Localizing achieved proper can ship critical worth. Executed mistaken, it turns into a black gap of assets with little to point out for it.
That can assist you get it proper, I turned to somebody who’s been deep within the trenches of localization: Erik Sarissky, Head of Worldwide Advertising and Product Localization at Ahrefs.
Erik has quadrupled natural site visitors to Ahrefs’ Spanish weblog in 18 months, going from 5.4k visits to over 22k visits. Along with worldwide advertising and marketing actions, he oversees localization efforts throughout 14 languages and consistently evaluates new potential languages.
One among his most eye-opening wins? After switching from Continental Portuguese to Brazilian Portuguese, Ahrefs noticed an explosion in site visitors to its native content material and touchdown pages.
So I requested Erik if he’d be keen to share how he did it—and what he’s realized alongside the method.
Whether or not you’re contemplating increasing into neighboring markets or exploring development distant, this information breaks down what really strikes the needle in localization web optimization and what’s simply noise.
Translation is a step. Localization is a system. Let’s break it down.
- Translation is about changing phrases from one language to a different, preserving that means and construction. Consider it like Google Translate, however achieved by a human. It’s correct, however literal.
- Localization adapts examples, references, idioms, visuals, and even the technique to really feel pure within the goal market. It’s the distinction between being a vacationer who memorized a phrasebook and somebody who really grew up there.
- Transcreation takes issues additional. It adapts the tone, emotion, and intent of a message so it hits dwelling in a special tradition. A intelligent English tagline may want utterly completely different wording to really feel intelligent in Spanish. It’s much less about constancy, extra about resonance.
“We already had good content in English,” Erik told me, “but there was no clear strategy. We were just translating everything without considering localization or keyword research. I knew that by fixing this, traffic would grow.”
And it did. That’s why a translation-only approach often falls flat. It misses the nuance, and with it, the traffic.
We have to begin with a wake-up name: localization is not non-compulsory, as a result of Google may do it for you.
Since early 2025, Google has been more and more translating English content material into native languages and serving it below a Google-owned subdomain (e.g., translate.google.com). This proxy translation exhibits up in AI Overviews, Featured Snippets, and even the common 10 blue hyperlinks.
The consequence? Customers by no means land in your web site. As an alternative, they view a Google-hosted model of your content material, and any clicks or hyperlink fairness are attributed to Google, not you.
“When you don’t localize your content material, Google may do it for you—and take the site visitors.”
If you wish to see if Google is auto-translating your web site pages, you should use Search Console and Ahrefs to seek out out:
In Google Search Console, go to Search Appearance → Translated Pages.
In Ahrefs’ Site Explorer, open the top pages report for translate.google.com and apply a “URL contains” filter, entering your domain name.
The fix? Localize your content—even minimally. A 300-word native-language version can often be enough to replace Google’s proxy page in the SERPs.
Attempting to localize for everybody directly is a recipe for burnout. As an alternative, do what Erik did: select markets primarily based on metrics.
“We base selections on billing, interface language use, model searches, GDP, and tech adoption. Indonesia stunned us—it’s the most effective by way of billing.”
Spanish was the pilot. We’ve been fairly lively in Spanish-speaking international locations, organizing all types of meetups and participating in conferences in Spain and Latin America.
Then, Japan as a result of we have been already fairly standard there (extra standard than our opponents, if I’m not mistaken). Indonesia emerged as an surprising champion, so it’s going to be subsequent in line. While you let knowledge lead, you make investments smarter and develop sooner.
Take a look at historic demand and potential future upside. Markets with fast-growing tech ecosystems or altering search conduct (like South Korea’s transfer from Naver to Google) could also be extra ripe than they appear.
That brings us to native key phrase analysis.
What works in English received’t all the time work in Spanish, Japanese, or German. Localized web optimization begins with native intent.
“We test site visitors potential, enterprise potential, and assign priorities. Then we resolve if it’s price adapting or writing from scratch.”
Right here’s what Erik’s referring to.
If you skip this step, you risk optimizing for keywords no one searches—or worse, competing on terms you can’t win.
To make things easier, you can start with your native language, add keywords to a list, and let AI show relevant translations with SEO metrics right inside the tool.
Here’s a sample result that instantly dispels doubts about the correct translation.
Next, Erik uses Ahrefs’ business potential framework to prioritize keywords. We score every marketing idea from 0 to 3 based on how essential our product is to the problem.
“When you’re simply translating the whole lot, you’re losing time,” Erisk advised me bluntly. The smarter strategy is to decide on what deserves translation and determine when creating from scratch is a greater funding.
For instance, Erik’s group begins by translating knowledge research instantly, as that is essentially the most environment friendly technique. They then create supplementary micro-studies inside these translated items so as to add native relevance.
“We created a workflow to assessment which articles to translate and when to jot down new ones. Typically translations don’t make an excessive amount of sense—it’s higher to jot down 100% new articles primarily based on native key phrase analysis.”
Earlier than you even begin localizing, step again and map your objectives. Contemplate content material age, market relevance, and search demand. Prioritize items that can carry out, not simply those you’ve just lately printed.
A very good place to begin would to take a look at the site visitors and hyperlinks a bit of content material will get. Simply plug within the URL of the web page in Ahrefs’ Site Explorer and look at these numbers in the Overview report.
To see which pages perform the best, head on to the Top pages (you can additionally filter by country) and Best by link report.
In case your localized content material reads like a careless translation, it received’t join. To actually resonate, it must really feel prefer it was written by somebody who lives in that market.
“After I took over the Spanish weblog, we began utilizing extra examples from the Spanish market and created our personal case research.”
So, the recommendation right here is: use native case research, adapt examples to suit regional norms, and reference the platforms your viewers really makes use of.
