The fastest way to insult a market is to translate a post into its language and assume the job is done. A literal translation carries the wrong idioms, a humor style that does not land, proof a local reader cannot verify, and a call to action pointing at a page that does not exist in that market. Localization is not translation. It is deciding, per element, whether to translate, transcreate, or rebuild so the post reads as if it were written by someone from that market.
This guide is about that decision. It is a sibling to multi-brand social media management , but the axis is different: there it is many brands in one market, here it is one brand across many markets. The voice stays the same; the language and cultural context change.
Three levels of adaptation
Every piece of a post falls into one of three buckets. Naming the level prevents both lazy translation and needless rework.
- Translate. Factual, neutral content where the words carry no cultural weight: a spec, a date, a feature name. Direct translation is fine.
- Transcreate. Anything where intent matters more than words: headlines, hooks, jokes, taglines, CTAs. You recreate the effect in the target language, often with different words entirely. “Slip into something more comfortable” cannot be translated; it must be reinvented.
- Rebuild. Content that does not transfer at all: proof, examples, references and offers tied to the source market. A US case study or a “free shipping in the lower 48” line gets replaced with a local equivalent, not adapted.
The single most common localization failure is translating something that needed transcreating, especially the hook and the CTA. That is what makes a feed sound machine-made.
A per-element decision table
| Post element | Default level | Why |
|---|---|---|
| Product spec / date | Translate | Neutral facts |
| Hook / first line | Transcreate | Must grab a local reader |
| Tagline / slogan | Transcreate | Intent over words |
| Idiom or joke | Transcreate or cut | Rarely transfers |
| Statistics / proof | Rebuild | Must be locally credible |
| Case study | Rebuild | Use a local customer |
| CTA + landing link | Rebuild | Point to the local page |
| Hashtags | Rebuild | Local tags differ entirely |
Notice hashtags: a popular tag in one market is dead or means something else in another. Never translate a hashtag; research the local one. Build a per-market set the way you would a hashtag strategy from website content .
Source-language and adaptation, not translation-only
A workflow that ends at translation produces flat content. A working localization workflow has a market reviewer in the loop:
- Lock the source post in your primary language, with the intent of each element noted.
- Adapt per market, applying the right level to each element.
- Local review by someone who lives in the market, checking that the hook lands, the proof is credible, and nothing is accidentally offensive.
- Approve and publish on the local channel, in the local time zone.
Step 3 is non-negotiable. A native speaker who is not from the market is not enough; idioms, humor and sensitivities are regional, not just linguistic. Mexican and Argentine Spanish are not interchangeable, and neither are UK and US English.
Things that quietly break across markets
- Time zones. Your best time to post workflow is per-market. A single global schedule posts to half your markets at 3 a.m.
- Platforms. The dominant network shifts by region. Do not assume the same channel mix everywhere; let each market’s multi-channel social content plan reflect where its audience actually is.
- Text length. German runs long, so designed templates overflow. Some scripts run short. Build visual templates with room to flex.
- Direction and layout. Right-to-left languages like Arabic and Hebrew flip the entire layout, not just the text.
- Formats and units. Dates, currency, measurements and phone formats are all local. A “12/06” date is read two different ways across markets.
- Color and imagery. Symbolism and stock-photo norms differ; a visual that reads as friendly in one market can read as off in another.
Keep voice constant, context local
The trap on the other side is over-localizing until each market sounds like a different company. The fix is a shared voice layer plus a local context layer. Brand voice, the personality, the values, the way you treat customers, stays constant everywhere and lives in your social media brand guidelines . Only the language, references, proof and offers localize. A reader in any market should recognize the same brand; they should just never feel they are reading a foreign import.
Measure per market, never in aggregate
A global average hides everything that matters. One market can carry strong numbers while another quietly fails, and the blended figure looks fine. Report per market and watch:
- Engagement rate per market, not pooled.
- Transcreation rework rate. High rework means the source intent was unclear or the adapter lacked context.
- Local CTA click rate. Low clicks often mean the link points at a non-localized page.
Where Utin fits
Utin is being built to scan each market’s localized website pages, so posts for that market start from proof, offers and language that actually exist there, rather than from a translated version of the home market. The shared brand voice carries across; the local context comes from local sources. Teams publishing across markets can register interest as an early pilot.