Chapter 1. The Business Model Decides How You Produce#

1.1 One story, several different businesses#

The first thing written down on a microdrama should not be character names. It should be how the project completes an exchange of value. Why is the audience allowed to keep watching? Why is a platform willing to distribute it? Where does the producer's money actually come from?

At least five distinct products hide under the single word microdrama: series sold by episode or as a bundle; free series funded by advertising; dramatized advertising made for a product or brand; content commissioned by a platform or rights holder; and motion comics or narrated comics built mostly from novel adaptation and voice-over. They can share subject matter. They cannot share rhythm, cut points, assets or metric systems.

The most common greenlight error is finishing the story first and looking for monetization afterwards. That leaves a structural misalignment nothing downstream can repair. A gentle coming-of-age story may complete well and still have no boundary sharp enough to charge at. A paid series that cuts at the most painful moment of every episode, converted directly into a free ad-funded version, may simply feel like endless teasing until the audience quits.

1.2 The paid route: managing withheld satisfaction#

The core of a paid series is not "as much suspense as possible." It is that the audience has accumulated enough emotional investment, and that a specific, imaginable, imminent satisfaction is sitting right in front of them. The moment before payment is what this system calls the paywall boundary.

An effective boundary has five components: a visible event, a specific question, emotion already accumulated, an answer about to arrive, and a satisfaction that is deliberately withheld. The boundary in E006 of Backlit Takeover is not the abstract promise that a secret about her parentage is coming. It is that the DNA report is already on the banquet screen, the result line is about to display, the lights are cut — and Lin Yun receives a live photo of her mother being held.

The audience is waiting on two answers at once: whether Lin Wei is the false heir, and whether Lin Yun's mother is still alive. The first delivers a public reversal. The second escalates the stake from status to family. The desire to pay comes from an event that has started and has not finished.

Paid design can also destroy trust. If every episode cuts before the answer and nothing is ever paid off, the audience works out that the show is only stalling. So each episode should repay one old debt and incur a larger new one. E007 must respond to the blackout and the photo immediately — it cannot change location and spend fifty seconds re-establishing.

A paid project records these at greenlight: the expected first-paywall episode, the boundary type, the evidence that must accumulate before it, how the first shot after payment repays the debt, and where the larger payoffs sit. A cut point is not an opening the editor hunts for once the script is done.

The core ledger for a paid project#

monetization:
  mode: paid_unlock
  first_boundary_episode: E006
  boundary_type: proof_interrupted_by_personal_danger
  visible_event: DNA result about to appear on the banquet screen
  accumulated_emotion: Lin Yun has absorbed repeated public humiliation and obtained evidence
  withheld_satisfaction: public confirmation of Lin Wei's real identity
  added_threat: live photo of Lin Yun's mother
  next_episode_repayment: prove when the photo was taken, then trace the location

1.3 The IAA free route: managing retention and ad tolerance#

In the IAA model the audience pays with attention rather than money. The work needs enough watching and enough return visits to create ad opportunities, without letting the ads and the breaks destroy the experience. It still needs suspense, but it does not have to stack every satisfaction behind a single paywall.

The IAA version usually benefits from more within-episode payoff. Lin Yun does not just produce the authorization; she uses it in the same episode to obtain access to the books. The closing question — why is the payee her dead father — arrives at the end. The audience got a status reversal in this episode and is willing to keep watching for the new question.

The danger in the free model is adding episodes mechanically. More episodes do not automatically mean more effective ad impressions. If the middle stretch repeats the same humiliation and the same comeback, completion and next-episode rate fall together. At greenlight, state what each story segment changes in the relationships, which secret it opens, and what new cost it introduces — rather than filling runtime with repeated events.

When one project ships both a paid and an IAA version, do not simply change the export settings. They may need different episode-end positions, different lengths, different recaps, different ad breaks and different payoff density. The safest method is to maintain the story at the beat layer and compile it into different episode arrangements.

1.4 The branded route: the story must manufacture a reason to act#

Branded microdrama is not a product placed inside a popular story. Real brand integration requires the product, service or brand value to participate in causality: if you removed the brand, the way the character solves the problem, the way a relationship changes, or the audience's understanding should all be affected.

There are at least three depths of integration. The first is environmental presence — the product is set dressing. The second is functional participation — the product helps a character complete an action. The third is value participation — the brand connects to a character's choice or identity. The deeper the integration, the more legal review, fact-checking and client approval it requires, and the less it reads as a crude insertion.

Suppose Backlit Takeover is serving an enterprise finance platform. The crude placement is the logo appearing on the boardroom table. Functional integration is Lin Yun using it to trace the anomalous payment. Value integration is when "every modification leaves an audit trail" becomes the mechanism by which the story exposes abuse of power. The third makes the product's function and the revenge mechanism the same event.

The commercial boundary for branded work is usually not payment but memory and action. The project must state whether it is raising awareness, explaining a feature, changing an impression, generating leads, driving purchase, or serving existing customers. Different goals determine how often the product appears, where, whether an explicit CTA is needed, and which data can measure any of it.

1.5 Commissions and submissions: study the buyer's production interface first#

When a platform or production company takes submissions, it is not only judging whether the story is good. It is judging whether the story fits its existing production and distribution machinery. Typical materials include character biographies, a synopsis, the first five to ten episodes, the paywall cut points, total episode count, and information about the creator. Buyers differ on episode length, total count and required IP status.

