After Five Minutes

Week 12 · SLOP8760

Phase 4 of 4 · The Episode

This week adds

What comes after the pilot: a curve to read, a bible to write, a calendar to hold, and a judgment about which of those scales.

Reading the curve

Shape matters more than the average. A steady curve with one sharp drop tells a different story than a slow, even decline ending at the same number.

What the platform actually counts

The curve is not one measurement. Each of these answers a different question, and only the first is about the opening.

Metric	What it counts
Intro	the share still watching after the first 30 s
Average view duration	average minutes watched, among engaged views

Source: YouTube Help, Measure key moments for audience retention — https://support.google.com/youtube/answer/9314415
Source: YouTube Help, Understand your YouTube engagement — https://support.google.com/youtube/answer/9313698

A season’s money arrives early

80%of total revenue, in the first month

Reported for Douyin’s own top-performing native micro-dramas, not for the market as a whole: 80% in month one, 15% across months two and three, 5% across four to six. A release calendar is not an administrative preference.

Source: NRTA reprint of Xinhua, 微短剧迎来新变化 (Feb 2025) — https://www.nrta.gov.cn/art/2025/2/28/art_3731_70276.html

The scope is in the sentence itself and is narrower than “the market”: “抖音原生头部微短剧的首月收入占总收入的80%，2-3月占比为15%，即使进入4-6月阶段，仍能保持5%的收入占比。” Douyin-native, top tier, on one platform. The shape of the curve is the transferable part; the exact percentages are not a number to quote about a show on any other platform, and certainly not about one made in this course.

A season is not just a pile of finished episodes; it’s a schedule someone commits to and a viewer can rely on. A release calendar states when each episode goes out and holds that date the same way an assessment’s due date gets held — a season that ships on an unpredictable schedule loses the audience a retention curve is trying to measure in the first place, regardless of how good any single episode is.

This slide is why. If four fifths of what a title earns arrives in its first month, the calendar is not paperwork around the work: it is most of the commercial decision, and a slipped episode is not a slipped episode, it is a slipped window. The same report notes the free, ad-funded model at about half the market by October 2024; the 2025 white paper puts that share at about 66.3%, which is the direction rather than a contradiction.

A “bible” is a written document containing the format, character descriptions and story lines or synopses for a number of episodes of a series.
WGA 2020 Basic Agreement, Art. 13.B.7.m

Source: WGA 2020 Theatrical and Television Basic Agreement — https://www.wga.org/uploadedfiles/contracts/mba20.pdf

The curve points, it does not explain

Every drop maps back to a decision made earlier. The platform names the shapes it reports — flat lines, gradual declines, spikes and dips — and it compares a video against others of similar length. It never says why a moment failed.

Source: YouTube Help, Measure key moments for audience retention — https://support.google.com/youtube/answer/9314415

What scales, and what doesn’t

The queue, continuity discipline and the gate scale close to linearly. Judgment does not. The form scales too: 124 micro-drama titles ran on satellite channels in 2025, up 254.3%, across 36 channels rather than ten.

Source: NRTA CVB, 2025年全国卫视频道微短剧收视报告 (Dec 2025) — https://www.nrta.gov.cn/art/2025/12/26/art_114_72199.html

Generation, continuity discipline, and the quality gate scale close to linearly: more episodes need more passes through the same queue, with the same roles, checked against the same reference material. What doesn’t scale the same way is judgment — deciding which hook actually fits a beat, deciding whether a drop on the curve is a pacing problem or a one-off, deciding what belongs in the bible and what was a one-episode accident. Adding more people to a queue multiplies throughput; it does not multiply the judgment behind any one decision, which stays a single person’s job no matter how many episodes are in production.

The number on the slide is the form arriving somewhere it did not used to be: broadcast television. Cumulative viewing across those titles is reported at 2.81 billion viewings. A form that started on a phone is now scheduled against a channel’s evening, which is a scaling problem of exactly the kind this slide is about.

Before class

Bring the finished pilot and one specific guess, based on its beat sheet, at where its own retention curve would drop.

This week’s exercise

Write a one-page account of the semester’s process: what you’d carry into the bible for a second episode, one change the retention-curve reading would justify, and what you now believe does or doesn’t scale past a single episode. Bring the finished pilot and this account to Wednesday’s Dailies.

Reading

YouTube Help: measure key moments for audience retention
YouTube
Intro tells you what percentage of your audience still watched your video after the first 30 seconds
YouTube Help: understand your YouTube engagement
YouTube
Average minutes watched among those who stayed to watch
TikTok Creator Academy: TikTok Analytics
TikTok
average watch time, watched the full video (percentage of viewers who watched the entire video)
WGA 2020 Theatrical and Television Basic Agreement
Writers Guild of America2020
A 'bible' is a written document containing the format, character descriptions and story lines or synopses
NRTA CVB: 2025 satellite-channel micro-drama viewing report
NRTA CVB2025
Linked, not quoted: the counts come from the report's tables rather than a sentence.

After Five Minutes#

Week 12 · SLOP8760#

This week adds#

Reading the curve#

What the platform actually counts#

A season’s money arrives early#

The curve points, it does not explain#

What scales, and what doesn’t#

Before class#

This week’s exercise#

Reading#