Pitch a Sketch Series Without Pretending It Has a Sitcom Engine
Pitch a Sketch Series Without Pretending It Has a Sitcom Engine
A sketch-series pitch usually fails in one of two directions. The first is a list: nine premises, no relation between them, each one funny enough on its own and none of them belonging to the same show. The second is an apology wearing a plot — a shared apartment, a boss, a season arc — bolted on because the writer suspects that disconnected pieces won't read as a series at all.
Both failures come from the same missing answer. A sketch show holds together because of what its comedy notices and how an episode is assembled. Neither of those is continuity.
So the pitch has two jobs. It has to name a selection principle sharp enough to rule things out, and it has to make an episode imaginable — not a story, but a stretch of viewing with a shape to it. Everything below is about those two jobs, and about why borrowing a sitcom's machinery gets in the way of both.
Recurring performers, recurring characters, a host, a framing device: all permitted. None compulsory. The question is never whether continuity is allowed. It's whether the continuity you're promising is the thing your comedy actually runs on.
A sitcom engine answers a question your show isn't asking
A sitcom engine is a situation that keeps producing trouble plus characters whose history accumulates. That's a real achievement. It means episode six lands differently because of episode two, and the writing room has a generator it can crank whenever it needs a story.
A sketch renews inside the piece. The piece arrives, does its thing, and ends. Nothing carries to the next one except what the audience has learned about how this show operates. That isn't a deficiency to be repaired. It's a different fuel, and it burns hotter per minute.
If you import the sitcom engine anyway, you're making two promises you don't intend to keep. First, that the audience will be rewarded for remembering. Second, that the next sketch has to fit a world — one apartment, one office, one set of relationships. A premise that would have been excellent on its own now has to be routed through people who have a shared past. You've traded range for the feeling of legibility, which is roughly the trade a pitch should never make.
Name what the show notices
"This one is dark and absurd and grounded" tells a reader almost nothing. Those are adjectives about the surface. A selection principle predicts. Given a new premise, it tells you whether the premise is yours or somebody else's.
Test a candidate principle by writing its rejection. If you cannot say what the show would turn down, you don't have a principle yet, you have a mood board. "Workplace absurdity" rejects almost nothing — cubicle, hospital, spaceship, all fine. "The paperwork that outranks the person" rejects a great deal, and the things it rejects are exactly the things that would blur the show.
Here's one to work with, invented for this article: small systems that demand absurd performances of certainty. It selects counters and forms and instructions and waiting rooms and helplines — any place where a small rule asks a person to sound sure of something nobody can actually verify. It excludes, say, a dinner-table family argument with no system in the room. That exclusion is the principle working.
The perspective is not a premise. It's the layer above the premises, and it's what lets a reader of your pitch supply a sketch you didn't write and get it right.
Give each sample its own comic action
A premise and a comic action are different things, and pitches routinely confuse them. The premise is the situation. The comic action is what the piece does with it — the behavior it exaggerates and the direction things move.
Take the invented series "Proper Channels," built on that perspective: two sketches and a recurring host.
Window Four. A service counter. A visitor has a printout confirming an appointment. The screen behind the counter does not show the appointment, and the attendant cannot accept the printout as proof because it lacks a counter stamp. The stamp is issued at Window Two. Window Two requires an appointment. As the exchange continues, the visitor's manner tightens into the attendant's own register — brisk, procedural, certain — while the attendant's confidence drains away, until the visitor is the one asking the verification questions and the attendant is the one answering them. The appointment is never proven. Nothing is resolved. What has moved is who sounds official.
The comic action: certainty transfers without any fact changing.
How to Open a Box. A silent instructional piece, fixed camera, hands and a box. On-screen labels appear, then subdivide. Lid. Lid seam. Seam edge, left. Seam edge, left, upper. Upper seam edge, interior-facing. Each label is more exact than the last, each one points the hands to a smaller area, and the precision itself is what stops the work. The vocabulary expands and the action contracts until the hands are resting on a fully labeled, entirely unopened box.
