SEA and PPC Campaign Translation: How to Localize Keywords and Ad Copy¶

An agency launches Google Ads across four new markets. They take an existing English campaign, run the keywords and ad copy through DeepL, swap the currency - and go live. One month later: $18,000 spent, CTR below 0.8% on most languages, cost per conversion twice as high as the English-language market. The geo-targeting is correct, the bids are reasonable, the campaign structure is solid. Something still went wrong.

The problem isn’t the settings. The problem is that translating keywords and ad copy is a fundamentally different task from translating documents or articles. People in different languages search for the same thing differently - different phrasing, different word order, different intent. Taking an English keyword list and translating it produces technically correct phrases that real users either don’t search at all, or search in a completely different context than you intended.

Here’s what goes wrong and how to fix it.

Why Direct Keyword Translation Fails¶

A concrete example. Someone in the US looking for sneakers types “best running shoes” - a marketing, recommendation-seeking phrase. Someone with the exact same need in Germany types “Laufschuhe Test” - they want product test results and comparisons, not marketing promises. Same product, same category, completely different search intent. If you translate “best running shoes” literally and plug it into a German-market campaign, you won’t hit the query people are actually using.

Another example: “billig” in German technically means “cheap.” But to a native speaker, the word carries a connotation of low quality. Translate “cheap prices” as “billige Preise” in an ad headline, and a significant portion of your audience will reject the ad before they ever click. The right word for that context is “günstig” - affordable, a good deal, without the quality implication.

As Ninja Promo specialists note:

“The most common mistake companies make when translating keywords or ad copy is assuming that a direct translation equals the keywords people actually search. Search behavior doesn’t translate.”

More examples of how languages diverge:

• In Mexico “car” is “carro,” in Spain it’s “coche.” Run a campaign with just one variant and you immediately lose half your audience
• Mexican online gamblers search for “bonos de apuestas” and “bonos de bienvenida,” not the technically correct translation “bono casino online” - result: irrelevant impressions and zero CTR
• German compound nouns aren’t direct translations of English equivalents: “Laufschuhe” isn’t a translation of “running shoes,” it’s a distinct search concept with its own volume and intent

The result of direct keyword translation: your ads appear for queries with low actual traffic, CTR drops, Google reduces Quality Score, CPC climbs. Same spend, fewer clicks, even fewer conversions. The budget keeps draining.

Keywords: Research in the Target Language, Not Translation¶

The right approach isn’t “translate keywords” - it’s “research keywords in the target language.” That’s a different task with a different process.

Separate research for each market. You don’t take the English list and translate it. You open Google Keyword Planner or SEMrush in the target language and region and find out how people on that market actually search for your product. You study synonyms, regional variations, question-based queries, compound constructions.

Native speakers as a mandatory step. No tool replaces a review by a native speaker. They’ll catch whether the phrasing sounds natural, whether there’s unwanted subtext, which local expressions actually appear in search. This review is not optional - it’s a required step before launch.

Long-tail keywords separately. According to Lunio’s research, long-tail keywords account for 40-60% of local traffic on regional markets and diverge between languages even more than short-form queries. Direct conversion of long-tail phrases from English almost always produces either low volume or different intent.

On tools. Google Keyword Planner is free and good for PPC - since 2025 it offers forecasting breakdowns by city, region, and device type. Downside: often shows volume ranges (“100-1K”) instead of exact numbers. SEMrush provides more precise data (not ranges), especially useful for detailed CPC research on specific markets. For markets where Google doesn’t dominate: Yandex Wordstat for post-Soviet markets, Baidu for China, Naver for Korea - completely different search demand dynamics on each.

Ad Copy Transcreation: Beyond Translation¶

With keywords the situation is relatively clear - you need separate native research. With ad copy it’s more complex: it’s not just words that matter, but tone, cultural context, emotional subtext and even the rhythm of a phrase.

A real case from practice. A trading platform for the Mexican market launched the headline “trading sin comisiones” - a technically accurate translation of “trading without commissions.” CTR: 1.2%. The team did native localization - changed it to “Invierte sin comisiones ocultas” (Invest with no hidden fees). CTR jumped to 3.8% and cost-per-lead dropped by 27%. The difference wasn’t the language - it was a message that resonated with how that audience thinks: they’re not afraid of commissions, they’re afraid of hidden fees.

As Verbo Labs notes:

“Culturally adapted campaigns can increase conversions by up to 30% while reducing the risk of brand misinterpretation.”

The difference between translation and transcreation in a PPC context:

• Translation - converting words from one language to another. Technically correct but often sounds mechanical. Native speakers sense this immediately and don’t click
• Transcreation - adapting a message so it triggers the same response in a new audience. Words may change completely, but the intent and emotion survive, adapted to the culture and language of the target market

The famous example of a translation failure: HSBC’s campaign “Assume Nothing.” In several markets it was translated as “Do Nothing.” The company needed $10 million to repair the reputational damage and relaunch the campaign.

