Many brand films look crowded but contain only one active character. The lead speaks while everyone else nods. The lead picks up the product and the room smiles on cue. When the decision is made, the supporting cast applauds. The frame is busy, but the relationships are inert.
A supporting character should not simply fill space. They should make the lead's choice visible. The person who offers a crucial object, resists an easy answer, redirects an awkward moment, or completes an exchange can change the story. Ideas such as trust, collaboration, curiosity, and care become credible through those exchanges—not through background smiles.
A supporting character is an action inside a relationship
Remove a supporting character from the script. If the lead still completes the same action in the same way, that character is probably set dressing.
A useful supporting role changes at least one thing: what the lead does next, how the audience reads the lead, why the product appears now, or where the scene's emotion turns. That change does not require dialogue. A delayed look can create resistance. Moving a chair can express care. A joke that nobody catches can reveal distance.
This is different from deciding how many seconds belong to people and how many belong to brand information. The more revealing question is: who does what to whom, and what changes because of it?
Give each supporting role a different verb
Ensemble scripts often confuse identity with dramatic function. The character list says colleague A, colleague B, customer, and family member, but it does not say what any of them moves forward.
Assign each supporting role a relationship verb. One person initiates by presenting a problem, object, or invitation. Another resists, making the decision cost something. A third redirects with information that changes the lead's approach. Someone witnesses, using a precise response to help the audience read the result. Another completes the action by turning it into an exchange between two people.
One person may perform more than one function, but a shot rarely improves when everyone performs the same one. When a lead opens a new product, four people leaning in with identical surprise feels choreographed. A more credible pattern might be: one person notices the sound first; one continues working; one hesitates before moving closer; one clears space. Different reactions make the group feel like a relationship, not a row of emotion icons.
Reactions need sequence, not synchronization
A reaction shot does more than respond to the lead. It labels what an event means. If one person worries, another is curious, and a third catches the object confidently, the audience receives three different signals: risk, novelty, and competence.
Build a reaction order: who notices first, who misunderstands, and who confirms the outcome last. The edit does not need to visit every face. It should preserve the reactions that change interpretation. A smile that arrives too early can dissolve tension; a glance that arrives half a beat late can clarify the turn.
Supporting performers also need permission not to react. In lived spaces, people do not complete a full expression for every event. One person keeps clearing the table, another pauses briefly, and a third looks toward someone else. Those unsynchronized actions make the lead inhabit a world where other lives continue.
Put the product inside the exchange
The One Club's 2026 page for Apple's A Critter Carol describes the iPhone 17 Pro as a supporting character. It is not a trophy waiting to be admired. The characters discover it, pass it around, and use it to make a music video together. The useful lesson is not to make the product visually small. It is to let the product participate in a relationship.
D&AD's retrospective on Gatorade REPLAY offers a related principle: the brand was built into the organizing idea rather than attached after the story. A brand-film script can test product presence with one causal sentence: because this person hands it to another person, the second person can now make a different choice.
The product still needs clear, accurate, well-crafted coverage. The difference is that the hero shot becomes the result of an exchange rather than a pause imported from outside the story. The audience understands what the product does within the relationship before studying how it looks.
Block the ensemble so relationships remain readable
When the cast grows, coverage often collapses into two options: squeeze everyone into a wide shot or cut through a sequence of close-ups. Better blocking establishes spatial direction first and then lets relationships alter it.
Give the lead a readable action line: where they enter, whom they face, and which direction the object moves. Place supporting characters at different distances. A nearby person can assist or block. A middle-distance character can exchange a look. A far character can express attitude by approaching or staying away. Even a two-second cutdown can then reveal who matters to whom.
Do not give every supporting role equal facial coverage. Prioritize three kinds of images: hands and bodies that change the action, eyelines that redistribute attention, and spatial changes after the exchange is complete. These action nodes are also easier to restructure for shorter formats than a collection of isolated expressions. ONCE's article on pacing in brand films extends that discussion from performance into edit structure.
For AIGC ensembles, decompose the relationship first
Multi-character generation becomes difficult for more than numerical reasons. Every additional subject adds identity, position, eyeline, occlusion, and interaction-object continuity. The 2026 primary paper MultiAnimate frames the problem around per-person identity, spatial features, and inter-person relationships. That helps explain why a prompt such as "several people interacting naturally" gives the system too little structure.
A more controllable hybrid workflow breaks the relationship into short action units: A hands an object to B; B does not accept it immediately; C watches B rather than the object. Live action or controlled references can secure the key exchange. Generative tools can extend environments, create transitions, or produce shots that do not carry critical causality. When the ensemble itself must be generated, lock character references, positions, and each beat's action target, then inspect hands, eyelines, and object ownership shot by shot.
Technology can expand a setting. It cannot decide who changes whom. If that relationship is vague in the script, richer images only create more places for attention to go wrong.
Start with a relationship-action table
Before storyboards, write one line for every principal character: what they want, what they do to the lead, what changes in the lead, and what role the product plays in the exchange. If a line says only "stands nearby," "watches happily," or "creates atmosphere," either give that character a real action or remove them.
Then rehearse the scene on a phone without lighting or art direction. Watch only the action chain. Is the initiation visible? Does the resistance feel real? Does a person cause the turn? Does the product arrive when the relationship needs it? A relationship that works in an empty room is worth taking into production.
If you are developing a cast-heavy brand film, ONCE can shape the character functions, ensemble blocking, product actions, and the relationship between the master film and cutdowns. Explore ONCE's brand video work and bring the current script or character outline to the conversation.
