Pitch a Factual Explainer Series Around Questions, Not a List of Subjects

By pitch.dogFoundation12 min read20 September 2026

Television

A deck that names twelve subjects has told a commissioner what your series is about. It has not told them what an episode does. Those are different pitches, and only the second one survives a commissioning meeting.

"Queues." "Maps." "Sleep." "Why your train ticket costs what it costs." Each of these is a slot. A slot can be filled by almost any script, which is exactly the problem: nothing in the pitch tells the reader how a viewer thinks differently at the end than at the start. The repeatable unit of an explainer series isn't the topic. It's the question, the evidence that tests it, and the revised understanding the episode can honestly offer.

What follows is a working method for building that unit, developed through two invented episode pitches. Neither has been produced, filmed, or tested. The framework itself is an editorial proposal about how to pitch, not a report from a series that used it — and the demonstrations below are design exercises whose results are precisely what the pitch proposes to find out.

Turn the subject into a question with a consequence

"Queues" is a label. "Why can two queues of the same length take different times?" is a question. The difference isn't decoration. The question names a puzzle a viewer can hold in mind while watching, and it tells the production team what to go and find out.

The second half matters as much as the first. A useful question arrives with a consequence: an answer would help the viewer understand something beyond the immediate case. If waiting turns out to be produced by something other than the number of people ahead, the answer also bears on what would have to be true for an extra till to help, and on why one slow transaction can matter more than the crowd standing behind it. A question earns its place by pointing at more than itself.

Do not, though, build the question on a claim you haven't tested. It is tempting to open with "We all believe the longer queue is the slower one." That's an assertion about viewers, and it needs evidence — a survey, prior research, or at minimum a producer-owned account of how audiences responded. What you can say without testing is narrower and still usable: a viewer waiting in a line may carry a rough rule that more people ahead means a longer wait, and the episode is designed to find out what that rule misses. If reporting later shows the rule is uncommon, the episode changes; the pitch hasn't broken.

The practical test of a question is whether it creates a specific reporting task. "Why is waiting so strange?" sounds like a series and assigns nobody anything. "Why can two queues of the same length take different times?" tells you to find two equal lines, or to build a comparison in which length is held still while something else moves. That is a job.

A provocative slogan is the failure mode. "The queue is a lie" is both unfilmable and unfalsifiable — no evidence would settle it — so the team ends up writing narration that asserts it instead.

Write the steps between the initial and better explanation

An episode's spine is a short chain of turns. For each one, the pitch should name three things: the explanation the viewer currently has, the evidence needed to test its limit, and the revised understanding that evidence might permit. Write them as a sequence rather than a list, and mark which steps are conditional.

A turn that only adds a fact without changing the explanatory path isn't a turn. It's trivia. "The average commuter spends a certain number of hours a year waiting" is a statistic. Nobody's model of why moves.

For the queue episode, an illustrative spine might run like this. The current explanation: a line of a given length takes a predictable time to clear. The evidence step: build a comparison that holds the starting count fixed and changes only one thing — how service time is distributed across two parallel lanes. Two lanes, the same number of tokens in each, one server per lane, and a stipulated service interval for every token. In the first run, the intervals are similar across both lanes. In the second, the same total amount of service work is redistributed so that one lane carries the long intervals and the other carries the short ones. Each lane begins both runs with the same number of tokens, and neither the starting count nor the total work changes between the two runs.

The revised understanding that comparison might permit: within this model, equal lane length doesn't determine when a lane clears. The distribution does. The light lane empties early while the heavy one is still going, even though both began with the same number of people.

Notice what is conditional here and what isn't. The design of the comparison is a proposal. The arithmetic — a lane's clearing time is set by the intervals assigned to it — follows from the stipulated conditions. But whether the visible difference between the two runs reads clearly on camera, and how much it matters, is unknown until someone builds it and looks.

If the model shows less than you hoped, the episode hasn't failed. It has learned that the mechanism isn't visible under these conditions, which is itself an explanatory turn worth filming — provided you can say what the null result rules out and what it doesn't. A model that fails to demonstrate something does not establish that the thing is false in the world. It only fails to demonstrate it here.

That is the discipline that keeps a series honest: the episode must not require every test to vindicate the narration you walked in with. Some questions stay open. A pitch that leaves one unresolved, and says so, is stronger than one that pretends the answer was always going to arrive.

Give images and sound explanatory work

Moving image earns its place in an explainer when it lets the viewer inspect something a sentence couldn't show them. So the first question for any planned shot is: what does this let someone see?

A tabletop comparison can do real work. If each lane opens both runs with the same visible number of tokens while the pacing of removal differs, the viewer can watch two identical starting counts resolve at different moments. A single consistent audio cue for each completed service — the same sound every time, in both lanes — lets the rhythm be heard as well as seen, so that one lane falling silent while the other grinds on registers in two channels. That is what the medium is for.

What decorative footage cannot do is establish a mechanism. Cutaways of people checking phones, sighing, shifting their weight establish impatience. They tell nobody why the line moved slowly. If the episode's argument is about how service work is distributed, no quantity of waiting-room atmosphere supports it.

The distinction the pitch must make explicit is the one between a model and evidence. The tabletop lanes are a model built to illustrate a relationship under stipulated conditions. They are not evidence that the relationship explains real queues. Those are two different claims, and a pitch that slides from one to the other will be caught by whoever reviews it. If you want to say something about a real checkout, you need real material — timing data, service logs, an account from someone who ran the counter — and that is a separate reporting task with its own dependencies and permissions.

