Write a Commercial That Can Travel Across Languages Without Losing the Idea

A line that runs one second in the original language can run three in another. That is the cheap problem, and you can usually solve it by cutting something. The expensive problem is that the line was the joke.

What travels between versions is a mechanism: the specific thing that makes a beat work for someone who already knows the language. Words, timing, performance, gestures, and product offers are the parts that may need to be authored again in each place. Those are different jobs with different owners, and treating them as one job is how a version ends up with correct words and no joke.

The practical consequence for a treatment: find the beat that carries meaning in language, say so out loud, and name what a qualified local writer will need to decide. Promising that a joke travels is not a creative decision. It is a prediction about a language you may not speak.

Find the part that depends on language

Before anything is translated, name the source of the effect. There are a few common ones, and they behave differently when they cross a border.

A verbal ambiguity lives in a word or phrase that carries two readings. A expected response lives in the audience's knowledge of what a person normally does when they hear a certain thing. Social context lives in the situation that makes the instruction natural — who is allowed to say this to whom, in which room, carrying what. A reveal lives in the cut that shows what happened while everyone assumed something else.

Here is an invented spot, made up for this piece, to keep the discussion concrete. Thirty seconds. A kitchen with too much on the counter. One character — call her A — comes through the doorway carrying a stack of flat-packed storage boxes, her face visible, and says, "Make room." The other character, B, is at the counter and says nothing. B hears the phrase as an instruction to build a room, turns from the counter, and starts building one. By the time A comes back, there is a chalk rectangle on the floor with an actual door leaning in it, a small lamp, a chair. A looks at it. Beat. Then she opens the flat-packs, assembles the storage, and the clutter disappears into it. B's little room stays, now with a shelf inside, and they both sit in it.

The load-bearing element in this spot is a verbal ambiguity plus an expected response. "Make room" has a settled meaning in a kitchen: clear some surface. The comedy is that B misses that meaning entirely and takes the words at their most physical. The audience laughs because they know the idiom and they know what anyone else in that kitchen would have done.

Now mark the exact moment where understanding depends on that source. It is not the second the line is spoken. It is the turn — the instant B turns away from the counter toward something else. Before that instant, nothing has happened. After it, everything has. That turn is the hinge of the whole thirty seconds.

This is what a transcript cannot carry. Hand a local team a transcript and they receive words. Hand them a marked beat — here is the situation, here is the line, here is the turn, here is what the audience already knows — and they receive a job. They can now tell you whether the mechanism exists in their language, and if it doesn't, they can tell you before you have shot anything.

Give the local writer a purpose, not a frozen sentence

The brief you send out is not "translate 'make room.'" It is a statement of situation and purpose, plus the constraints that are genuinely fixed. Something like: a kitchen that needs space; one character wants it cleared; the second character takes the request in its most literal direction and acts on it; the misunderstanding must land in under two seconds of screen time, and the product must resolve the clutter that prompted the line. Then the question: in this kitchen, is there something a person would naturally say that means clear the counter and could be taken as an instruction to build?

There are three honest answers, and a treatment should be written so that all three can come back.

Yes, there is such a phrase, and it does the same job. Then the beat is basically held, and the work moves to performance and timing.

No, but there is a different misunderstanding that produces a comparable action — the wrong reading sends B off to do something else visible and funny. Then the beat is rewritten and the surrounding shots may need to move.

No, and there is nothing here. Then the verbal gag does not exist in this language, and the version needs a different joke. That is a real answer, not a failure of effort.

What the local writer should not be handed is a rendering of the source line to approve. And what should not be counted as approval is a machine translation. Machine translation can tell you what a phrase means. It cannot tell you whether anyone says it in a kitchen, whether it sounds like a person or a slogan, or whether the double reading survives the trip. Use it the way you'd use a rough pencil sketch: something to react against, for the person who will actually write the line. The industry word for that work is transcreation — writing the beat again in another language rather than rendering the words — and it is a writing job, not a checking job.

Two things worth keeping out of the brief entirely: invented cultural traits, and confident claims about what a given audience will find funny. Neither is yours to supply. The person with local fluency and local creative responsibility supplies those, or the beat stays open.

Trace an adaptation into performance and picture

Take two routes for this spot and follow them all the way into what gets photographed.

Route one keeps the wordplay. A qualified local writer finds a phrase that means "clear the counter" and can plausibly be taken as an instruction to build. A still delivers it on camera, face visible, in the same medium shot. Everything else about the spot is untouched.

Route two drops the wordplay. There is no line. A comes in, sets down the flat-packs, and measures a gap in the air with both hands to show how much shelf width is needed. B reads the measured width as the width of an opening and starts framing a doorway at that size. The chalk rectangle, the door, the lamp, the chair, the reveal, and the product resolution all follow.

Route two solves the obvious problem and creates a subtler one. The language dependency in the dialogue is gone, which removes the lip-sync risk entirely. But B is now a person who takes measurements from hand gestures and builds things, and the spot has to establish that before the turn, or the turn looks arbitrary. In the original version, B's oddness arrived free with the mishearing. Here you may spend a second or two of setup you did not previously need — a pencil behind the ear, a tape measure on the counter — and that comes out of the same thirty seconds.

