Is Google ‘Stealing’ Your International Search Traffic With Translations? Here’s What We Found

Is Google translating our content without sending people to our site? Here’s what’s really happening and how we can protect our traffic.

Google does show machine-translated pages sometimes. But this mostly happens when local content is missing or not good enough. The big worry is that people get answers without visiting our site. But from what we have tested, it depends on which browser and device they use—and where they are.

In many cases, especially with Chrome, users still visit our website. Google just translates the page inside the browser. So, we don’t lose traffic or credit. But with Safari or Firefox, it is different. Google may use a proxy link (translate.goog) that keeps users inside Google’s system, not on our website.

So, we need to look at this problem in two parts:

• What users see in search results.
• What happens when they click the result.

Without clear tests across browsers and countries, we can’t say for sure that this is a big problem for everyone. Still, we can take smart steps now.

First, we should find out which of our pages show translated URLs.

Then, check which browsers our visitors use.

This will give us better information and help us decide where to focus our localization work.

In short, this problem is real—but it is not the same for everyone.

If we understand the full picture, we can act smartly instead of just worrying.

What is the article saying?

It says Google is showing translated versions of English pages right in search results. These pages are on a Google proxy domain, not the publisher’s real site.

So, what happens?

• Publishers lose traffic and data about how users behave.
• Google controls what users see and how it looks.
• This makes many publishers feel like Google is pushing them out.

What did Google actually say?

The important part is:
“This only happens when our systems decide that there is no good, local-language content for the user’s search.” This means Google does this only when there is no better content in that language.

What is Google doing with translation proxies?

Google now uses a system that shows translated pages when a user searches in a language with no strong local content. This is not the same as when we use Google Translate ourselves. It happens on Google’s servers, and publishers may not even know. The translated page shows a special Google domain that looks like the original site. For example: www-searchengineland-com.translate.goog It may say “Translated by Google.” But the user never visits our real site.

As a result:

• We lose traffic, data, and ad money.
• Users think they are on our site, but they are not.
• We may not even know our content is used this way.

Even though the goal is to help users get answers in their own language, it raises big questions about control, credit, and fairness for content creators.

What did we find?

We set up a test system that acted like a real local user.

We changed browser language settings (in Chrome, Firefox, and Safari), set Google’s search location to the right country, and used VPNs to match local IP addresses.

From this, one thing was clear:
The problem of Google cutting out the publisher is real—but how big it is depends on the browser, device, and user settings.

In our tests, Chrome users went to the real publisher’s site.

Google’s built-in translation worked on the page, and the publisher still got data, ads, and branding.

But for Safari and Firefox users, it was different.

Google’s translation proxy (translate.goog) was used more often. When users land on these proxy pages, even links inside keep them inside Google’s system. That means users may look at many pages without ever visiting the real website. This causes big losses—less traffic, no user data, no ads, and no way to bring users back later.

This shows a big truth:
The browser a person uses is not just a design choice—it can affect business results. We now have to look at traffic by browser to see where we may be losing visitors.

Simple averages or headlines can be wrong. The real effect changes a lot based on the user’s browser, device, and location.

Action 1: Translated search results in both AI and normal search

We found many search results where Google translated English content into local languages—both in normal and AI search results.

Our test was to check if Google translated content when no local page was found.

We did not compare the quality of local vs. English content.

We tested with searches in Spanish, Indonesian, Thai, and Turkish.

Example: Spanish search in Mexico for local art exhibits

We thought there would be little local content, so Google would translate.

Chrome experience: Google showed many AI answers that were clearly translated.

When we hovered over the link, the URL was from the real publisher—not Google Translate.

Safari experience
One AI result was shown and translated.

Hovering over the link showed a Google Translate URL, not the real publisher.

The URL had settings to translate and open a new tab.

• hl=es means Google shows Spanish based on browser settings.
• sl=en means source content is in English.
• tl=es means target content is Spanish.

client=sge means the result came from Search Generative Experience (SGE).

Firefox experience
We saw AI results and two translated sections. The link showed a Google Translate URL with similar settings.

Unlike Safari, it did not ask to open a new tab.

Action 2: What happens when users click the result?

We saw different results depending on the browser.

Chrome experience

Clicking the link sent us to the real publisher’s site. Chrome’s built-in translation started working right away.

We could still use Google Translate toggle on the page.

