Localisation means reshooting, not translating

The hook. The first three seconds are built on a recognizable situation, and situations differ across markets. "Couldn't withdraw money for a week" works where that's a common experience and sounds odd where banks work fine.

Pacing. A Russian phrase is roughly a quarter longer than its English equivalent; Spanish is a third longer. A dubbed track either doesn't fit the edit or gets delivered as a tongue-twister, and the viewer hears the rush before understanding the message.

The face. The audience reads "one of us" versus "outsider" in a split second, before any text. An actor obviously shot for a different market undermines trust in exactly the thing the video was made to build.

We transfer the structure, not the lines. The second-by-second script stays the same: where the hook is, where the demo is, where the CTA is. The content of each segment gets written from scratch, with a local creator or casting director.

We shoot with local faces. It's more expensive than dubbing and cheaper than opening a market twice — the second time after the first attempt didn't work.

Reshooting a batch for a market runs roughly 60–70% of the original shoot cost: the brief, structure, and editing decisions already exist; you're paying for the crew and the cast.

Dubbing costs 10%. The four-to-six-times price gap looks decisive right up until you compare retention: on reshot material it's typically 1.5–2x higher, and the cost per action is lower by the same factor. The savings on localization have a negative return.

There are exceptions. Educational formats without a face on camera, interface demos, anything that relies on the screen rather than the person — these translate fine. There, localization really does come down to the audio track and captions.

The rule is simple: if the video relies on trust in the speaker, reshoot it. If it relies on what's shown, you can translate it.