accountable:

The six steps of the process

1. Step 1 · Pick markets with data, not with intuition

   Open Search Console, go to Performance and switch to the Countries tab. There you have clicks, impressions and average position for every market where you already exist. Sort by impressions and keep the top three or four that also make commercial sense. If a country gives you impressions but you do not sell there, do not measure it: you will burn quota and attention on a number you cannot act on. This step is free and it usually dismantles assumptions — it is common to discover the market you thought was secondary delivers more impressions than the main one.

2. Step 2 · Accept that one keyword per market is not the same keyword

   Before configuring anything, review the vocabulary. «Renta de autos» in Mexico is «alquiler de autos» in Argentina and «arriendo de carros» in Colombia. Measure the Mexican term in Colombia and you get a position that is technically correct and commercially irrelevant, because almost nobody searches that way there. No tool automates this step, and it is the single most frequent explanation for an international report that does not line up with sales.

3. Step 3 · Declare country and language explicitly

   Every keyword is configured against a specific market: gl for the country in question and hl for the language. A rank tracker asks you for this when you create the keyword and honors it on every later check, which is exactly what the browser cannot guarantee. Rankiamos uses DataForSEO as its SERP data provider; on Pro you can also measure Bing, where the same keyword on Google and on Bing counts as two units.

4. Step 4 · Always measure in the same time window

   SERPs move during the day. Comparing a 9 a.m. measurement with an 8 p.m. one introduces a variable you do not control, and it will slip into the report looking like a ranking change. A scheduled automatic check solves this without you having to think about it. On Rankiamos the Pro plan checks daily and the Free plan once a week; on both you can re-measure a keyword by hand 6 hours after its last check when you need a clean cut (30 units a day on Pro, 13 on Free).

5. Step 5 · Start recording before you need the data

   This is worth saying bluntly because it hits this use case directly: your own history starts the day you add the keyword, it is not retroactive. If you are launching in Mexico next month, add that market's keywords today. With no baseline, three months from now you will not be able to prove you improved; you will only be able to claim it. Search Console does store the past, but only where you already had impressions and with the limits we covered.

6. Step 6 · Report per market, never as an average

   An average of your position across Mexico, Colombia and Spain is a number that describes no reality at all. If you are fourth in Colombia and thirtieth in Mexico, the average says seventeen and helps you decide nothing. One line per market, always. It is more awkward to present and it is the only way the report survives a hard question.

One nuance about Search Console worth being clear on before mixing the two sources: the position it reports is a weighted average of every impression in the period, not the position at a specific moment, which is why it almost never matches what you see in the SERP. It is a distinction that confuses a lot of people, and we explain it in detail in the article on Search Console average position versus real position. The two sources answer different questions and complement each other: Search Console tells you where you already exist, the tracker tells you where you stand today.

One more variable crosses with country and should not be confused with it: measurement on mobile and on desktop can differ considerably inside the same market, and that has its own analysis in the comparison of mobile versus desktop rankings. What matters here for your budget is that in Rankiamos each device consumes its own quota unit: measuring on mobile and on desktop is two units, not one (Free measures one device of your choice). The same goes for each additional location: it also consumes quota on its own.

International SEO in LATAM: 17 Spanish-speaking countries that are not one market

The most expensive mistake in international SEO across LATAM is not technical, it is a mental map problem. Nearly every tool and nearly all the literature treat «Spanish» as if it were a market, the same way they treat «English» as if the United States and the United Kingdom were the same thing. They are not, and in Spanish the dispersion is wider: we are talking about 17 Spanish-speaking LATAM countries, plus Spain, plus the US Hispanic market, each with its own SERP, its own competitors and its own vocabulary.

The practical consequence is that the same purchase intent is written differently depending on the country, and Google treats those variants as different queries. This is not a style detail: it is the difference between measuring the keyword that brings you customers and measuring one nobody types in that market.

With that table in hand, the question «what Google position do I hold in LATAM?» turns out to be badly framed. There is no such thing as a position in LATAM. There is a position in Mexico for the Mexican term, one in Colombia for the Colombian term, and so on. Any tool that hands you a single number for the whole region is averaging realities that cannot be averaged.

The three blocks and why they are measured separately

Spanish-speaking LATAM. Seventeen countries with wildly different competition levels. Mexico and Colombia concentrate most of the region's Spanish-language search volume, while markets like Uruguay, Paraguay or Bolivia have far less contested SERPs: the same keyword that leaves you outside the top 20 in Mexico can put you in the top 5 in a small market with identical content. That asymmetry is a concrete opportunity for new sites with little authority, and you only see it if you measure country by country.

Spain. It is covered, and it comes with a warning: it shares the language but not the market. Different vocabulary, different competitors, and a media and directory ecosystem that barely shows up in LATAM SERPs. If you sell on both sides, that is two separate configurations, not one with an accent. And if your content is written in neutral LATAM Spanish, in Spain you will be competing against sites written in the local variant, which shows up in CTR even when the position is good.

The US Hispanic market. This is the block most people ignore and the one that delivers the most surprises. These are Spanish-language searches made from US territory, with a SERP that is neither the Mexican one nor the English-language American one. A first-party data point as an example: in Rankiamos's own Search Console, the United States delivers more impressions than Colombia over the last 90 days — for a product built for LATAM. If your business touches Hispanic audiences in Miami, Houston, Los Angeles or New York, that market gets measured separately or it does not get measured at all — and keep in mind that Rankiamos measures the United States in English today, so that Spanish-language SERP is not something it tracks.

A final warning about ambition: do not measure all 17 countries just because you can. Every market you add consumes quota, takes up space in the report and demands interpretation. Two or three well-measured markets, with the right keyword in each and a clean historical series, are worth more than 17 columns nobody looks at. Start with the ones already giving you impressions in Search Console and grow from there.