Pitch a Cooking Series Through Decisions, Not a Menu of Dishes

By pitch.dogFoundation13 min read20 September 2026

Television

You have the host. You have the dishes, and they photograph beautifully. What you have in the document is a menu: episode three, focaccia; episode four, Sunday ragù; episode five, the dumplings his grandmother made. A commissioner reading that knows what the show is about and has no idea what they will watch.

The substitution is small enough to make on one line. Name what someone has to work out, then say what would tell them they have worked it out. Bread becomes how do you follow a method that only exists in someone else's hands. Dumplings become what do you do when the person who knows is on the phone and can't see your counter. The food stays in the pitch. It stops being the subject.

The two episodes described below are invented for this piece. No filming, testing or maker's account stands behind them, so read them as shapes to copy rather than as evidence about how a real production went.

Replace the dish title with something someone must work out

A subject is not a question. Focaccia is a subject. How do you follow a method nobody wrote down is a question. A pitch needs both, and only the question can organize an episode.

For the question to have an edge, it has to belong to a particular person who wants a particular thing. Not "people want to learn bread," but this woman wants to be able to make it without her father standing in the room. Now the missing knowledge is specific. She probably has the ingredients. What she doesn't have is the moment-by-moment judgment about when a thing is finished — the part of a method that lives in hands and usually comes out of a mouth as one short word. Her father says "ready." She has heard the word her whole life and it has never told her what to do.

One caution about how you describe that gap. Unwritten means she has nothing to consult. It doesn't mean nobody anywhere ever wrote it down, and a pitch that inflates the gap into a lost family recipe is making a historical claim it can't support in order to get a scene. Pitch the gap that exists for this person.

The same move works when nothing is missing and two things are wanted at once. Two relatives are planning one meal for other people, in one kitchen, with one oven, on one afternoon. One thinks a welcoming table is generous in quantity — plenty of food, something for everyone. The other thinks a welcoming table is one everybody sits down at together, which means the dishes all have to be ready at the same minute. Nobody is withholding knowledge. Each has a purpose, and plenty of food, all of it hot, out of one oven at one time is more than the afternoon holds. That is also an episode question: what does each of them mean by welcoming?

The fastest way to make either of these look like television is to add a clock, a rival or a wound. Don't. A countdown gives your editor a rhythm and gives the cook nothing to notice; its presence tells the viewer that the real subject is the countdown. A rivalry turns a judgment into a contest, and cooking quietly becomes a scoreboard. A family wound, invented for a pitch, is a promise you will have to manufacture on the day of the shoot, in front of people who may not want to perform it. Real constraints already exist — a meal that has to reach the table for other people, an ingredient that is only good in one season. Use those. Just don't bolt a second one on to make the first feel urgent.

Put the judgment where the viewer can encounter it

Separate decisions from tasks. Chopping an onion is a task: the cook knows what finished looks like, and the camera is recording competence. Whether the onions are cut small enough for this dish, for this person is a decision, and a viewer can watch it being made. Pitches run on tasks. Episodes run on decisions.

A filmable decision has a moment, a duration and a witness. Look for four kinds of beat — a beat being any moment you would point at if you were describing the episode to someone in an elevator.

Comparison: two things held against each other, or one thing held against a memory. Hesitation: a hand that stops, a sentence that starts twice. Interpretation: someone reads a tone or a face and decides what it meant. Revision: the cook changes course. Revision is the strongest of the four, because a settled method doesn't revise. If a plan has no room for a wrong turn, it is a demonstration rather than an unfolding judgment — and if your outline says she gets it right on the second attempt, you have scheduled the outcome and there is nothing left to watch.

Then decide what the picture carries and what has to be said out loud. A camera holds a pause perfectly. It cannot hold the standard the cook is comparing against — another loaf, another kitchen, another person's hands. That is the part the cook has to explain, and it isn't narration laid over the image. It is the content. Get it while the person still knows what they were just doing, not in a tidy summary weeks later, which is a description of a settled thing and hides the settling.

This is where instructional and observational episodes part company, and both are legitimate shows.

In an instructional sequence, the method was settled before anyone arrived, the promise is that a viewer can do it afterwards, and the cook's explanation confirms what the picture shows. In an observational sequence, the method is the thing under discussion, the promise is that a viewer will see a decision and its consequence, and the explanation is the substance. The rule worth keeping is that neither pretends to be the other. Shoot a settled procedure and cut it as though the outcome were in doubt — a beat before the answer because a beat before the answer feels better — and you have made a claim about the people on camera that isn't true. Go the other way, shooting a real decision and then handing the viewer a neat list of steps as a payoff, and you have broken the promise you opened with. If you want to teach, say so in the pitch.

Give the host something consequential to do

The test is a trace, and it takes one sentence. Take the host's best moment and follow it forward. What does somebody do differently because it happened? If you can't answer, you have an interview, not a role.

In the dough episode, the host asks both people to point at the moment they would call "ready." The father points at something the daughter wouldn't have picked, earlier than her moment. Because that difference is now out loud, her next question changes from "how long?" to "what are you looking at?" — and the rest of the episode is about attention instead of duration. Small question, hard consequence, whole episode.

In the shared meal, the host asks each of them what the meal is for. The answer about plenty turns out to describe a table people keep getting up from: the dishes come out as they finish, hot ones hot, and that is what generous means to her. The answer about everyone together turns out to describe holding the first dishes under a cloth until the last one is ready, so nobody misses the first bite. Neither answer is about the menu. Once both are on the table, the argument moves onto the schedule, which is where the disagreement actually lives. Neither of them changes their mind. The disagreement just stops being about the wrong subject. That is a real host contribution even though nobody has learned anything new about food.