Two practical items belong beside the shot list. Feasibility: who builds the model, in what time, with what materials, and at what cost? Review: who checks the subject claims before filming? For this episode that means someone who understands queueing well enough to say whether the stipulated conditions isolate what you claim they isolate. A model that changes three things at once proves nothing about any one of them, and that flaw is far cheaper to find on paper than on set.

Prove the approach can produce a different episode

A second episode is where a series pitch either demonstrates range or reveals that it has one idea wearing different nouns. Swap "queues" for "maps," keep the tabletop, and you haven't written a second episode. You've written a reskin.

Take a genuinely different question: why do two maps of the same journey give different impressions of distance?

The discipline is identical — question, evidence, revised understanding. Almost nothing else is. There is no tabletop, because there is no mechanism to construct. The subject is a document, not a process. The evidence step is comparative and archival: take two published maps of the same route and let the viewer inspect where the same stretch of ground occupies a visibly different share of the whole. That comparison has to be normalised — the maps brought to a common scale, or measured by relative distances inside each drawing, such as the ratio between two legs of the journey, so that overall scale cancels out. A bare size difference would not count: the same ground taking up more space on a larger sheet is reproduction size, not a different impression of distance, and the episode should say so before someone mistakes one for the other. The visual method changes with the subject — aligned side by side with a shared endpoint, perhaps a transition that slides one drawing onto the other so the divergence is felt rather than described. The sound design is probably thinner, because attention is resting on a still image.

That contrast is the point. If your format only works when you can build a model, it isn't a series format. It's one episode with sequels.

There is a temptation worth naming. Both episodes could be flattened toward a single convenient thesis — something like "representations aren't reality" — and made to rhyme. Resist it. The queue episode is about how waiting is produced under stated conditions. The map episode is about how a drawing encodes a purpose. Forcing them toward a shared moral makes the series feel coherent in a deck and repetitive on screen.

A second subject that needs archival comparison instead of a constructed demonstration tests whether the approach actually flexes. A third that needs neither — a question best answered by interviewing people — would test it further, and might show where the format ends.

Show the unresolved work inside the pitch

Commissioners are not fooled by confidence, and they aren't much impressed by it either. What they need is an accurate picture of what they're buying: a method of explanation and a viewing experience, not a conclusion that later reporting is obliged to confirm.

That means labelling the state of every claim. Three labels do most of the work: researched, hypothesis, proposed for filming. A compact episode block might read:

Episode 1 — Why can two queues of the same length take different times? Researched: the principle that clearing time depends on more than headcount in a single-server system (awaiting subject verification). Hypothesis: that a tabletop comparison holding the starting count equal while redistributing service intervals between two lanes will make the difference legible on camera. Proposed for filming: the two-lane build, the shared audio cue per completed service, the visible starting token count held equal across both runs. Open: what real checkout data would show; whether the effect survives arrivals spread over time rather than all present at the start.

That block does something a topic list cannot. It tells the commissioner which part of the episode already stands and which part is an errand. It also protects the team: nothing in the pitch promised the demonstration would behave as hoped, so a surprising result is a development, not a breach.

Then add the second question, briefly, to establish range — and be honest that its path is less developed. "Why do two maps of the same journey give different impressions of distance?" carries a different evidence dependency: it needs published maps, rights to use them, a way of normalising them to a comparable scale, and a subject review of whatever claims the episode makes about each drawing. Naming that dependency in the pitch beats discovering it in week three.

The pitch's final job is to show one explanatory path in enough detail that the reader can inspect it, and one further question that proves the path isn't a one-off. Two questions, two genuinely different routes from question to evidence to qualified understanding. If reporting later overturns the narration you were hoping for, the episode is still makeable — it simply becomes a different episode. Say so in the pitch. That is the promise that keeps a format alive across a season.

Frequently asked questions

Why is a list of topics not enough for an explainer series pitch?

It names what the series is about but not what an episode does. A topic is a slot that almost any script could fill, so nothing tells the reader how a viewer thinks differently at the end. The repeatable unit is a question, the evidence that tests it, and the revised understanding the episode can honestly offer.

What makes a question strong enough to build an episode around?

It names a puzzle a viewer can hold in mind while watching, and it creates a specific reporting task. It also arrives with a consequence: an answer would help the viewer understand something beyond the immediate case. An untested claim about what viewers believe should not be the opening; you can instead say a viewer may carry a rough rule and the episode is designed to find out what it misses.

In the queue episode, what does the tabletop comparison test, and what does it not prove?

It holds the starting count equal while redistributing the same total service work so one lane carries the long intervals and the other the short ones. Within that model, equal lane length does not determine when a lane clears; the distribution does. But the tabletop is a model, not evidence that the relationship explains real queues, so real checkout claims need real material such as timing data, service logs, or an account from someone who ran the counter.

How does the map episode differ from the queue episode, and why include it?

The map episode asks why two maps of the same journey give different impressions of distance. It uses comparative archival evidence — two published maps normalised to a common scale or measured by relative distances inside each drawing — rather than a constructed tabletop model. Including it shows the approach can flex instead of being one episode with sequels.

Researched, hypothesis, and proposed for filming. Labelling each claim lets the commissioner see which part already stands, which part is an errand, and what remains open. It also protects the team: a surprising result becomes a development rather than a breach of what the pitch promised.