The comic action: exactness becomes the obstacle.
The host. A recurring figure in short beats between pieces, who introduces each one by confidently describing the wrong procedure — announcing the correct way to reseal a box before the desk scene, explaining how to request a stamp before the box scene. The confidence never wavers. That's the recurring element, not the information, which is reliably wrong. The gap between what he says and what follows is the show's thesis in miniature, running on a thirty-second timer.
Three mechanisms: a reversal, an accumulation, and a fixed form with mismatched content. The range isn't in the settings — a counter and a table aren't far apart. It's in what each of them does to its audience. One asks you to track who has authority. One asks you to read faster than the frame can fill up. One asks you to hold a mismatch in your head and wait.
Notice that only one of the three runs on setup, escalation, punchline. The silent piece has no punchline; it has a failure to complete. If you require every sketch to land the same way, you'll throw away the ones that give an episode its texture.
And notice the failure mode this section is guarding against: a fourth premise in which a different desk demands proof of the receipt that proves the receipt. New job title, new object, same mechanism as Window Four. That's not range. That's a costume change.
Decide what returns, and say what doesn't
Recurrence is a tool with a specific job. Say what yours is doing.
The host above returns for three reasons. He marks the joins between pieces, so the audience knows when one thing has ended. He gives them a place to breathe. And he restates the show's game in compressed form, which is useful precisely because episodes are otherwise discontinuous — a reader who has forgotten the principle gets it back in half a minute.
What does not return: anyone's history. The host's authority is identical every time he appears. Nobody wonders what happened to him last week, and the moment they do, you're writing a different show.
The same distinction applies to a recurring cast. If four performers play new people in every sketch, the audience builds familiarity with a performer's register, not with a character's past. That's a real pleasure and it is not continuity. Recognition and accumulation are different sensations, and a pitch that treats them as the same thing will end up explaining a fictional household that the sketches never visit.
You are also allowed to return nothing at all — no framing device, no recurring figure, only the perspective doing the work. That's a legitimate answer to the question, and it's a stronger one than a host who exists because you felt the episode needed glue.
Arrange the episode, then argue for the arrangement
An order is not a container. It's a claim about where the audience's attention should be spent, and two defensible orders of the same material will produce different shows.
Take the two sketches above plus three host beats — an opening, a middle, a close. Same content twice.
Order A: host, Window Four, host, How to Open a Box, host.
Order B: host, How to Open a Box, host, Window Four, host.
In A, the episode opens in talk and stays there for two long stretches. The audience settles into language, then the silent piece arrives as a drop in volume and an increase in scrutiny — the mode changes from listening to reading. The close then snaps back into speech. The episode drains toward silence and gets pulled out of it.
In B, the host opens, then hands off to the silent piece almost at once, before the audience has settled into the show's talking rhythm. The frame has to be learned on the spot. The desk scene then arrives as an outpouring of talk, and the episode builds toward its loudest point and stays there.
The host's mismatch also changes timing. In B, the host misdescribes the piece and the piece follows immediately — the audience catches the wrongness in real time, watching the gap open. In A, the opening misdescription has to be carried across the desk scene before it pays off. Same joke, delayed. Whether that delay is a pleasure or a fumble depends entirely on how strong the desk scene is.
Neither order is a tested running order. Nothing here has been read aloud, staged, or shown to anyone. The order is an argument you make to a commissioner, and the honest version says so.
One more thing about this section: it will tempt you to state a sketch count and a runtime, because that's what pitch documents look like. Don't, unless a specific commission has told you its numbers. Length and count are facts about a slot, not about your show.
Reject premises for two different reasons
Which brings the perspective back in, because its real use is rejection, and there are two distinct ways to reject.
The returns-counter premise fails on mechanism. It fits the perspective perfectly — small system, performed certainty — and it duplicates Window Four's action. Passing the perspective test is necessary and not sufficient. A pitch needs range of mechanism, not range of setting, and the easiest way to produce a repetitive show is to keep finding new counters.
