How to Write a K-Drama: The Structure of a 16-Episode Series, Pillar by Pillar

The shape in one table

This is a working model, not a rule; dramas vary it constantly. But if you map recent miniseries against it, the pillars are usually within an episode of where the table puts them, because the model comes from the economics of the form rather than from theory.

Why pairs: two episodes a week

Broadcast miniseries air two episodes a week on consecutive nights. The writer therefore thinks in pairs: the Monday episode sets, the Tuesday episode pays, and the Tuesday ending has to hold the audience for a week rather than a day. Structurally this means:

• Odd episodes open, even episodes close. The hook at the end of an odd episode is a question; the hook at the end of an even episode is a shock or a turn.
• Pairs are the real units. Eight pairs, each with an arc. Outlining sixteen episodes is easier as eight two-hour movements.
• The pillars land on even episodes. 4, 8, 12, 16. The week ends on the turn.

Streaming originals released all at once do not have the weekly gap, and their structures are looser, but writers trained in the broadcast form carry the pairing into them, and the binge viewer feels it as rhythm.

The ending-hook rule

Every episode ends on a hook, without exception, because the form's economics are weekly retention. The hook is a concrete moment, not a summary: someone standing at the door; the moment she learns; the object in the wrong hand. Writers' rooms and contest readers will judge a treatment on the list of sixteen ending moments before they read a page of script.

The hooks need variety. Sixteen cliffhangers of the same type (a reveal, a reveal, a reveal) numb the audience. Rotate: a reveal, a decision, a physical danger, an arrival, a line of dialogue that changes a relationship, a discovery by a secondary character, a flash-forward. The ending is also where the drama's signature image or line returns; the strongest finales echo the opening of episode one, and the echo is planned from the start. The prompt in our library that builds a sixteen-episode table asks for an ending moment per row for exactly this reason.

The pillars in detail

Episodes 1–4: premise and lock

Episode one is a premise pilot: it must establish the protagonist, the antagonist or opposing force, the world and the tone, and end on a hook that makes episode two necessary. Many dramas open with a cold open drawn from later in the story (a flash-forward to the crisis) before the title, a technique the TV pilot guide covers.

By the end of episode four, the central relationship is locked, meaning the two leads cannot walk away from each other, whether the relationship is romantic, adversarial or both. In a romance, this is the point where the audience knows they are the couple even if the characters do not. In a thriller, it is the point where the protagonist is committed to the case. The first reversal often lands here too: what looked like the situation is not.

Episodes 5–8: escalation and the midpoint

The middle of the first half is where the drama's pleasures live: the genre's promised scenes, the growing attraction, the case's early wins. The goal seems reachable. Then episode eight reverses it.

The midpoint reversal changes the goal, not just the odds. The protagonist wanted to win the case; now she wants to protect the witness. The couple wanted to keep it secret; now the secret is out and the question is different. In a romance, the acknowledgement of feeling (a confession, a kiss, a choice) commonly sits at or near the midpoint, which is why it is so often around episode eight; the structure puts it there, and the audience has learned to expect it.

Episodes 9–12: consequences and the worst crisis

The second half opens with the cost of the midpoint. Forces that were background become foreground; the antagonist gains; secrets that were the audience's become the characters'. Episode twelve is the low point: the separation, the arrest, the death, the betrayal that seems final. Broadcast writers sometimes call the stretch after this the hardest to write, because the audience has to believe the situation is unrecoverable for three more hours.

Episodes 13–15: the long fall and the truth

Thirteen and fourteen play out the crisis; characters act from what they believe, and what they believe is partly wrong. Episode fifteen is where the truth surfaces: the misunderstanding that drove the separation, the identity of the real culprit, the reason for the sacrifice. The reveal is the penultimate episode, not the last, because the finale needs room for the consequences.

Episode 16: resolution and echo

The finale resolves the plot line and the relationship line and returns, in some form, to the opening: the same place with a different meaning, the same line said by the other person, the same image with the missing person present. K-drama finales are often quieter than Western audiences expect; the climax was episode fifteen, and sixteen is the landing.

Two lines, one structure

Most dramas run an A line and a B line, and the genre decides which is which. In a romance drama the relationship is A and the plot (the company, the case, the illness) is B; in a thriller or a legal drama the plot is A and the relationship is B. Both lines hit the pillars, and the craft is in making one pillar serve both: the midpoint case reversal is also the moment the two leads have to trust each other.

Supporting couples and secondary plots run their own smaller arcs, usually resolving an episode or two before the leads' so the finale is clear.

Compressing to 12 or 8

Streaming originals commonly run 8 to 12 episodes. The pillars move proportionally and the pairing loosens:

Shorter series lose the escalation stretch (5–8 in the 16 model) most, which is why 8-episode dramas tend to be genre pieces with a strong plot engine and 16-episode dramas tend to be the ones with room for a relationship to grow. Choose the length by what your story needs in its middle, not by what is fashionable.

Outlining a drama with this model

1. Write the premise, the protagonist's goal and the genre in three lines.
2. Write the sixteen ending moments first, as concrete images, and check them for variety.
3. Place the pillars: what locks at 4, what reverses at 8, what breaks at 12, what surfaces at 15, what echoes at 16.
4. Fill each pair with an arc of its own.
5. Assign the A and B lines to each pillar.
6. Then, and only then, write episode one.

A model can draft the table from a premise and a goal, and its first version will be the probable one; push it for alternatives at each pillar. If you draft with AI, keep it at the structure and prose-treatment level; Korean dialogue needs a native writer, as the international writers' post explains.

Try it in SparkLore

SparkLore's prompt library includes a sixteen-episode structure prompt that builds the table above from a premise, a goal and a genre, with the pillars fixed and an ending moment per episode, and a cold-open prompt for the first ninety seconds of episode one. The free AI script generator writes a scene in Korean drama format, and the Studio holds the whole series with its cast in one bible and exports to HWP-compatible DOCX. Sign up to outline a full series; plans are on the pricing page.