As Weglot describes the difference:

“Translation is word-for-word conversion from one language to another that is literal, quick, and can sound like it was written by a robot. Transcreation is a step beyond localization where the original message is rewritten entirely in a new way to match cultural meaning, emotion, and intent, even if the words are completely different.”

For agencies this means: if you hand ad copy to a translator and ask them to “translate it” - you’ll get technically correct text that converts poorly. The right brief for the person doing the work: “Adapt this message for the target market, preserve the ad’s intent, and stay within the character limits. Don’t translate word for word.” That’s a different task and a different level of work.

Google Ads Character Limits and How Different Languages Complicate Things¶

There’s a technical constraint that frequently becomes the bottleneck in localization and gets overlooked until the text is already written.

Google Ads enforces hard limits: - Ad headline: 30 characters - Ad description: 90 characters - Responsive Search Ads: up to 15 headlines at 30 characters, up to 4 descriptions at 90 characters

The problem: most languages expand text relative to the English source. Words get longer, function words appear that don’t exist in English, grammar requires additional elements.

For real localization on German-speaking markets this plays out like this: you have a headline “50% off premium software” - 26 characters. The equivalent in German “50% Rabatt auf Premium-Software” - 32 characters, doesn’t fit. You either cut or rephrase from scratch to hit the limit.

Practical solutions: - Write original English headlines shorter than the limit (18-22 characters instead of 28-30) when you know you’ll be localizing into expansion languages - Give the person doing the work a clear brief: “Priority is fitting within the character limit while preserving intent. Don’t translate literally if it doesn’t fit.” - For CJK markets (China, Japan, Korea) plan for limits that are effectively half the standard from the start - Check character limits in the actual Google Ads interface, not manually - it has a built-in counter

One nuance about Germany: market research shows that part of the German-speaking audience uses the English word “trainers” when searching for sports shoes rather than “Sportschuhe.” That’s another argument for native research: native speakers know which hybrid search terms actually exist.

Quality Score and the Financial Impact of Poor Translation¶

This is probably the strongest argument for those who think “good enough, let’s just launch.”

According to Spider.af:

“76% of PPC budgets are wasted on ineffective campaigns, with most businesses losing three-quarters of their ad spend to preventable errors.”

Poor keyword and ad copy translation is one of the main drivers of that waste. Here’s how the chain works:

1. Keywords are translated, not researched → ads appear for irrelevant queries
2. Native speakers see mechanical text → don’t click → CTR drops
3. Low CTR → Google lowers expected CTR (the first Quality Score component)
4. Lower Quality Score → Google raises the minimum CPC bid for the same ad positions
5. Higher CPC → fewer clicks for the same budget
6. Fewer clicks → fewer conversions → customer acquisition cost climbs

Moving Quality Score from 5 to 8 delivers a 37% CPC reduction - not theoretical, but a documented effect Google publicly acknowledges. Same positions, same auction, you pay a third less.

A retail test showed an 18% CPA reduction after proper localization compared to direct translation. If your current cost per conversion on a market is €80, proper localization could drop it to €65-66 without changing the budget - purely through better copy and more relevant keywords.

PPC audits typically uncover 20-40% of budget being wasted and available for reallocation. A significant share of that comes from irrelevant keywords and low-CTR ads. On a campaign with a €10,000 monthly budget, that’s €2,000-4,000 per month you can recover just by doing localization right.

Campaign Structure for Multilingual SEA¶

Beyond translation and localization quality, there are structural decisions that are critical for multilingual campaigns to work correctly.

Separate campaigns for each language. Google Ads only allows language targeting at the campaign level - not the ad group level. If you mix languages in one campaign, you can’t correctly track which language converts or optimize bids separately. Separate campaigns aren’t optional - they’re a requirement for proper analytics.

Separate landing pages for each language. Non-negotiable. Sending traffic from a German-language ad to an English page means high bounce rates and zero conversions. Someone who arrived from an ad in one language expects to see content in that same language. And Google Ads factors landing page quality and relevance into Quality Score calculations - a language mismatch between ad and page directly lowers QS and raises CPC.

Localized negative keywords. Each market needs its own negative keyword list. Terms irrelevant on one market may be key terms on another. Direct conversion of negative keywords between languages fails for the same reasons direct keyword translation fails.

A/B testing per market, separately. Never run an A/B test that mixes multiple languages - results will be unreadable. Each market tests independently: its own control, its own variation, its own statistical threshold.

Bidding strategy by market. Automated bidding works well for stable markets with 30+ conversions per month - the algorithm has enough data to work from. For new markets, start with manual CPC or maximize clicks without a CPA constraint. Once you’ve collected 30+ conversions, switch to Target CPA or Target ROAS.

