How Freelance Video Editors Sell Translation as a Deliverable

Why This Is a Good Fit for Solo Operators

Freelance editing has a structural problem: income is capped by hours, and hours are capped by being one person. The usual escape routes are raising rates until clients resist, or subcontracting, which turns a craft business into a management business.

Localization is a third option that fits the solo model unusually well.

The work is largely asynchronous — you set a batch running and do something else. It scales without proportional hours, because the fifth language costs a fraction of the first. It uses assets you already have from projects you already delivered. And it is sold to existing clients, who are the cheapest clients to sell to and who already trust you with their material.

Most importantly, it is a deliverable rather than a service. You are not selling more of your time. You are selling a set of files that did not exist before, at a price the client evaluates against what those files are worth in a new market rather than against your hourly rate.

That last point is the commercial unlock. An editor who charges by the day is competing on rate. An editor who delivers nine language versions of a client's product film is selling market access.

What to Offer

Keep the menu short. Three products cover almost every situation and prevent the endless customisation that destroys margin on small engagements.

Subtitle package. Translated, timed subtitle files in the formats the client needs, plus optional burned-in styled versions for social. Fastest to produce, easiest to sell, and the natural entry point. Many clients who hesitate at dubbing will buy this immediately.

Voiced version. Translated narration replacing the original voiceover, laid over the existing music and effects, delivered as a finished mix. This is the product with the clearest value story for the client.

Full localized cut. Voiced audio plus localized on-screen text — titles, lower thirds, captions, end cards. This is where you have a real advantage over the client doing it themselves, because it needs your project files and your craft.

Resist offering per-asset customisation beyond these. A defined menu is faster to quote, faster to deliver, and reads as a professional service rather than an improvisation.

Preparing Your Own Projects

The habits that make localization profitable are habits you adopt now and benefit from later.

Keep text as editable text layers. Never flatten titles and lower thirds into the render if you can avoid it. This single practice is the difference between a localized graphics pass taking twenty minutes and taking a day.

Design with expansion room. German and Finnish run notably longer than English. Lower-third boxes sized exactly to the English string will break in several target languages. Building in headroom costs nothing at design time.

Keep music and effects separate from voice in your mix. If you deliver with a clean instrumental and effects bed, voiced versions are straightforward. Where you only have a full mix, dialogue separation can recover a usable bed, but starting with stems is always better.

Archive a textless version of everything. The single most valuable thing you can keep for future localization work.

Keep your project files organised and backed up. Obvious, routinely neglected, and the difference between quoting a past project confidently and declining because you cannot find the assets.

Adopt these across all your work regardless of whether localization is currently in scope. Every project delivered this way becomes a future upsell that costs you nothing to preserve.

The Workflow, Step by Step

A repeatable sequence keeps this from becoming ad-hoc work you underquote.

Confirm the source is locked. Never localize before picture lock. Clients who ask for parallel workstreams to save time are creating rework for both of you.

Generate and clean the transcript. Automated transcription gets you most of the way. Your job is fixing proper nouns, product names, and the filler and false starts that will translate badly.

Get the client to approve the transcript. This is the highest-value step in the entire process and it costs one email. It surfaces the "actually we should reword that" conversation once, in the source language, rather than after you have generated audio in six languages.

Build the terminology list. Product names, people's names, brand terms, technical vocabulary, and anything that must not be translated. Ten minutes here prevents systematic errors across every language.

Translate, then review the text before generating audio. Correcting text is fast. Regenerating audio is not. Any workflow that generates voice before the translation is reviewed is manufacturing rework.

Generate the voice. Match voice character to the original — a warm narration should not become a clipped corporate read.

Localize the graphics. Your text layers, your target-language strings, checked for fit at every safe-area boundary.

Mix and QC. Levels, sync, loudness to the client's delivery spec.

Deliver with a clear file naming convention that includes language codes. Clients with multiple language versions will thank you, and it prevents the wrong file going to the wrong market.

Quoting Without Losing Money

The most common freelance mistake here is quoting hourly. Localization time is dominated by waiting, and you cannot bill for waiting.

Quote per finished minute, per language, per tier. Something like a base rate for the subtitle tier, a higher rate for voiced, and a higher rate again for the full localized cut — with a separate one-time setup fee covering transcript cleanup, terminology, and graphics templating.

The setup-plus-per-language structure matters commercially. It reflects the reality that the first language carries most of the work, and it makes the fifth language feel cheap to the client — which is exactly the behaviour you want, because your marginal effort really is low.

Things to build into the quote that people forget:

Transcript cleanup time, which is real work.
One round of client revisions per language, with additional rounds chargeable.
Reviewer cost if you are including native-speaker review.
Processing cost, which is small but nonzero.
A rush multiplier, because clients will ask.

And be explicit in writing about what is not included: source content changes after transcript approval, additional languages, re-recording, and creative rewrites for cultural adaptation. Those are all legitimate additional work and the boundary needs stating before it is tested.

QC in Languages You Do Not Speak

This is the part that makes editors hesitate, and it is more tractable than it feels.

