Your brand receives two proposals for 12 product videos from production partners in China. One includes 12 distinct concepts. The other develops two concepts into three aspect ratios and two languages. Both can produce 12 files; they are different purchases.
ONCE recommends defining the deliverable unit before comparing its price. Separate the creative assignment, production requirements, version set and acceptance conditions. A file count alone cannot describe the work or promise a business outcome.
A commercial shift, with a defined evidence base
On September 29, 2026, WFA published a summary of research with Agency Mania Solutions based on responses from 69 multinational companies. It reports widespread fixed-fee or output-based arrangements in production, with variation by discipline and region.
This is not a census of Chinese production suppliers or proof that hourly pricing has disappeared. The public summary was checked on November 2, 2026; the members-only report was not reviewed. Its expectation that AI will accelerate commercial change is a prediction, not an established industry-wide result.
The useful buying question is what replaces time as the description of the work you intend to receive.
Define output separately from outcome
TrinityP3’s output-pricing discussion distinguishes a priced deliverable from fees linked to outcomes such as leads or sales. For video procurement, an accepted creative package and a promised sales result are different commitments.
A film can serve a commercial objective without its production agreement including media buying, pricing, inventory or sales follow-up. If performance incentives are part of the engagement, define their metric, data source, observation period, responsibilities and adjustment conditions separately. No universal incentive percentage follows from this article.
ONCE’s deliverable unit card
This proposed buying method attaches a short definition to each priced item. It is not an industry rate card.
| Field | What to specify | Ambiguity removed |
|---|---|---|
| Creative assignment | Product, audience, customer question and number of agreed directions | Whether alternate openings represent distinct concepts |
| Production requirements | Existing footage or new filming, talent, locations and demonstrations | Whether another version includes another shoot |
| Version set | Duration, aspect ratios, languages, captions, voice and file list | What the quantity actually counts |
| Use and acceptance | Purpose, usage permissions, approver, revisions and acceptance criteria | Whether a returned file is usable and what changes are included |
Buyers can compare like-for-like scope. Suppliers can explain why a new demonstration, actor or creative direction changes the assignment. The card does not require every supplier to charge the same price.
Count a hypothetical version tree
Assume two independent concepts, each supplied in 16:9, 9:16 and 1:1, in English and German. If those three formats include the required master specifications, the package contains 2 × 3 × 2 = 12 files. Do not add two masters again outside that count.
This is scope arithmetic, not an ONCE package, client case or cost benchmark. It describes two creative assignments and 12 files. It establishes neither that only two fees are appropriate nor that every file deserves a new-concept price. Reframing, copy adaptation, voice work and language review can be substantive tasks.
If the German market needs a different product promise or use case, assess whether it introduces new creative or filming work. Calling it “translation” does not resolve that question. Conversely, an already agreed export specification should not silently become another story-development charge.
AI changes execution; the comparison still needs a stable unit
Agree acceptance and selection before scaling generation. Multiple model candidates are not automatically multiple purchased deliverables. Equally, output pricing need not mean unlimited generation until every stakeholder likes a result.
Specify the candidates included within an approved direction, how feedback is consolidated, which corrections are included and which changes require a revised fee and timetable. A supplier’s failure to follow confirmed product facts differs from a client’s later change to the product promise.
Production teams still need internal cost and time records for planning, filming, selection, correction, review and handoff. A fixed output price does not remove the need to understand delivery economics. Those internal records and the client-facing acceptance unit serve different purposes.
Pilot the unit before turning it into a subscription
Ask a shortlisted production partner to price a small, real batch using the unit card. Confirm the concept count, filming basis, file list and acceptance process. After delivery, compare actual work and changes with what the card described.
Where a China-based manufacturer supplies the product facts, name its approver before scripting. Where your market team adds a claim, language or usage requirement, record the change before commissioning adaptations. This keeps product decisions from arriving as unexplained edits at final review.
Use the records to decide which work can be priced in advance and which still needs a specific estimate. A recurring need may support a service menu; a highly uncertain assignment may need a different arrangement. No single compensation model suits every brief.
For ONCE’s move toward productized and managed content services, the valuable asset is a comparable definition of work. Accumulated records could improve pricing and capacity planning; no measured margin or scale gains are claimed here. Explore ONCE’s video production articles for related planning decisions.
Sources and limits
The WFA public summary was published September 29, 2026. The TrinityP3 page carries date markers for May 27, 2025 and March 21, 2026. Both were checked November 2, 2026. The unit card, version tree and procurement recommendations are ONCE’s analysis, not findings attributed to either source.