In every test, clicking in Chrome sent us to the real page, not translate.goog.

So, Chrome did not block traffic to the publisher.

Safari experience

Clicking the link opened a new tab using Google Translate.

The address bar showed the proxy link:
This used Google’s translation tool on a different domain, not the real site.

Every time we clicked inside, it stayed in the translation system.

Firefox experience

Firefox worked like Safari.

Clicking the link opened a new tab with Google Translate. The URL showed a Google proxy page, not the real website.

Publisher Recommendations

We should check how much traffic comes from Safari, Firefox, and old Chrome versions that don’t use auto-translation when content doesn’t match user language.

If local content is missing or low quality, we may have chances to reach users in countries where good local content is not available.

Interesting finding:

We had trouble repeating Erik’s search using the Spanish word for [respiratory system].

We were surprised because we thought Google would not need translated content in Spain—and we were right.

In our tests, Google showed only local Spanish sites in AI Overviews and top results.

Then we saw Erik was using a SERP simulator set to ES-419—that means Spanish for Latin America, not Spain.

When we changed our VPN to places like Bolivia, Peru, and the Dominican Republic, we saw the same translated results Erik saw.

But in Mexico or Argentina, we did not see it.

It seems these countries have better local Spanish content, so Google did not need to translate.

This shows Google’s translation tool depends not only on language but also on local content quality by country.

Why is Google using its own domain (translate.goog) instead of the real publisher’s domain?

Google likely uses its own domain to have better control, smoother performance, and fewer legal problems.

Using translate.goog lets Google:

• Fully control how translated pages look and work.
• Create a clear legal gap between Google’s page and the publisher’s original page.
• Give a steady user experience across all devices, browsers, and sites—without needing help from the publisher’s code.

As we said before, this method isn’t perfect for us as publishers. But for Google, it is a system that works on a large scale while meeting legal and tech needs.

A small joke – but maybe not:
What if Google made a paid service called Translate.goog Pro™? They could give publishers access to traffic, data, and user stats—for a fee. If users already stay on Google’s platform, maybe we’ll have to buy our own data, like a paid software service.

Is this product made for ease and safety – or is it a mistake?

It reminds us of an earlier problem with AI Mode. Google had added something called rel=”noopener” to outgoing links by mistake, which hurt tracking. That bug was later fixed. But here, Google’s proxy setup seems like a planned move. It aims to make translation fast, safe, and easy for users. But it doesn’t help us as publishers—we lose control, traffic, and important data.

This isn’t just a small issue—it’s a deeper design choice that might continue unless Google is pushed to change or offer new tools.

So, who benefits more: the users or Google?

It’s easy to say Google wants more control and profit. But it’s more complicated than that.

Yes, Google keeps users on its system and collects useful data. But it also helps people read content in their own language—quickly and without extra steps.

So the translation feature helps both Google and users. The real problem is us, the publishers.

We make the content, but we may lose the credit, the data, and the visitors. So, while the proxy isn’t “stealing,” it is a system that doesn’t help everyone equally.

Still, users do get a lot:

• They get content in their language.
• They don’t need extra tools or settings.
• They can read content that might not be available to them otherwise.

In the end, this isn’t just about who wins—Google or users. It’s about how value and credit are shared as search tools change and grow. We understand why some publishers feel left out.

But we also have to ask—was that traffic really ours if no local version of the content existed?

That said, Google could have worked more with publishers. They could have helped us keep credit and made a fairer system.

Who wins and who loses?

Every time Google changes something, some effects are small and others are big. The start of auto-translated proxy pages is no different.

Even though Google says it helps users without content in their language, this change shifts who gains the most, who can adjust fast, and who gets left behind.

To see the full picture, we need to ask:

• Who really benefits?
• Who might lose?
• Who still has time to make changes?

What do publishers lose?

While users benefit and Google grows its system, we as publishers face losses. This is not a small problem—it affects money, site data, and future plans.

Let’s break it down:

• Traffic: People stay on Google and don’t visit our site.
• User Data: We don’t get info like time spent or actions taken.
• Ads and Leads: No visits mean no ads shown or leads captured.
• Personalization: We can’t retarget or show content that fits users’ interests.

These losses are real—but how bad they are depends on the content type. For example, an info article may be okay without clicks, but a product or sales page needs users to visit and act.