Right here’s Erik displaying the outcomes of localizing a information to affiliate internet marketing by utilizing native examples:
And that is an instance of how they created a completely new article about advertising and marketing guide suggestions by asking Spanish audio system for his or her favorites, quite than merely translating the prevailing English model.
The outcomes:
Erik’s group just lately began utilizing this strategy in Japan as properly. Right here’s an instance of a case examine they created that includes a Japanese company’s success with Ahrefs.
However content material relevance isn’t nearly utilizing native examples; it’s additionally about voice and opinion. Based on Erik:
“In the future this shall be absolutely changed by AI, and what’s going to matter shall be opinionated content material by native authors.”
That includes native authors not solely boosts authenticity but in addition social attain.
And right here’s the bonus: content material like this naturally earns hyperlinks.
“We don’t observe any particular strategy [to link building]. These native hyperlinks come naturally, and Web site Explorer exhibits regular development.”
When your weblog content material feels prefer it was created for the native viewers—not simply translated at them—it builds belief. That belief turns into shares, mentions, and backlinks. Invest in content quality and cultural connection. Local links will come, as you can see in the screenshot below.
Erik shared a case examine displaying how republishing articles can restore and even enhance search efficiency, particularly when rankings have plateaued or declined over time.
Search engines like google and yahoo usually favor freshness, and updating alerts that your content material continues to be related.
What to give attention to when republishing:
- Replace outdated data. Overview stats, examples, screenshots, and hyperlinks. Bringing the whole lot updated exhibits Google (and customers) that the article continues to be related and reliable.
- Enhance inside linking. Use Ahrefs’ internal link opportunities feature to connect your updated article to relevant content across your site. This helps distribute link equity and improves crawlability.
- Use People Also Ask as a content guide. Check Google’s “People Also Ask” section to find common related questions and add answers to your article. This makes it more comprehensive and aligned with search intent.
- Enhance with visuals and data. Add infographics, charts, or updated statistics to support your points. This makes content more engaging and can improve dwell time and backlink potential.
AI can pace up the fundamentals, but it surely nonetheless stumbles on nuance. You’ll want people to iron out the main points.
We applied an AI pre-translation system primarily based on a longtime glossary, prompts, and translation reminiscence, since AI has the aptitude to study and self-improve.
He added:
“The [Spanish] Weblog is run by people for now, however we’re additionally testing AI pulling articles, translating them, and importing to blogs.”
Begin with automation the place it helps most—bulk translation, glossary administration. However all the time depart room for human QA.
You want editors who can spot format glitches, catch damaged hyperlinks, and flag mistranslations or cultural mismatches that AI may miss. Construct in clear suggestions loops between translators, reviewers, and content material homeowners to verify high quality improves with each iteration.
By the best way, this is applicable to all forms of content material localization, together with product UI. For instance, we had a humorous mistranslation in Ahrefs’ interface for fairly a while. As an alternative of “don’t evaluate”, it stated “don’t purchase”.
Model guides and glossaries are essentially the most thrilling, however they’re your finest protection towards inconsistency.
“Glossary, type information, SOPs… that’s how we guarantee consistency.”
That is additionally the proper time to assume extra strategically and resolve issues like whether or not we wish to translate product names and authentic ideas.
To take care of high quality throughout languages, doc the whole lot. Create a shared glossary of key phrases, set up tone and formatting guidelines, and construct an editorial type information that translators and native writers can depend on. Erik shared how each of these paperwork search for Ahrefs.
Maintaining each model aligned with the supply? That’s a significant ache. However there’s hope.
Erik discovered a cool use case for the Content material adjustments function in Ahrefs’ Site Explorer (you can find it in the Overview report). If the English blog has made any changes to the content that resulted in more traffic, he’ll see exactly what it was. Then he can just repeat the same thing for localized versions.
Behind each profitable localization technique is a strong technical basis. That features getting hreflang tags proper, adjusting inside linking buildings, and guaranteeing localized media recordsdata don’t break the consumer expertise.
- Select your URL construction properly: Subfolders like “instance.com/fr/” are usually higher for web optimization as a result of they consolidate all of your area authority in a single place. Nation-code top-level domains (ccTLDs) like “instance.fr” can really feel extra native and reliable to customers in that nation, however cut up your web optimization efforts throughout a number of domains, making it tougher to rank globally.
Don’t redirect customers by IP: Routinely sending guests to a localized model of your web site primarily based on their IP can confuse Googlebot, which crawls largely from the US. Meaning your localized content material may not even get listed. As an alternative, present a language or area switcher so customers can select their most popular model with out being pressured to do so. - Contemplate a CDN: A content material supply community helps serve your web site from servers nearer to the consumer, which reduces loading occasions—particularly essential for worldwide audiences. Quick ship higher consumer experiences and this counts for Google, so this small infrastructure change can have large efficiency impacts.
- Implement hreflang correctly: The hreflang tag tells Google which model of your content material is supposed for which language or area. If it’s misconfigured, Google may serve the mistaken model in search outcomes. Ahrefs’ Web site Audit software helps you catch points right here by visually mapping your hreflang construction and flagging damaged hyperlinks between language variations.
We’ve received a full information on the technical points right here: 10 Worldwide web optimization Greatest Practices + Guidelines.
Last ideas
The perfect methods are born from expertise, particularly errors.
“I wouldn’t redo something, as a result of if I didn’t attempt, I wouldn’t know. Principally, FAFO.”
Don’t goal for perfection from day one. Check. Study. Adapt. Each market will train you one thing new. Monitor what works, double down on wins, and let go of approaches that stall.
Bought questions or feedback? Let me know on LinkedIn.