Three things a host can do inside that space: ask an experienced cook what they are noticing at the exact moment they notice it; try to say the cook's explanation back in their own words and get corrected — the correction is the scene; or find the place where two people are using one word to mean two things and stay there long enough for each of them to hear it. What a host can't do is adjudicate. If your host is deciding whose dough or whose dinner is correct, you have made a judging show, and the two relatives stop being the subject.

Enthusiasm is not a role either. "Amazing" and "look at that" are the same noise a producer could add in post. Neither is "now she's adding the—," which is a caption the edit can write better.

Keep the host's history and the participants' relationships as proposals until they are established. A pitch that says your host trained under someone, or that the two relatives have barely spoken in years, is making claims about a résumé and about real people. You can pitch the shape — someone with experience, someone without — without asserting biography you haven't confirmed, and without putting a participant's private history on screen as the reason to watch. If a family wants to tell that story, it is theirs to offer.

Finish the dish, then answer the episode question

Three things end at once in most pitches, and they shouldn't: the plate, the episode question and the season.

The plate is the easiest and the least able to end anything. A finished dish is a fact. Two people can understand each other's judgment and still not make the same bread. The daughter may be able to say exactly what her father attends to and still not have it in her hands this year. If the episode question was what does he notice that she doesn't, that is answered, and it was answered in an afternoon. If the pitch promised she learns to make it the way he does, the episode now owes the viewer something it can probably only fake. So choose the question at pitch stage by asking what a single episode can honestly settle.

The shared meal works the same way in reverse. What does each of them mean by welcoming is answerable, by asking. Will they agree about dinner might not be answered this year, and the pitch shouldn't promise it. Two people can both want the same generous evening and mean different things by it. Making that legible is a real ending. Somebody winning is a different show.

Then the season. What stays constant is the promise: someone has to work something out, and you get to watch them work it out. What varies is the kind of decision. Here is the trap — swap the list of dishes for a list of decision types and you get: episode three, an unfamiliar ingredient; episode four, a shared meal; episode five, someone else's method. That is the same pitch in a new vocabulary. Variety a viewer can feel comes from different shapes, and the two episodes here have different shapes: one is an investigation into what another person notices, the other a negotiation about what a meal is for. Each generates its own beats. If two episodes in your outline produce the same beats, you have written one episode twice.

The check this method still needs is simple, and it isn't done in this piece: watch an authorized episode and a maker's account of the same episode side by side, and mark every place the account explains something the footage doesn't show. Those explanations are your beats. Anything else is a description of a result. Keep practical cooking, nutrition and safety instruction out of a pitch like this — not because it is unwelcome, but because it needs its own qualified sourcing, and it isn't what these episodes are doing.

Here is what the two episodes look like written out. Neither has been made.

A woman in her forties wants to make a dough her father has made for years without measuring anything. There is no written version for her to consult — that is the gap, not a claim about the recipe's history. Over an afternoon she tries to get him to say what he is doing. He says what he has always said: "ready." She still doesn't know which moment he means. The host asks each of them to point at the instant they would call ready, and the two answers don't match. From there the episode follows their attention rather than their clock — what he is watching, what she is watching, and what changes once she stops asking how long. The dough gets made and eaten. The question the episode set out to answer, what he notices that she doesn't, gets answered. Whether she can now do it herself is left where it belongs.

Two relatives are planning one meal for other people in one kitchen with one oven, and they agree about almost nothing, because they mean different things by welcoming. One wants plenty and something for everyone, served as it comes out; the other wants everyone at the table at the same moment, which means holding the first dishes until the last one is ready. Both are trying to be generous. The host asks each of them what the meal is for, and the two answers turn out to describe different evenings: a table people keep rising from, against a table where the early dishes wait and cool. The argument moves off what to cook and onto when to eat, which is where the disagreement actually lives. The meal happens. The disagreement doesn't end, and the episode never promised it would. What a viewer gets instead is a clear view of what each choice costs the other person, which is more than the same two people shouting about a menu would have given them.

Now delete the food. Take out the dough, the meal, the kitchen, everyone's names. What is left should still tell you what you are going to watch: one person trying to find out what another person notices; two people discovering that they mean different things by the same word. If the paragraph falls apart when the dish goes, the dish was the pitch.

Frequently asked questions

A menu names subjects but not what a viewer will watch. The article argues for replacing each dish title with a question someone has to work out, plus what would show they have worked it out.

How can a pitch describe an unwritten method without overclaiming?

The gap is specific to this person: she has nothing to consult. It is not a claim that nobody ever wrote the method down or that a family recipe was lost. Pitch the gap that exists for her, not a historical claim used to manufacture a scene.

What makes a decision filmable rather than just a task?

A task records competence, as when a cook chops an onion and knows what finished looks like. A decision is a judgment with a moment, a duration, and a witness. The four beats are comparison, hesitation, interpretation, and revision; revision is strongest because a settled method does not revise.

How do instructional and observational episodes make different promises?

In instructional sequences, the method is settled before shooting, the promise is that a viewer can do it afterward, and the cook's explanation confirms what the picture shows. In observational sequences, the method is under discussion, the promise is that a viewer will see a decision and its consequence, and the explanation is the substance. Neither should pretend to be the other.

What useful role can a host play—and what should a host not do?

A host can ask an experienced cook what they are noticing at the exact moment they notice it; try to say the cook's explanation back and get corrected; or find two people using one word to mean two things and stay long enough for each to hear it. A host cannot adjudicate without turning the show into judging. Enthusiasm alone is not a role, and biography or relationships should stay proposals until established.