That trade is worth understanding before you compare the two approaches to performance at all.

Voice replacement over shared images is cheap and often right. It suits material where the mouth is not carrying the meaning: A's back is to camera, the line is off-screen, the face is small in frame. In this spot, the line is the hinge and A's face is visible at the hinge, so a replacement is a risk you can measure but not eliminate. Route two has no line to replace.

An exchange that needs locally performed material — Route one — is more expensive and often right. It requires the actor and the day. It may also require that the setup was shot to survive a differently shaped conversation, which is a decision you make in preproduction, not in the edit.

Which brings up the cheapest version of this decision, and the one most often missed. If versions are already expected when the spot is shot, you can protect the dub before you have one: put A's line off-camera, or turn her toward the doorway as she says it, or frame wider so the mouth is not the only carrier of the beat. Each of those costs you the face at the exact moment the spot turns. That price is easier to weigh in preproduction than in a mix session, when the alternative is a reshoot or a line that never quite sits in the mouth.

And silence is not a universal solution. Route two removes a language dependency and puts the weight on a behavior — a person building a small framed room inside a kitchen — that the audience has to read as absurd. Absurdity is a social judgment, not a visual one. Whether this reads as funny, or as strange, or as something a person might actually do, is exactly the kind of question a qualified local reader answers and a treatment writer cannot. Route two is not automatically the safe option. It is a different bet.

Keep commercial inputs separate from creative equivalence

Some parts of a version are not creative problems at all, and they should not be quietly absorbed into the creative one.

Local product availability is a fact with an owner. Whether the storage line is sold there, in what configuration, in what package, with what name on the box, is not something a treatment can settle by describing it well. If the product isn't there, or the packaging differs, you may not have a version of this spot — you may have a different spot that shares footage.

Offers, prices, and required wording are the same. They have their own reviewers and their own approval chain, and their length is an input to the edit rather than a translation of anything. An end card that runs four seconds in the original and needs eight in a market is not a language problem. It is four additional seconds taken from the beat you just re-timed, and it belongs in the timing conversation from the beginning rather than after the first cut.

The useful artifact at this stage is not a translated script. It is a short note that records what is shared, what is authored locally, and what is still unresolved. For the invented spot above, it looks like this.

Nothing in that table claims a version has been approved. It says which questions have answers and which ones are waiting.

The shared purpose, the open question

Strip the spot down and the shared purpose is small and portable: two people, a kitchen with too much in it, and a product that finally makes the space. The misunderstanding is the detour. That much can be agreed in any market, because it is a situation, not a sentence.

The open question belongs to a local writer and reads roughly like this: in your kitchen, is there something one person would say that means "clear the counter," and could another person hear it as an instruction to build? If yes, we need the line and a sense of how long it takes to say. If no, we need to know that, and we need to know whether the alternative is a different misunderstanding or a different joke.

What waits on that answer is specific. How A's line is shot, and whether her face stays in the frame at the turn. Whether B speaks at all in this version. How long the pause before the reveal runs, since the pause is now measured against a line that may not exist yet. Whether the end card leaves room for the beat, or the beat has to lose a second to make room for it.

None of those should be filled in with plausible-sounding lines just to make the treatment read as finished. An unresolved language beat is a live question with a named owner, and saying so is more useful to everyone downstream than a version that looks complete and isn't.

Frequently asked questions

What is the difference between translating a line and transcreating it?

Translation renders words. Transcreation rewrites the beat in another language. The body describes it as a writing job, not a checking job: a local writer needs the situation, purpose, constraints, and the marked turn, not a frozen sentence to approve.

Why is a transcript insufficient for adapting a language-dependent joke?

A transcript gives words. What the local team needs is the mechanism—the situation, the line, the turn, and what the audience already knows. In the kitchen example, the hinge is not the line itself but the instant B turns away from the counter, so the local writer can say whether that mechanism exists or must be replaced.

What three answers might a qualified local writer return?

Yes, there is a phrase that does the same job; no, but there is a different misunderstanding that produces a comparable visible action; or no, and nothing here works, so the version needs a different joke. The piece says a treatment should allow all three to come back.

What trade-offs arise if a version drops wordplay for a visual misunderstanding?

Route two removes the lip-sync and language dependency, but it must establish why B takes hand-measured gestures as building instructions before the turn, which may cost extra setup seconds. It also shifts weight onto a behavior whose absurdity is a social judgment, so a qualified local reader—not the treatment writer—is needed to say whether it reads as funny, strange, or plausible.

Which commercial inputs should be kept separate from creative equivalence?

Local product availability and packaging are facts with owners; offers, prices, required wording, and end-card length have their own reviewers and affect the edit. The useful artifact is not a translated script but a note or table recording what is shared, what is locally authored, and what remains unresolved.

Further reading

• Translate a Series Pitch Without Promising a Different Show — For the note that hands a beat to a local writer: how to mark where meaning sits in address, register, or joke target rather than in the words.
• Plan Commercial Cutdowns Before the Long Version Eats Every Beat — Your re-timed beat and a longer end card compete for the same seconds; this shows how to plan runtimes around what a first-time viewer must know.