Which subtitle work should stay in-house, and which should go out?

Two ways the same decision goes wrong

A title team with strong in-house editors can still miss a release window. The editors are not the problem; the calendar is. When a flagship series and a catalog refresh reach picture lock in the same fortnight, the same few people are expected to segment, translate, time, check, revise and package every language version, and no subtitle tool adds reviewer hours in the week they are needed.

The opposite failure is quieter. A team outsources everything, then spends release week answering questions only it can answer: how a character's name is rendered, whether a joke is adapted or kept, which cut the files were timed against. The invoice went down, and the title team's workload moved into its inbox, where nobody had budgeted for it.

Both teams asked whether subtitles should be in-house or outsourced, and both answered for the whole operation at once. The better question is narrower. For each stage of the workflow, does it need the title team's context, or can it be written down precisely enough for someone else to do and for the title team to check?

Map the slate before choosing a model

Start with the coming run of releases, not with a cost comparison. For each title, record the languages it ships in, the date picture and audio are locked, the platform delivery date, how likely late changes are, and which tracks are required. Translated subtitles, same-language captions and subtitles for the deaf and hard of hearing (SDH) serve different viewers and carry different content, so list them separately; the subtitle QC brief guide sets out why one file rarely serves all three.

Then read the map by week instead of by title. The shape that matters is concurrency: how many language versions have to be reviewed, corrected and exported inside the same window, against the review time the in-house team actually has in that window after its other work. As an example, a team that comfortably handles one weekly episode may face a dozen language versions of a flagship season locking within the same few days. That peak, not the annual volume, is what the operating model has to survive.

Mark the titles likely to change late. A stable catalog master and an episode still in the edit suite create very different work, whoever does it, because every picture change after timing means retiming, re-review and redelivery.

Split by stage, not by title alone

Subtitle production is several jobs that are easy to treat as one. Break it into the stages the team actually runs: creating the timed source template, translating into each target language, timing and conforming to the platform specification, quality control, revisions after picture changes, packaging and delivery, and final sign-off.

Some of those stages carry title context that is expensive to transfer. The style decisions for a show, character and place names, tone calls on humour or sensitive lines, and the judgement of whether a revision is good enough to ship sit most naturally with the people who own the title. Final sign-off in particular is hard to hand over without also handing over accountability for the release.

Other stages can be specified. Translation from an approved template with a glossary and style guide, conforming to a written platform specification, first-pass QC against that specification and format validation can all be defined tightly enough to be done elsewhere and checked on return. Owning the timed template in-house is often the pivot: it fixes cue boundaries once, and each target language is produced against the same structure. The trade-off is that one template forces every language into the same cue breaks, which does not suit every target language.

Three boundaries that hold up in practice

The first keeps the editorial core in-house, meaning the template, the style decisions and final sign-off, and sends target-language translation, timing and first-pass QC out. It suits teams with strong editors and more languages than they can staff.

The second splits by title. Flagship and fast-changing titles stay inside, where a late edit can be settled across a desk, and catalog work with stable masters goes out. It suits libraries being relaunched in new markets.

The third keeps steady-state volume in-house and routes peak overflow out under the same specification, so the outside team works to the rules the inside team already uses. It only works if that specification is written down before the peak rather than explained during it.

None of these is a default. The right boundary is the one that the slate map, the team's real review time and the access rights an outside team can be given all support.

Make the handoff exact before the peak arrives

Whatever leaves the building needs a defined route back. The title team should be able to say who owns timing changes after picture lock, who answers terminology and sensitivity queries, how a correction to one line is carried into every episode and language where it recurs, which file version is current, and what counts as accepted. Without those answers, outsourcing moves work rather than removing it.

Test the handoff on a sample episode before committing a season to it. A sample shows whether the specification is complete, whether queries arrive in a form the title team can resolve quickly, and whether returned files drop into the existing sign-off step without rework. A gap in the transfer costs far less to find on one episode than across a whole release.

This is the part of the decision MoniSa works inside. Subtitle work is scoped to the stages a team chooses to hand over and can start from the team's own timed template or from the source media. The specification is agreed on a sample episode, and QC runs as a pass separate from the subtitling itself, so returned files arrive for the title team's own sign-off rather than replacing it.

What to have on one page before choosing the split

Put these together before deciding which stages stay and which go.

• The coming slate by week, with languages, lock dates and delivery dates for each title
• The required track types per title: translated subtitles, same-language captions, SDH
• The review time the in-house team really has in each peak week
• The workflow broken into stages, each marked as needing title context or specifiable
• A named owner for queries, late picture changes and final sign-off
• A sample episode chosen to test the handoff before a season depends on it

Signs the current split is costing more than it saves

These point to a boundary in the wrong place, not to a failing team or supplier.

• In-house editors spend release week on timing and packaging instead of editorial calls
• Outsourced files come back with queries only the title team can answer, and nobody owns answering them
• After a late picture change, nobody can say which subtitle file version is current
• The split was set by budget line rather than by where decisions actually get made
• Flagship and catalog titles are handled identically despite carrying different change risk

What to send MoniSa for a scope discussion

The title team keeps its editorial and sign-off decisions; these let a scope discussion start from the real slate.

• The slate map, or the part of it being considered for handover
• The stages the team plans to keep and the ones it would send out
• The platform specification, house style guide and any glossary or character list
• A sample episode with its current picture-lock status and required track types
• Who answers queries and who gives final sign-off on the title team's side