Most defects in localized video are technical, not linguistic, and you are already qualified to catch them:

Audio sync against picture at the head, middle, and tail of every version.
Loudness and true peak to the delivery spec.
Dialogue sitting correctly against the music and effects bed.
No truncated, missing, or duplicated segments.
Every speaker present and distinct where the source had multiple speakers.
Subtitle timing, reading speed, and line length within limits.
Subtitles not colliding with on-screen graphics.
Graphics text fitting its container at every boundary, with no clipping or overflow.
Accented and non-Latin characters rendering correctly everywhere, including in file names.
Consistent frame rate and timecode across the set.

For linguistic quality, be straightforward with the client. Two honest positions work:

Offer a reviewed tier where you engage a native speaker and price it in. Build a small roster of reviewers for the languages you deliver most.

Or offer an unreviewed tier and say plainly that linguistic review is the client's responsibility — many clients have native speakers in-house and would rather review it themselves.

What does not work is silently delivering unreviewed output and letting the client assume otherwise. They find out eventually.

Pitching Existing Clients

You do not need new clients for this. You need a different conversation with the ones you have.

Lead with a finished sample, not a proposal. Take a project you already delivered for them, localize sixty seconds of it into one plausible market language at your own cost, and send it. "Here's your product film in Spanish — took me an afternoon" is a fundamentally different message from a rate card.

Pick the language deliberately. Look at where the client actually does business, where their site has traffic, or where their competitors are active. A relevant language proves you understand their business; a random one proves you have a tool.

Frame it as asset value, not as production. The client already paid to make the film. You are proposing to multiply what it can do. That framing supports a much better price than positioning it as an editing add-on.

Make additional languages the easy yes. Once one language is agreed, the marginal price should feel small enough that adding three more is not a new decision.

Put it in every future scoping call. Ask what markets they operate in during the initial conversation, before the project is even shot. Sometimes the answer changes how you shoot it, which is worth more to them than the localization itself.

Setting Expectations Honestly

Three things worth saying to clients before they hear them from someone else.

Synthetic voice is clear, natural, and appropriate for narration, explainers, training, and product content. It is not a performance. For a brand film that lives on emotional vocal delivery, say so and quote a human alternative. Clients remember who was straight with them.

Output quality depends on input quality. Overlapping speech, background noise, heavy accent variation, and unscripted rambling all degrade results. Clients who know this beforehand read an imperfect result as a source issue rather than your failure.

Some content should not be localized. Content with a short shelf life, content dense with untranslatable references, or content aimed at a market the client has no ability to serve. Advising against a sale occasionally is what makes your recommendations credible the rest of the time.

Protecting Yourself Contractually

Solo operators carry risk that larger firms absorb structurally, and localization introduces a few exposures worth addressing in your terms rather than discovering later.

Define the deliverable precisely. Number of languages, output formats, whether graphics localization is included, how many revision rounds, and what constitutes a revision versus a new request. Vague scope on a multi-language job compounds across every language.

State the review position explicitly. If linguistic review is not included, say so in writing and say who is responsible for it. This is the single most likely source of a dispute, and a sentence in the quote prevents almost all of them.

Fix the source at transcript approval. Make client approval of the source transcript a contractual gate, with changes afterwards chargeable. Without this, a client who rewrites a line after you have generated six languages will expect the rework for free.

Disclaim on client-supplied terminology. If the client provides product names and technical terms, their accuracy is the client's responsibility.

Address rights briefly. Confirm the client holds the rights to the content and to any music in it, and that they are authorising translated distribution. You are not in a position to verify this and should not carry the risk of it being wrong.

Cap your liability at the value of the engagement, as standard practice for a solo operator.

None of this requires an expensive lawyer. A page of clear terms attached to every quote resolves the overwhelming majority of problems that arise in this work.

Frequently Asked Questions

Do I need to speak the target language to deliver this work?

No. Technical QC — sync, levels, timing, character rendering, graphics fit — is language-independent and catches most real defects. For linguistic quality, either price in a native-speaker reviewer or state clearly that linguistic review sits with the client. Both are professional positions; silently skipping review is not.

How should I price it if I normally charge day rates?

Switch to per finished minute, per language, with a separate setup fee. Localization time is mostly waiting, which you cannot bill hourly. The setup-plus-per-language structure also matches your actual effort curve and makes additional languages an easy yes for the client.

What if the client's original project has no separated music and effects?

Automated dialogue separation can recover a usable bed from a full mix. It is not identical to a proper stem under close listening, but for corporate, training, and explainer content it is entirely serviceable. Flag the limitation during quoting if the content is premium brand work.

Which languages should I offer?

Start with the ones your existing clients actually need, which is usually a short list. Do not build a menu of thirty languages you have never delivered. Depth in three or four languages where you have reviewers you trust beats breadth you cannot quality-check.

Should I tell clients the voice is AI-generated?

Yes, in writing, in the quote. It is a defining property of the deliverable, some clients have their own disclosure obligations, and finding out later damages trust badly. Making it explicit in your tier descriptions turns it into a clear feature of the offer rather than an awkward revelation.