Every brand operating across Belgium reaches the same budget conversation. Someone asks what it would cost to do this in three languages, and the answer that comes back is usually wrong in both directions.
It is not three times the cost. It is also not the marginal extra that suppliers sometimes imply. The money moves in a specific way, and understanding where it goes is what makes the decision possible.
Where the money actually goes
The shoot barely changes
If two versions are planned before the day, the second one costs ten extra minutes per piece. Same location, same light, same setup, the person says it again.
That is the cheapest multilingual content anyone can buy, and it is only available before the shoot. Deciding afterwards means a second shoot at full price, which is where most of the perceived expense of multilingual work actually comes from.
The edit genuinely doubles
This is the honest cost. Two languages mean two timelines, two sets of captions, two exports per format, two rounds of checking. There is no trick that avoids it, and a supplier who prices it as a small surcharge is either absorbing it or planning to cut corners on the second one.
The approval is where the weeks go
Here is the part nobody budgets. Each language adds a person who has to read and approve it, and those people do not work in parallel. The Dutch reviewer waits to see whether the French version changed, and the French reviewer waits for the Dutch one to finish.
A one week approval becomes three, not because anyone is slow, but because the queue is serial. This is the actual cost of multilingual content in Belgium, and it is paid in calendar rather than in invoice.
| Where you might expect the cost | What it really is |
|---|---|
| The shoot | Ten extra minutes if planned in advance |
| Translation | Real but small compared to everything else |
| The edit | Genuinely doubles, no way around it |
| Approval | Triples the calendar, and nobody plans for it |
| A second creator | Only needed when both faces must be native |
The last row is the one worth thinking about deliberately. Sometimes one bilingual person can carry both versions and the audience accepts it. Sometimes the accent decides the credibility, and then it is two creators or nothing.
The split is not demographic
The most common mistake is to allocate budget by population share. It produces a plan that satisfies a spreadsheet and matches nobody's behaviour.
What decides the language of a purchase is the category, not the buyer's mother tongue. Technical and professional purchases frequently happen in English regardless of who is buying. Some retail categories are researched in one language and bought in another. And a large part of the country consumes content in a language that is not the one it fills in on a form.
The practical version is to ask, for one product, in which language the last ten enquiries arrived. That is a better guide than any national statistic, and every company already has the data in its inbox.
What that changes
It usually means the honest answer is not three equal versions. It is one primary language done properly, a second done properly where the evidence supports it, and a third done as captions or not at all.
That is a defensible decision made on evidence, rather than an expensive gesture of even-handedness that leaves all three versions slightly underfunded.
What it costs to skip a language
Skipping is a legitimate choice and it is not free. The costs are specific and worth naming before making the decision.
A brand that speaks only one language in a bilingual market is read as taking a side, whether or not it intended to. In some categories that is irrelevant and in others it is fatal, and the difference is not predictable from outside the category.
There is also a search cost. Content that does not exist in a language cannot be found in it, and in a country where people search in different languages depending on the topic, that gap is silent and permanent.
And there is a distribution cost. Retail partners, resellers and public sector buyers frequently need material in a specific language to use it at all, so a missing version does not just underperform, it removes an entire channel.
When skipping is right
When the evidence says the audience does not exist. When the product is genuinely sold only in one region. When the alternative is three bad versions rather than one good one, which is the most common real situation.
Where machine translation is fine and where it is not
It is fine for a factual caption: opening hours, a price, a product name, a practical instruction. Nobody is persuaded by a caption, they are informed by it, and an occasional stiff phrase costs nothing.
It is not fine for anything spoken, anything with a joke in it, or anything where a wrong word changes an obligation. A translated spoken script keeps the rhythm of the original language, and native speakers hear that immediately even when every word is correct.
The practical rule is that translation is acceptable where the reader is looking for a fact, and unacceptable where they are deciding whether to trust you. That line is easy to apply and it saves most of the budget people spend on translating the wrong things.
The rule that resolves most of it
Do the practical content in every language you serve, and the atmospheric content in one.
Practical content is what people search for and act on: prices, hours, how a service works, what happens if something goes wrong. It has to exist in the language of the person who needs it, and it is cheap to produce because it does not require performance.
Atmospheric content is the brand film, the launch piece, the thing with music. It carries less specific information, it travels better across languages with captions alone, and it is by far the most expensive thing to produce twice.
Most Belgian brands do exactly the opposite. They produce the expensive atmospheric piece in three languages and leave the practical answers in one, which is the least efficient possible allocation of the same budget.
