Choose a launch video length from the message

List what the viewer must understand: the situation, the product action, the result and the next step. Add a category explanation only if the viewer cannot interpret the demonstration without it. Those jobs need different amounts of attention depending on the interface and copy.

Consider an illustrative request-assignment product. The task is to show an incoming request receiving an owner. A viewer familiar with shared queues may need little setup. Someone unfamiliar with the category needs to understand why the owner field matters before a highlight around it becomes meaningful.

These are creative planning options, not measured optimal durations. All three preserve the same promise: make the next owner visible. Longer versions add explanation rather than repeat the short version more slowly.

• 15 seconds: 0–3 show an unassigned request; 3–8 select the owner; 8–12 hold on the assigned state; 12–15 display name and next action. Use very little copy.
• 30 seconds: 0–5 establish the handoff problem; 5–11 orient the queue; 11–20 demonstrate assignment; 20–25 inspect the result; 25–30 close. This adds context and a readable proof hold.
• 60 seconds: 0–8 establish the task; 8–18 show the starting conditions; 18–34 explain assignment; 34–44 show confirmation; 44–53 address one relevant limitation; 53–60 close. This supports a less familiar viewer.

Read the actual copy while looking at the planned screen. Mark the moment the viewer must stop following movement and inspect a label. A scene may need a longer hold even when narration is short. Conversely, a large, simple state change may need less explanation than a sentence describing it.

If the short version requires tiny text or hides an eligibility condition, do not solve the problem by accelerating everything. Remove a secondary message, move nonessential detail to supporting material or choose a longer outline. Preserve any condition needed to interpret the headline honestly.

Choose the shortest version that completes its intended communication job. Record the rejected alternatives: for example, the 15-second cut assumes queue familiarity, while the 60-second cut includes a permissions explanation unnecessary for the launch audience. That record makes later runtime requests easier to evaluate.