Localized ad extensions. Phone numbers should be local, physical addresses local, the USPs in sitelinks adapted to the market. Ad extensions affect both Quality Score and CTR.

To better understand how paid channels interact with organic in an overall agency marketing strategy, see our article on content marketing for translation agencies.

Tools for Researching and Localizing PPC Campaigns¶

Several tools meaningfully simplify work on multilingual campaigns.

Google Keyword Planner - free, built into Google Ads. Good for initial research and forecasting. Since 2025 it provides forecast breakdowns by city, region, and device. Downside: often shows volume ranges (“100-1K”) rather than exact numbers. Switch the region and interface language to get correct results for each market.

SEMrush PPC Keyword Tool - provides more precise volume data (not ranges), lets you build campaign structure directly in the tool, good for CPC research by region. For agencies that regularly launch campaigns across multiple markets, usually a justified investment.

Manual native research. Not replaceable by any tool: searching Google in the target language, reading autocomplete suggestions, browsing “People also ask” and “Related searches.” This gives you understanding of the real search context that no tool’s dataset captures.

Alternative platforms. For markets where Google doesn’t dominate: Yandex Ads (Yandex Wordstat for keyword research) for post-Soviet markets, Baidu for China, Naver for Korea. These platforms have different audiences, different bid dynamics, and different approaches to keywords altogether.

Brand glossaries. If you regularly work across multiple markets - maintain a glossary: client product names, brand terms, industry-specific vocabulary and their approved equivalents in each language. This ensures consistency across campaigns, ads, and landing pages and saves time on every new launch.

Native speaker as final gate. Regardless of tools - before launch, have all copy reviewed by a native speaker. Not optional. Even excellent transcreation can have subtext or sound awkward in a specific context - only a native speaker catches that before you spend the budget.

Multilingual PPC Campaign Pre-Launch Checklist¶

If you need a quick reference - here’s the minimum checklist before launching a new language campaign.

For a deeper look at how to build marketing content for translation and language services - see our article on translating marketing materials for international markets.

FAQ¶

What’s the difference between translation and transcreation for PPC ads?¶

Translation is a word-for-word conversion from one language to another. Transcreation is rewriting the message so it triggers the same response in a new audience - the intent and emotion survive, but the words may change completely based on cultural context. For PPC ads this means the copy can look very different from the original while delivering significantly higher CTR and conversion rates.

Can you use DeepL or ChatGPT for PPC keywords?¶

As a starting point for research - yes. As a final keyword list - no. Automated translation doesn’t know local search volumes, regional variations, or actual search intent. Any draft output needs validation in Google Keyword Planner for the target region and review by a native speaker before launch.

How many characters are in a Google Ads headline and description?¶

Headline: 30 characters. Description: 90 characters. For double-byte character languages (Chinese, Japanese, Korean) each character counts as 2 - the effective limit is half. For Responsive Search Ads: up to 15 headlines and up to 4 descriptions.

Do you really need separate landing pages for each language?¶

Yes, and it’s not optional. Sending traffic from an ad in one language to a page in another language guarantees high bounce rates and zero conversions. Additionally, Google factors landing page relevance into Quality Score - a language mismatch between ad and page directly raises CPC.

How does ad copy translation quality affect Quality Score?¶

Quality Score has three components: expected CTR, ad relevance, landing page quality. Mechanical or awkward translation reduces CTR (native speakers see the copy isn’t natural and don’t click). Lower CTR → lower expected CTR component of QS → higher CPC. Moving QS from 5 to 8 delivers roughly a 37% CPC reduction - at a €10,000/month budget that’s €3,700 saved monthly.

How much more does full localization cost compared to simple translation?¶

Native keyword research + ad copy transcreation + native speaker review typically costs 40-70% more time and budget than simple translation. But consider: proper localization reduces CPA by 18-27% (documented in real campaigns), significantly boosts CTR, and cuts spend on irrelevant impressions. Over a 2-3 month horizon, proper localization pays for the additional cost and then some.

Which languages are hardest to localize for Google Ads?¶

Two tiers of difficulty. First - languages with high text expansion: German (+25-40%) and languages with complex declension systems (Polish, Hungarian, Czech) where a single concept can have 6+ forms, all of which need to appear as separate keyword variants. Second - languages with non-Latin scripts (Chinese, Japanese, Korean), where each character counts double and the effective headline limit is 15 characters instead of 30.

How do you measure whether localization outperforms direct translation?¶

Run both in parallel - the old campaign (direct translation) and the new one (native localization) - with equal budgets for 2-4 weeks. Key metrics to compare: CTR per language, CPC, Quality Score, CPA. Differences in CTR and Quality Score are typically visible within 1-2 weeks; differences in CPA emerge in 3-4 weeks once you have enough conversion volume for a valid comparison.