Should we localize creative for each market or translate it?
Our answer
Have a native speaker rewrite ad copy and on-image text for each non-English market. Use machine or Meta's translation only as a first draft. English-speaking markets need only light edits.
- Confidence
- Low
- Strongest proof
- 2 Meta rules
- 2 datasets, 5 practitioners
- Sources
- 7 sources
- 9 statements
M
LCMI
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentHalf-localized creative performs poorly and wastes budget. MHI Growth Engine, an agency, advises that non-English markets get cultural review by native speakers, not just bilingual staff.
Translated campaigns underperform, and machine translation should serve only as a first draft for a native rewrite (Junto, a French performance agency).
Meta's built-in voiceover translation works only from English to Spanish. Meta's rules say the translated voiceover cannot be edited, and it does not translate on-screen text or captions.
Our judgment
Junto writes about paid search, so applying its advice to Meta creative is our extension. Where the line falls between native rewriting and full localization follows MHI's trigger of a cost per purchase within 30% of the US baseline and practitioner agreement, untested. For a US-only seller none of this applies.
Our judgment, built from Junto and MHI Growth Engine: our reading of these sources, not a finding from any one of them.
What the sources say
9 statements from 7 sourcesAn operator says one client's creative hit rate rose from 15% in the US to 28% after moving testing to the UK, where conversion rates were 30% higher. Winners moved to the US by post ID.
“Moved all creative testing to the UK. Hit rate jumped to 28%”
About paid search, not Meta: a practitioner says translated campaigns underperform and advises rewriting ads from the value proposition up, with machine translation only a first draft. A 60% CPA cut is unverified.
“Translated campaigns underperform for predictable reasons”
ElevenLabs says fully localized creative performed well but took weeks and did not scale, and dubbed-only videos sounded local but looked American. It ran on Google and Bing, not Meta, and gives no figures.
“sounded local but looked American”
An AI dubbing vendor reports its own Google and Bing campaigns in seven languages, not Meta. Non-English campaigns lifted conversions 17.6%, with $3.78 million incremental value at 7.16 ROAS, without a controlled design.
“Non-English campaigns delivered a 17.6% lift in conversions compared to English campaigns alone”
Meta's developer docs list opt-in Advantage+ creative features that translate ad text and text inside images, each now opted into individually.
Meta's developer docs say the translate_voiceover feature supports English to Spanish only, on single videos in Feed and Reels. Text overlays and captions are not translated.
“cannot be edited or regenerated”
A Canadian and US retailer says Canadian shoppers increasingly ask whether products are local, and it pushes Canadian brands in its Canadian marketing. It now runs a separate US site that ships from the US.
“we've seen a huge uptick in customers looking for those sorts of things like, where's this coming from? Is it local?”
Moburst's Intercept claims AI voice-over translation plus intent-matched variants cut multilingual CPA by 19% using Meta's Advantage+ translation tools. No sample or method is given.
“By using Meta’s built-in translation and AI voice-over tools combined with modular master creative, a team of two to three people can manage multilingual campaigns across five or more languages.”
Kamal Razzak of MHI Growth Engine, with no data, says English-speaking markets need light adaptation, while non-English markets need full translation with native-speaker review, not just bilingual staff.
“half-localized creative performs poorly and wastes budget”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer