Why five
One episode tells you almost nothing, because a single upload is mostly a measurement of algorithmic luck. Twenty tells you a great deal and costs twenty episodes to find out.
Five is enough to establish a world, a cast, a question and a release cadence, and to see a shape in the numbers across days rather than a spike. It is also small enough that being wrong is survivable, which is the property that makes it a test rather than a commitment.
- Three-second hold on each episode
- Per-episode completion
- Episode-to-episode conversion, which is the number that pays
- Pilot-arc completion across all five
- Return rate across release days
- Language-track comparison on the same episode
- Whether the recurring characters are actually recognised
Plan the season. Produce the pilot.
The order matters and it is frequently got backwards. Beat-sheet the whole season, then produce five, because the five have to plant what the thirty pay off. Beat-sheeting only the five produces a pilot that works and a season with nowhere to go.
The corollary is that the plan has to survive being cut short. If the arc only works when everything is produced, you have removed your own ability to revise or stop.
- 01Beat the whole arc
In less detail toward the end, but end to end.
- 02Produce five episodes under full governance
Same gates, same canon, same rights and delivery discipline as a season. A pilot produced casually tests nothing about the season.
- 03Publish on a cadence
Return rate across release days is not measurable if everything appears at once.
- 04Read the curve, not the average
Where people leave is actionable. How many left is not.
- 05Decide at the gate
Continue, revise, or stop - with the season batch authorized only on the first.
Stopping is a good outcome
Frequently asked questions
What if my pilot arc does not meet the gate?
Then you revise or stop, and both are legitimate. Revising usually means the hook placement, the episode length or the cliffhanger type rather than the premise - which is the kind of thing a curve tells you and an opinion does not.
Can I skip the pilot and commission a season?
The gate exists to protect you from that decision, and we would rather not take the engagement than take it knowing the evidence has not been gathered. Our own commercial units are structured this way for the same reason.
How long should the pilot run before deciding?
Long enough to see return rate across several release days, which means a cadence rather than a single drop. The specific window depends on your audience and your platform.
Does the pilot cost extra compared to the season?
The pilot arc is a commercial unit in its own right, followed by a season batch after the gate. Pricing is scoped per production and we do not publish it, so the honest specifics come from a scoping conversation.
Do branded campaigns have a pilot arc too?
They have a paid story-campaign pilot, which serves a similar function against a campaign metric agreed in the brief rather than a retention gate. The structure is deliberately parallel.
