The Full Video Production Process Step by Step

The complete path a commercial video takes from first conversation to final delivery, what happens at each stage, and who has to approve what.

Commercial video follows a process that has been stable for decades, and the arrival of AI generation has changed the middle of it without changing the shape. Understanding the whole path is useful for a client because it shows where decisions are cheap and where they become expensive, and it explains why a studio pushes hard for approvals at moments that feel premature.

The first stage is the brief, and it is worth more attention than it usually gets. A workable brief answers who the audience is, what they currently believe, what they should believe afterwards, where the video will be seen, how long it can be, what must legally appear, when it is needed and who signs it off. That last item causes more delay than any other. A project with three unnamed approvers discovers them one at a time, each with different opinions, usually after the edit is locked.

The second stage is concept and script. A concept is the idea that makes the message memorable, and a script is the specific execution of it in words and time. For a sixty second film the script is roughly one hundred and forty to one hundred and sixty words of voiceover, which surprises clients who arrive with four hundred. Cutting happens here or it happens later at greater cost. Approving a script means approving a runtime, and a client who approves a long script has approved a long film.

The third stage is the storyboard or shot plan, which converts the script into images and timing. Each panel names what is on screen, how long it holds, what the camera does and what the audience should understand. This is the most valuable document in the project because it is the last point at which changes are nearly free. Moving a scene on a storyboard costs an afternoon. Moving it after rendering costs weeks. Storyboarding also exposes problems that a script hides, such as a line of narration that has nothing to show against it.

The fourth stage is pre production logistics, which runs in parallel: casting, location recce, permits, scheduling, equipment, catering, insurance and the call sheet. For AI and 3D work the equivalent is asset preparation, look development and the still frame exploration that fixes the visual language before motion begins. Either way this stage is where a project is quietly won or lost, because problems found here are administrative and problems found later are creative emergencies.

The fifth stage is production. In conventional work that means the shoot days, which are intense, expensive and unrepeatable. In generative work it means the iterative building of shots against the approved boards, which is less dramatic but no less disciplined. Both share the same rule: capture more than the edit needs, because the edit will need something nobody predicted. Research on hybrid project management by Mirzaei et al. (2025) makes a point that applies directly here, that methodologies work best when customised to the specific project rather than applied uniformly, which is exactly the judgement a producer exercises when deciding how much structure a given shoot needs.

The sixth stage is the edit, and it happens in a defined order for good reasons. The offline edit establishes structure and timing with unpolished material, because it is pointless to grade a shot that will be cut. Once the structure is approved, the picture is locked. Then sound design, music, voiceover, colour grade, graphics and subtitles are applied to a timeline that will not move. Clients who request colour changes before picture lock are asking for work that will be thrown away, and a studio that agrees is not doing them a favour.

The seventh stage is review, and it is the stage most damaged by informality. Feedback should arrive consolidated, in one document, with timecodes, from one named person who has already collected internal opinions. Feedback that arrives as separate messages from separate stakeholders over ten days produces contradictory instructions and burns the revision allowance without improving the film. The studio's obligation is to state clearly what a round means; the client's is to run their internal process before sending it.

The eighth stage is delivery, which is more than exporting a file. A finished project usually needs the master at full quality, platform specific versions at the right aspect ratios and durations, a silent version, a subtitled version, sometimes a version per language, thumbnails or cover frames, and the project archive. Agreeing this list at quotation stage rather than at the end prevents the familiar situation where the film is approved on Friday and six unbudgeted versions are requested on Monday.

The ninth and least practised stage is measurement. Someone should look, four weeks later, at what the film actually did: watch time, completion, enquiries, sales conversations, whatever the brief said it was for. Kim et al. (2025), studying playback interactions in mobile video viewing, found that how viewers actually behave during playback carries information that headline view counts obscure. That is the argument for looking at retention rather than reach, and for feeding what is learned into the next brief instead of starting the next project from zero.

References

Mirzaei, M., Mabin, V. J., & Zwikael, O. (2025). Customising hybrid project management methodologies. Production Planning & Control, 36(9), 1188–1205. https://doi.org/10.1080/09537287.2024.2349231

Kim, E., Oh, S., & Park, S. (2025). An empirical study of user playback interactions and engagement in mobile video viewing. IEEE Access, 13, 78272–78289. https://doi.org/10.1109/ACCESS.2025.3566402