So writing one universal script and sending it everywhere is inefficient. Keep one content master and generate a different submission pack per buyer. The master holds the complete facts; the pack changes only format, length, sample range and commercial markers. Never maintain divergent versions of the story's truth across several documents.

Submissions also require protecting information. With an unverified counterparty, send the outline and partial script first. The full script, the central reversal and the editable assets go out only once identity, purpose, payment and confidentiality are settled. Registration, version timestamps, correspondence records and licence documents are not post-deal administration — they are a chain of rights that starts at greenlight.

1.6 Motion comics and narrated comics: what narration removes, and what it adds#

Motion comics and narrated comics reduce live-action shooting, and narration lets you compress time, exposition and non-essential action — which suits adaptations from long source material and high-concept subjects. But "it has narration" does not mean it is easier to produce. It moves the pressure onto voice consistency, the refresh rate of the image, asset reuse, and the provenance of the rights.

Generate a long narration sentence by sentence and the timbre and tempo will drift. Illustrate mechanically per sentence and it becomes a slide deck. Without stable character assets, the same person can change face every few seconds inside one passage. And adapting a novel without a licence can expose far greater risk precisely when the project finally gets traffic.

At greenlight a narrated comic must decide what the narration is: neutral storytelling, a character's own account, commentary with attitude, or an unreliable narrator. That decision drives the rhythm of the prose, the voice casting, the relationship between image and language — and how easily you can produce other-language versions later.

1.7 Use a matrix to decide the production route#

Fill this in before any script is generated:

If a project chooses two models at once, write down which is primary and which is secondary. You cannot say you want paid conversion, and premium brand feel, and organic spread, and ideally a platform sale — and then refuse to rank the conflicting goals. A showrunner can only make useful trade-offs when the goals have a priority order.

1.8 SOP for choosing the business model#

First, write the final value exchange without intermediate words like traffic or impressions. Second, write the primary action the audience must complete. Third, fix the primary metric for that action and two guardrail metrics. Fourth, decide what the story structure must add in order to produce that action. Fifth, list the viewing experience that structure is most likely to damage. Sixth, define the pilot scope and the stop conditions.

For the paid trial of Backlit Takeover the primary metric is unlock conversion at the first boundary; the guardrails are E001 completion and post-payment E007 completion. If conversion is high but E007 sheds viewers badly, the cut point successfully induced payment without repaying the promise in time. That is not a healthy win.

commercial_decision:
  primary_mode: paid_unlock
  secondary_mode: IAA_recut
  primary_action: unlock_E007
  primary_metric: first_boundary_conversion
  guardrail_metrics:
    - E001_completion
    - post_unlock_E007_completion
  pilot: E001-E007
  stop_conditions:
    - three different hook families all fail to hold the first three seconds
    - viewers understand the conflict but stay indifferent to both the mother and audit promises
    - continuity repair cost exceeds the pilot budget ceiling

1.9 Common failures and the fault tree#

Symptom: high ad click-through, heavy drop-off in episode one. First check whether the ad promised a subject or relationship the episode does not deliver immediately. Then check whether episode one spends too long explaining backstory. Check image quality last. Do not reflexively regenerate a prettier opening shot.

Symptom: decent completion, low conversion at the paywall boundary. Check whether the boundary question is specific, whether the audience is already invested, whether the withheld satisfaction is worth paying for, and whether the post-payment promise is credible. If the audience is merely calmly aware that another episode exists, adding a hard cut to black will not manufacture desire.

Symptom: low brand recall. Check whether the brand genuinely participates in causality, whether the key function is understood at the point of peak attention, and whether the brand name is drowned by story information. Adding more inert logo exposure usually just increases irritation.

Symptom: the narrated comic produces fast, but completion is falling. Check whether the narration holds one unvarying tempo, whether the visuals are merely per-sentence illustration, whether information or composition changes every three to five seconds, and whether the characters lack a recognizable state. The problem is likely in the audio and visual beat, not in the synopsis.

1.10 Checklist, exercises and deliverables#

A greenlight review must answer: where the revenue or value comes from; what the audience's primary action is; where in the story that action is driven; what the primary and guardrail metrics are; which kind of resentment the structure is most likely to provoke; which versions must be produced; and whether the stop conditions were written down before a large generation budget was committed.

Exercise: take one story — a wrongly accused nurse uncovers a hospital's secret — and rewrite it as four projects: paid, IAA, branded and narrated comic. One page each. But each version must change the first-episode payoff, the closing question and the core metric. Changing only the project name does not count.

Deliverables for this chapter: the business model decision table, the value exchange statement, the metric tree, the version matrix, the first commercial boundary hypothesis, and the pilot stop conditions.

A note on sources#

The observations here about submission materials and the ranges for episode length and episode count come from readable Zhihu sources recorded for this project. The judgments about the difference between paid and ad-funded acquisition, about creative and cut-point combinations, and about popularity not equalling profit come from the in-depth Xiaohongshu corpus. Specific platform rules change, and a real project must re-verify the current official requirements. This chapter does not treat earnings figures in creator posts as generally reproducible promises.