Our challenge now is to know which traffic really matters—and which is just about being visible. More importantly, we should have a say in how our content is used in this new system.

Didn’t Google already show signs of this with MUM?

In 2021, Google introduced MUM (Multitask Unified Model). One main feature was that it could understand content in many languages.

Google said: “It can learn from sources not written in the language you searched in.”

At that time, most SEO people thought it meant better understanding of questions—not changing how content is shown. But some, like Olaf Kopp, guessed where this might lead. He hoped Google would still reward the original creators of good content. As international SEO experts, we believed MUM would bring more changes. Now we see it happening in AI Mode with Gemini:

• Cross-language answers.
• Translated featured snippets.
• And now, full proxy-translated search results.

So, proxy translation is not new—it is the next step in MUM’s plan.

What could Google have done instead?

Before saying this system is bad, let’s think about other ways Google could fill the content gap—and the good and bad sides of each.

Option 1: Show the best local result, even if it’s not perfect

This means showing content in the user’s language, even if it’s less detailed.

Pros:

• Supports local websites.
• Pushes publishers to make better content for their region.
• No need for machine translation or Google proxies.

Cons:

• The answer may be less helpful.
• Users might get frustrated if the info is weak or incomplete.

This fits Google’s old idea: show local results first. But when good local content is missing, Google fills the gap. That may feel like Google is cutting us out—but it also shows local content needs more work.

Option 2: Show a translated snippet with a link to the real page

Here, Google shows a short translated preview in search results and links to the original page.

Pros:

• Users can still visit the publisher’s site.
• Publishers keep traffic, data, and brand value.

Cons:

• If the preview has all info, users may not click.
• The publisher’s work helps the user but brings no benefit to the publisher.
• User risk: If someone clicks, they might see a page they can’t read.
• Their browser may not translate automatically, causing them to leave quickly and hurt the publisher’s image.

Option 3: Show content only in the original language (like English)

This means Google shows content only as it is, usually in English. It avoids translation but is hard for people who don’t understand English well.

Problems:

• Many users can’t read or understand the content.
• The experience is bad for users in countries where English is not common.

All options have trade-offs. Showing local content helps the local web but may give weak answers. Linking to the original page gives credit but might confuse users. Showing only English keeps quality but blocks many users.

Google’s proxy translation isn’t perfect, but it is a practical way to fill the gap when local content is missing—not a trick to cheat publishers.

Option 4 (Mix method): Local preview + user clicks to translate page

Instead of fully translating the page on Google’s proxy, Google could try a middle way:

• Show a short translated snippet in search results in the local language.
• Link directly to the real page on the publisher’s site—not through a proxy.
• Give users a clear “Translate this page” button if their browser doesn’t do it automatically.

Possible benefits:

• Publishers still get visits, ad money, and user data.
• Users see useful content in search results.
• It builds trust and lets users choose.

Challenges:

• Google needs to change how results look a bit.
• Some users might not use the “Translate” button, so fewer see the full translation.
• It doesn’t fix AI Overviews or snippets that show too much info without clicks.

How Google could test it:

• Test in some countries using data from Google Search Console and Translate use.
• Compare how often users click real sites.
• See how many use the translation toggle.
• Check bounce rates and time spent.
• Get publisher feedback on credit and traffic.

This is not just a middle ground—it’s an upgrade. It keeps the “Translate” button, respects users’ needs, and lets publishers keep credit.

Does content type matter?

Yes—it depends on what users want. Some content isn’t made locally because people don’t search for it much. If someone wants a quick answer like “What is anemia?” or “Best time to visit Kyoto,” they may not click the page—they just read the snippet. But deeper content, like product guides or comparisons, needs real visits. In these cases, being translated without visits hurts the publisher.

So we need to ask: Which pages need local versions, and which do not? Many pages shown through Google’s proxy are basic info pages. These often get very low clicks—less than 2%. In those cases, visibility doesn’t mean real value.

But if trust or sales matter, not getting visits means losing real business. So we should focus on pages where localization truly helps.

What needs to change?

If this feels unfair, it’s because publishers were not part of the process. Google tries to solve a user problem—but without asking us for help.

We don’t want to stop translation. We want it managed better with:

• Clear credit: Always link directly to the original page—even if translated.
• Data access: Let publishers see how their content is used in Google’s translation.
• Opt-in: Publishers can choose to provide local content or help with translations.
• Fair rules: The industry should set clear rules about content use, especially with AI.