A different premise fails on perspective. Two roommates arguing about whose turn it is to buy dish soap: no system, no performed certainty, no absurd demand. It might be a very good sketch. It isn't this show's sketch. That's not a quality judgment and the pitch shouldn't pretend it is.
Learn to tell the two rejections apart and you can explain your own selections without insulting anything.
Premise is not a sketch, and a rough order is not an episode
A premise is the situation, in a sentence; its comic action is what the piece does with it. A developed sketch has beats: the printout, the missing stamp, Window Two, the reversal, the last line. What you put in a pitch is usually the premise plus the intended direction of change, enough for a reader to see the mechanism. A pitch may also carry one fully written sketch to demonstrate voice, which is a legitimate and different move. Either way, say which one the reader is holding.
If you want the pitch document to be checkable by someone else, keep it to five things: the perspective, in a sentence that rejects something; two to four premises, each with its comic action named; a short statement of what recurs and what stays free; one proposed order with a reason; and a plain note on what's actually been written. That last line costs you nothing and buys you a great deal.
What remains, after all of it, is craft work rather than pitch work. The three premises need enough beats to prove their mechanisms are genuinely different and not just differently described. The order needs reading aloud to find out whether the changes in pace land. Episode length, performance response, and whether a silent piece can hold attention in the middle of a real show are all open until there is material. This article is editorial reasoning about the form; no creator's account of selecting and ordering sketches was consulted, and the three premises above are invented to make the method visible.
The test at the end is simple enough. Replace one sketch in your pitch with a premise you haven't written yet. If the pitch still reads as a show, the perspective is doing the work. If it collapses, you were describing a plot.
Frequently asked questions
What two failure modes does a sketch-series pitch usually fall into?
One is a list of premises with no relation between them, each funny enough alone but none belonging to the same show. The other is an apology wearing a plot, such as a shared apartment, boss, or season arc bolted on because the writer suspects disconnected pieces will not read as a series. Both miss what the comedy notices and how an episode is assembled.
How do you test a selection principle for a sketch show?
Write its rejection. If you cannot say what the show would turn down, you have a mood board, not a principle. 'Workplace absurdity' rejects almost nothing, while 'the paperwork that outranks the person' rejects a great deal. The perspective is the layer above the premises, and it should let a reader supply an unwritten sketch and get it right.
In 'Proper Channels,' what are the comic actions and mechanisms?
Window Four runs on certainty transferring without any fact changing: the visitor tightens into the attendant's official register while the attendant's confidence drains. How to Open a Box runs on exactness becoming the obstacle: labels subdivide, vocabulary expands, action contracts, and the box stays unopened. The host is a recurring figure who confidently describes the wrong procedure, and the confidence never wavers. The mechanisms are a reversal, an accumulation, and a fixed form with mismatched content; only one of the three runs on setup, escalation, punchline.
What are two different reasons to reject a premise?
A premise can fail on mechanism: a returns-counter premise fits the perspective but duplicates Window Four's action, so passing the perspective test is necessary but not sufficient. A premise can also fail on perspective: two roommates arguing about dish soap has no system, performed certainty, or absurd demand. It might be a good sketch, but it is not this show's sketch, and that is not a quality judgment.
What changes when you switch the running order, and what should a checkable pitch contain?
Order A opens in talk and stays there, so the audience settles into language before the silent piece arrives as a drop in volume and an increase in scrutiny; Order B hands off to the silent piece early, before the show's talking rhythm is learned, and builds toward its loudest point. The host's mismatch also changes timing: in B the wrongness is caught in real time, while in A it is delayed across the desk scene. A checkable pitch can keep to five things: the perspective in a sentence that rejects something; two to four premises with each comic action named; what recurs and what stays free; one proposed order with a reason; and a plain note on what has actually been written. No running order here has been tested, staged, or shown to anyone, and sketch count or runtime should not be stated unless a specific commission gave its numbers.