Observation, Investigation, or Intervention: What Is Your Factual Show Doing?

By pitch.dogFoundation12 min read20 September 2026

Television

Most factual pitches open with the same promise: this is real. Real people, real place, nothing staged. The word real is doing a great deal of work in that sentence, and most of it is unearned. It tells the person reading your pitch nothing about what your production actually does to the world it films. Observation, investigation, and intervention are three different actions with three different standards of proof, and only one of them is what people usually mean when they call a show "just observational."

The fix is smaller than it sounds. Take one representative episode and answer a single question: what would happen if your team never showed up? Whatever remains is your raw material. Whatever you add — a question pursued, a condition introduced — is your method. Write both in ordinary sentences, then check that the episode you've described actually does what those sentences promise.

One warning before the example. Everything below is invented: the workshop, the volunteers, the numbers on the shelf, the proposal, and any outcome. None of it was filmed, tested, or observed. It exists to make three distinctions concrete, and you should treat it the way you'd treat a paper sketch of a set.

What would happen if you weren't there?

Picture a volunteer-run repair workshop. Call it Sable Street, and remember that it isn't real. It opens one afternoon a week. People carry in kettles, lamps, a food mixer with a stripped gear, a child's toy that stopped making noise. At the intake table a volunteer writes down the item, a first name, a phone number, the fault as the owner describes it, and which bench will take it. A session takes in around twenty-five items. Volunteers get through most of them. The rest go on a shelf in the back room.

At the point our fictional team arrives, the shelf holds forty-one items gathered from the last three months of sessions — roughly 312 intakes. That shelf is the whole example, because all three routes begin at the same Saturday, with the same queue at the door and the same intake table, and then diverge on the next thing the production does.

Route one. The team films the session. It does not touch the intake process, does not suggest which items get priority, does not ask anyone to do anything differently. Remove the team and the queue, the intake table, the benches, and the shelf all carry on exactly as they were. The production's action is presence: choosing which sessions to film, where the cameras point, what happens when a camera operator can see a volunteer about to slice a thumb open on a broken lamp base, and who gets asked to explain themselves afterwards.

Route two. The session still runs. The shelf still fills. What doesn't happen, without the team, is anybody sitting down with the intake log and the shelf and pursuing the question of why those forty-one items are still here. The production's action is pursuit: a question, followed through evidence, on camera.

Route three. The session runs. What doesn't exist is the collection window printed on the intake slip, the named volunteer whose job is to make the follow-up call, and the second-visit slot at the end of the afternoon. The production's action is introducing conditions and then filming what follows.

If you can't put your show into one of those sentences in plain words, you probably have two methods in one format and haven't noticed yet. That's worth finding out before you're in the edit.

And notice what route one does not claim. Filming an untouched session is not the same as having no influence on it, and a pitch that leans on invisible neutrality will embarrass itself the first time someone asks why the volunteer with the worst bench manner never appears. The useful distinction isn't did the production change anything. It's what does the production propose to do, and would the material exist without it.

Give each method its own evidence burden

Each route can prove something different, and the honest pitch says which.

Observation needs a defensible account of what was recorded. "We filmed six Saturdays and the shelf stayed full" is a claim the footage can carry. "Repair workshops are drowning" is not. Filmed sessions are evidence about the filmed sessions — which days you chose, how many, and what the camera couldn't see. That's a modest claim, and it's a strong one when you state it precisely, because it can't be knocked over.

Investigation needs a question posed before the answer is known, evidence anyone can check, and room for the answer to be no. In our fictional workshop, the team sorts the forty-one shelved items by reading the log — first by whether the margin carries a note about a part or a return visit, then by whether the contact line can be read at all. That gives fourteen items with an outstanding note, nine with a number nobody can decipher, and eighteen with a legible number and nothing flagged. Anyone can re-read that log and disagree with the count, which is exactly what makes it evidence.

But the log is a record of what a volunteer managed to write at a busy table. It is evidence about the intake process, not a census of why people didn't come back. The phone calls are a different kind of thing again: accounts given later, shaped by who's asking and what they'd rather you think. Both are useful. Neither is the other.

Here is the part that separates an investigation from a conclusion with cameras. If the owner calls establish that nobody was ever rung back, the episode has to be able to say so. The stronger version of the show is that finding. Why it happened is a second finding from a second source: if the volunteers say, at their benches, that each of them assumed somebody else made the call — or the coordinator says it for them — the episode has to name that source, because the calls can't. A format that can only air the answer "people abandon their things" wasn't investigating.

There's a boundary here too. The moment the team stops ringing people to ask what happened and starts ringing them to say "come Saturday and we'll hold a slot," it has moved into route three.

Intervention needs two things spelled out: the conditions you introduced, and what the resulting behavior can support. The collection window, the printed slip, the named caller — those are the conditions. What follows is behavior under those conditions, in this workshop, over the sessions you filmed. It cannot support "this arrangement reduces uncollected items," and it certainly can't support it beyond Sable Street. It also cannot be presented as the workshop's ordinary operation. If you show a tidier shelf in episode four and don't say who built the collection table, viewers will file it under how this place works, which is a claim your footage did not earn.

Mark the moment a hybrid changes method

The interesting version of Sable Street is a hybrid, and hybrids are fine. What gets shows into trouble is a hybrid that never announces itself.

Sequence it honestly. Two Saturdays filmed without the team touching anything: that establishes what a session looks like, roughly twenty-five intakes, and the shelf. Then the turn — the team asks the coordinator whether it can spend a fortnight on the shelf itself, working the log and making calls with each owner's agreement. That's the shift from observation into investigation, and it happens because the first two sessions raised a question the team couldn't answer by watching.

Then a second turn. The calls suggest that the missing step is a person who owns the follow-up. The team proposes a collection window and a named caller. If the coordinator agrees, three more Saturdays get filmed with the arrangement running.

Three moments now need to be visible to the audience, and each is a sentence, not a scene:

Who raised the question. Was the shelf the team's curiosity or the coordinator's long-standing complaint? Those produce different shows.

Who proposed the change. If the team proposed the collection window, say so. If a volunteer said "what if we just printed a time on the slip," then the workshop introduced the condition and the team observed a change — a different production action entirely, and one you should not claim credit for. If it was both, in stages, say that too. Partly ours is a legitimate and disclosable answer.

When the change took effect, and which sessions show it. "The first two Saturdays show the workshop as it runs. The last three show it running with a collection window we suggested."

Get any of those wrong and the episode misrepresents its own method while every individual shot remains accurate. That's the specific failure this whole exercise is for: not fabrication, but a true sequence with a false frame around it.

Revise the promise when people decline or redirect it

Now run the awkward versions.

The coordinator says no to the collection window. What's left? The two observed sessions and the investigation into the shelf — including the finding that nobody owned the follow-up call. That is still a show, and arguably a tighter one. A pitch that only works if the yes arrives is a pitch with a single point of failure, and it's the one part you don't control.

Someone at intake doesn't want to be filmed. A volunteer doesn't want their bench on camera. Four of the eighteen people with legible numbers decline to talk about why their item is still there. Report the four refusals and stop. A declined call is not a withheld explanation, and an object sitting on a shelf is not evidence that its owner doesn't care. This is where a factual pitch most often invents a motive it never had grounds for.

The investigation finds the opposite of what you assumed. Suppose the log shows the workshop tried a collection window two years ago and it collapsed — the volunteer who ran it moved away, the window clashed with the last bus. Your intervention is now a rerun, and the episode must say so. It also gets more interesting, because the question shifts from will this work to why did it fail before, which is a question the footage might actually answer.

A participant redirects rather than refuses. The coordinator agrees to a collection window but not to filming the calls. Fine — you can show the window and the shelf and say plainly that the conversations aren't in the episode, and what claim that means you aren't making. A redirect narrows your evidence; it rarely kills the show.

Two things to keep separate throughout. A planned method is not a promised outcome: your pitch describes what the production will do, not what the workshop will become. And none of these three labels is a consent form. Calling your show an intervention makes the welfare, permission, and review questions louder, not quieter, and those are their own project-specific work for qualified people. Naming your method is the beginning of that conversation, not a substitute for it.

Say what changed, and who changed it

Strip it back to three sentences you could defend in a commissioning meeting. What the production does: watches, pursues, or introduces — pick the verb, and don't hedge it with real. What the viewer sees: a session as it runs, a question worked through a log and a phone, or a change with a start date. What the resulting material can support: a description of filmed sessions, a finding with its evidence attached, or behavior under conditions you introduced.

Then write the transition sentence, and put it in the episode:

For the first two Saturdays we filmed the workshop as it normally runs. From the third week, at our suggestion and with the coordinator's agreement, the workshop added a collection window and a named caller. Everything after that point shows the workshop running with that change in place.

Adapt the wording; keep the structure. Who changed what, when it started, and what the audience is looking at on either side of the line. If a hybrid can't produce that sentence, it's hiding the one fact that determines what its own footage means.

The workshop is invented. Nothing above was filmed, tested, or observed, and no participant, arrangement, or result exists. What you can take from it is the question, and it works on real pitches the same way: name what your production does to the situation, name what would have happened without you, and make sure the episode says the same thing.

Frequently asked questions

How do I tell whether my show is observation, investigation, or intervention?

Ask what would happen if your team never showed up. Observation is presence: the session, queue, intake table, and shelf would carry on. Investigation is pursuit: without the team, nobody sits down with the log and shelf to ask why items remain. Intervention introduces conditions: without the team, there is no collection window, printed slip, or named caller.

What does each method need to prove?

Observation needs a defensible account of what was recorded, not a broad claim about all repair workshops. Investigation needs a question posed before the answer is known, evidence anyone can check, and room for the answer to be no. Intervention needs the conditions introduced and what the resulting behavior under those conditions can support, which is not the same as the workshop's ordinary operation.

What must a hybrid announce when it changes method?

It must make visible who raised the question, who proposed the change, when the change took effect, and which sessions show it. A hybrid can be fine, but it cannot hide the fact that determines what its own footage means. A true sequence with a false frame around it is the failure.

How should I handle refusals or redirects from participants?

Report refusals and stop. A declined call is not a withheld explanation, and an object on a shelf is not evidence that its owner does not care. A redirect narrows your evidence but rarely kills the show; it should be stated rather than filled in with an invented motive.

No. A planned method is not a promised outcome: the pitch describes what the production will do, not what the workshop will become. None of the three labels is a consent form. Welfare, permission, privacy, and review questions become louder with intervention and are separate project-specific work for qualified people.