This is not stopping new tech—it’s sharing value fairly. If we keep making helpful content, we should also get credit and rewards for it.

Did Google ever help creators localize content?

The article says Google stopped helping creators localize and took the traffic. But did Google really help before? Yes, Google gave advice like using hreflang and country targeting. They also had tools like the Translate widget or API, but those tools were closed or made paid. And they didn’t fully help with translation and localization. Now Google is doing something new.

Instead of showing what we made, They create translated versions on their own site. This is not a broken promise—it is a new way of working, and it changes how we see our role in search.

Is Google being unfair about auto-translated content?

At first, it seems Google breaks its own rules. For years, Google told site owners not to use automatic translation without checking. They said it was low-quality and could hurt rankings. But now Google is doing the same— translating content automatically and showing it in search.

without telling us or letting us check the translation. It does feel unfair.

If your content is being translated, is that good?

There is another way to see this. If Google translates our content for other countries, Maybe it means people there want what we have. This gives content teams and international SEOs a big choice.

In May 2022, Google added a “translated results” filter in Search Console.

This helps us:

• See which pages are translated.
• Check impressions, clicks, and CTR.
• Find new markets with growth chance.

In 90 days, daily translated impressions grew from 10,000 to 22,000.

Most growth came from:

• Turkish searches in Turkey (60%).
• Hindi searches in India (30%).

But here is the problem:
97% of these searches were just for information.

CTR was only 2%.

This means we are visible, but users do not click.

In today’s world of AI and zero-click searches, this is a big problem.

So what does that visibility mean?
It’s true Google shows our content across languages. That shows our content is useful and wanted. But if almost no one clicks, and everything stays inside Google’s system, are we really getting any value? Maybe this is a roadmap. These translations could show where we should make local pages. But only if we believe users will click and stay. Otherwise, we are just trying to get back data Google already took.

Is this a loss or a new chance?
Right now, Google uses our content without credit or pay. If this spreads, more sites may lose value to AI summaries and proxy translations. Whether this is a missed chance or misuse of power depends on what happens next—and if Google stays a helpful tool or becomes a gatekeeper.

Where do we go from here?
We are at an important point for global SEO, content plans, and how we work with big platforms like Google. Google’s proxy translation helps users, but it also shows a bigger issue— Some content creators don’t get the value they deserve.

What should publishers do now?
Don’t panic—find out what is happening: If Google auto-translates our content, we should check why. Look at which search terms and countries are involved. Google Search Console’s translated content filter shows patterns. We should check if our pages appear in AI Overviews or proxies.

Check browser use on your site:
From tests, Chrome users often go straight to our site.

But Safari and Firefox users may get stuck in Google’s translation layer.

We need to check traffic by browser and device to know the risk.

Markets with many Safari users may be hit hardest.

Fix the content gap—but do it wisely:
Start by translating pages that bring good traffic or sales.

When Google translates our pages, that shows what is missing.

We shouldn’t see it only as a threat. It is also a warning.

Those who act fast may gain real benefits.

Ask for better links and more control:
We need to ask Google for better credit—like clickable links to our sites.

Google should give tools to choose if we want our content translated this way.

A simple opt-in or opt-out in Search Console would help a lot.

Where things should go next

If we talk to Google about sharing content, AI training, or licenses,

we must include proxy translation in those talks.

Who owns the translated version?

Who gets the traffic, data, or money?

This is not just tech—it is about fairness.

Lawsuits over content use and fair pay may soon include this issue.

Proxy translation could be the next big topic in the fight for fair content rules.

The real issue: gaps, needs, and fair value

This all begins with gaps—gaps in local content, in access to information, and now, in trust.

Google is trying to fix a real problem:
People need information in their own language when local content is missing. This fits with Google’s goal to make all information easy to find and useful. But by doing this, Google is cutting us off from our own readers. When users stay inside Google’s translation layer, we lose traffic, data, and income. It’s not stealing—but we are not being treated fairly either. The internet used to be about sharing value. That balance is now broken.

What we need is not to stop translation— but to create a new system where everyone is included. A system where we, the content creators, help fill the gaps. And where platforms like Google give credit and work with us, not around us. Proxy translation might be the future of search. But it should also start a new conversation— about what a fair, open, and global internet really means.

FAQs: