Turn Documentary Research Into a Plan for What to Film
Turn Documentary Research Into a Plan for What to Film
Take one finding from your research and ask where its consequences could become visible or audible. Is someone making a decision, repeating a task, comparing records, returning to a place, or trying to resolve an account that does not fit? That is the beginning of a filming plan. “We will explore the issue” is not yet one.
The useful distinction is between evidence that supports a finding and material through which a viewer can understand it. Sometimes they are the same thing: a recording captures the event you are investigating. Sometimes a written record establishes a fact, while an activity happening now offers a way into its consequences. And sometimes neither observation nor the available records can carry the proposed question. Then the next action is more reporting, not a more confident description of the film.
Here is a way to make that choice without scripting reality in advance.
Start with a finding, not a subject
Consider a fictional documentary about a repair workshop. All workshop records, people, and proposed encounters in this example are invented teaching material, not a real project or a reported case.
“The difficulty of keeping old machines running” is the subject. It suggests a world, but it does not tell you what the film should follow.
A usable finding is narrower:
Three machines appear in the workshop's surviving job cards for a parts replacement and again for a later repair visit. The cards do not establish why the machines returned.
This sentence contains something worth investigating without assigning a cause. The replacements may have failed. A different component may have broken. A return visit may be normal servicing rather than a failed repair. The records might describe different problems in language too brief to resolve them.
Write down what supports the finding before turning it into a scene. In our exercise, that means the identifiers connecting the visits, their dates, and what each card actually records. It also means noticing what is absent: a complete history for every machine, a diagnostic report, or an account from the person using it.
“New parts are making the machines fail” is a different claim. You cannot bridge the gap by filming a row of broken machines. The image would be vivid; the causal problem would remain.
Now identify the consequence that makes the finding matter to the film. Perhaps the interesting question is how the workshop decides whether to repair a machine again, replace it, or stop accepting the job. That question leads toward a process rather than a verdict. It also gives you a reason to investigate ordinary work, not just wait for a spectacular failure.
Find the activity that could expose the question
Ask what a camera and microphone could register that would help a viewer follow that decision.
For the workshop, possibilities include a customer explaining the fault at intake, a technician comparing the machine with an earlier job card, a test that rules out one explanation, or a conversation about whether another repair is worth attempting. Each possibility offers a relationship: an account meets a record; a diagnosis changes; a practical limit enters the decision.
A beautiful shot of tools does not perform the same job. It can establish texture, rhythm, or a place you want the viewer to spend time in. Keep it for those reasons. Do not ask it to demonstrate why the machine returned.
The same distinction applies to an interview with an authority. A technician describing the general difficulty of repair may give essential context. That description does not establish what happened to these particular machines. Decide whether you need context, an explanation of a specific record, or an unfolding decision before making a contributor list.
A proposed opportunity should therefore name both an activity and a question. “Observe intake and diagnosis to understand how a repeat visit changes the next repair decision” is more useful than “Film workshop life.” It tells the recipient what you intend to follow and gives you something against which to judge the material you obtain.
Three routes through the same finding
The job cards do not dictate a single kind of documentary. They open several routes, each with a different evidentiary limit and a different next action.
Follow the process as it happens
Suppose the workshop is willing to discuss filming, but no access agreement or participant consent has been established. You could propose following a returning machine from intake through diagnosis and the decision about further work.
The material might include the customer's description, the technician's tests, comparisons with an earlier card, and the conversation in which somebody decides what happens next. The film would be following how people investigate a problem, not displaying proof of a conclusion already reached.
This route needs an actual opportunity. The relevant machine may not return during your filming period. The diagnosis may be inconclusive. A participant may decline. A decision may occur somewhere the camera cannot go. Those are conditions of the proposal, not defects to conceal with present-tense scene writing.
A useful next action is to ask how repeat visits are handled and where that process could be observed. It is not to request that someone reproduce a confrontation for the deck's benefit. Should a demonstration later become useful, identify it as a demonstration rather than passing it off as a spontaneous event.
What can this route answer? It can show how a particular diagnosis and decision unfold. It cannot, by itself, establish the failure rate of every replacement part the workshop has used.
Build an openly authored sequence from the records
Now suppose filming access is unavailable, but the project has usable, appropriately cleared records. You could construct a sequence that follows one machine across its dated cards.
The first card records a replacement. A later card records another visit. Rather than animate a confident arrow labeled “part fails,” show the interval and the difference between the descriptions. Let the missing diagnosis remain missing. An authored explanation can make the limits of the record intelligible instead of disguising them.
The sequence could move from the full card to its date, machine identifier, and fault description, then place the two visits beside each other. A narrator could distinguish what is recorded from the question it raises. The construction is the filmmaker's; the cards are its evidence. Do not add invented mechanical sounds that appear to be recordings of the historical failure, or present a staged workbench as footage from that visit.
This route changes the next action. You need to inspect the record closely, identify what the audience must be able to read, and establish whether the proposed reproduction is available to the production. You may also need further reporting to interpret a notation. Possessing a copy and being able to use it in the film are separate matters.
The result can be a filmable account of what the surviving records reveal and withhold. It is not an observational scene rescued by decorative motion.
Report further before choosing either form
A third version of the exercise gives you only photographs of several cards with cropped machine identifiers. You cannot reliably establish that the later visit concerns the same machine.
In this version, the original finding is not yet supported. Neither a workshop shoot nor an elegant record sequence fixes that. First obtain fuller records or another reliable way to connect the visits. The next interview might exist entirely for research: to understand the filing system, test your reading, or locate missing material.
This can be the least visually exciting next step and still be the correct one. A filmmaker does not owe the pitch a scene before the reporting supports it.
Nor must further reporting always produce the original film. It may reveal that the supposed repeat repairs were routine inspections, or that the real decision occurs with the machine's owner rather than at the workshop. The plan should be able to move with that finding. Otherwise the research has become a casting exercise for a predetermined explanation.
Let available material change the proposal
There is a documented example of research altering an archive-led production in Todd Douglas Miller's account of Apollo 11. In his 2019 Smithsonian interview, Miller explains that the team initially planned to scan the available film and other assets; research at the National Archives brought a large-format collection into the project. He also describes using footage from other missions for some experiences the Apollo 11 material did not cover. That second point matters as much as the discovery: material selection involved a reconstruction decision, not merely finding an illustration.[^1]
CNN's published transcript of an aired preview contains a different, small-scale example of what archival material can carry. Aldrin and Armstrong negotiate his position, the hatch, and the steps during the lunar exit. In the surrounding interview, Miller discusses finding technical details and human moments through research. The excerpt supplies a practical exchange rather than a retrospective explanation of its importance.[^2]
These are two bounded observations, not a claim to have inspected the complete film's edit. The Smithsonian interview documents the maker's account of his choices; the CNN transcript documents the words in that aired excerpt. Neither establishes a universal recipe for documentary structure.
For the workshop film, the transferable question is simple: what does the actual material allow you to follow? A job card, a diagnostic test, and a remembered account carry different kinds of knowledge. Choose the relationship you can substantiate, not the one the pitch initially made easiest to imagine.
Write the proposed material without promising the response
The deck needs enough specificity for a recipient to understand the film. You do not have to replace every verb with “might.” You do have to distinguish material held, access established, and work being pursued.
Here is a pitch passage for the first version of our fictional exercise:
Surviving job cards connect three machines to a parts replacement and a later repair visit; they do not establish why the machines returned. We propose following intake, diagnosis, and the decision about further work, asking what a return visit changes for the customer and technician. We hold copies of the cards. Filming access and participant consent are being pursued; no return visit or particular outcome is secured.
The uncertainty has not removed the film. You can still picture its activity and understand its question. What has disappeared is the promise that a technician will admit fault, a customer will become angry, or an investigation will deliver the answer the filmmaker expected.
Compare that with “We follow a frustrated customer as the workshop confronts the consequences of its failed repairs.” Unless those events and that cause are established, the sentence spends evidence the project does not have. A small disclaimer elsewhere in the deck will not undo it.
Keep the access statement close enough to the proposal that the recipient can tell the difference. This is especially important when a page includes a convincing location image: seeing the workshop in a photograph does not tell the reader whether you may film its staff or customers.
Choose the next action the question needs
Before commissioning another mood board, take one finding through this short written sequence: what is supported; what consequence matters; what activity or record could make it legible; and what must happen before you can obtain that material.
Then choose an action with an observable result. Ask to inspect the complete cards. Arrange a research conversation about intake. Request access to follow a diagnosis. Test whether a record sequence can preserve the distinction between a second visit and a failed repair.
You are not trying to predict the best scene. You are deciding where the film can encounter something it does not yet know—and giving the pitch recipient a clear account of why you are going there.
Sources
[^1]: Air & Space / Smithsonian, “‘Apollo 11’ Takes a Fresh Look at the First Moon Landing”, March 8, 2019. Todd Douglas Miller's answers to the opening question about the archive discovery and the later question about material from other Apollo missions. Maker's retrospective account; not independent footage analysis. Inspected September 19, 2026 (UTC).
[^2]: CNN, CNN Newsroom, June 23, 2019, 8 p.m. ET hour, segment 07. Final Apollo 11 interview and the second “BEGIN VIDEOTAPE” excerpt following the 20:55:08 marker, from Aldrin's question about his feet through the lunar-view exchange. Published broadcast transcript, not a timecoded transcript of the complete film. Inspected September 19, 2026 (UTC).
Frequently asked questions
What turns a documentary research finding into a filming plan?
A finding becomes a plan when you ask where its consequences could become visible or audible, such as someone making a decision, repeating a task, comparing records, returning to a place, or trying to resolve an account that does not fit. The plan should name an activity and a question, not just say the film will explore the issue.
What is the difference between evidence that supports a finding and material through which a viewer can understand it?
They are sometimes the same, as when a recording captures the investigated event. Sometimes a written record establishes a fact while a current activity offers a way into its consequences. If neither observation nor available records can carry the question, the next action is more reporting.
In the repair workshop example, what can the job cards establish?
The cards connect three machines to a parts replacement and a later repair visit, with identifiers and dates, but they do not establish why the machines returned. The replacements may have failed, another component may have broken, or a return visit may be normal servicing. Filming broken machines would not bridge that causal gap.
How should a pitch handle unsecured access or an uncertain outcome?
It should distinguish material held, access established, and work being pursued. For example, a pitch can say it holds copies of cards while filming access and participant consent are being pursued and no return visit or particular outcome is secured. A small disclaimer elsewhere does not undo a sentence that spends evidence the project does not have.
What are the three routes the article proposes through the same finding?
One follows the process as it happens, such as intake through diagnosis and decision, if access and opportunity exist. Another builds an openly authored sequence from records, preserving missing diagnoses and distinguishing what is recorded from the question it raises. A third reports further before choosing either form, especially if the finding is not yet supported.