Write a Factual-Series Narrator Who Can Say Only What the Material Supports
Write a Factual-Series Narrator Who Can Say Only What the Material Supports
A factual series almost never fails because its narrator knows too little. It fails because the sample voiceover sounds like someone who was everywhere.
You can hear it in a pitch deck. The voice is calm, the sentences are short, the transitions land, and somewhere in the middle the narrator explains what a person was thinking in a room nobody filmed. The read is good. The read is not the problem. Confident delivery has never improved a piece of evidence; it only makes an unsupported sentence easier to approve.
So the working rule here: every narrated sentence is also a claim about how the narrator knows it. Before you choose a voice, mark what each sentence rests on — something the camera shows, something a person said, context your production has established, or your own interpretation. Then ask which of those a narrator can honestly carry.
This piece does that with one small, deliberately thin set of materials. Everything below is fictional and stipulated up front, so the example's arithmetic can be checked. No real script or recording was audited here; the method is what's on offer.
Audit one short passage against its actual material
The materials. Fictional, and thin on purpose:
- M1 — one shot. Forty seconds, time-stamped to the session's posted opening time: Tuesday, 9:00 a.m. The workshop's side door is closed, the handle doesn't turn, and the audio track carries nothing from inside.
- M2 — recorded Thursday. An interview with Dave, the volunteer who holds a key. He says his bus didn't come and that he got there late.
- M3 — same week. A conversation with Maria, a second volunteer. She says she assumed the session had been called off.
- Established context. The workshop is volunteer-run, and the posted schedule lists Tuesday at nine.
- Absent from the record. Any interior footage. Any image from after 9:00 a.m. Any statement by anyone about whether the session happened. Any account of how Maria formed her assumption. Any evidence about what anyone intended.
The draft voiceover:
By nine, the workshop should have been rattling with chains and pumps. Instead, the door was locked. The Tuesday session had been abandoned months ago — the team had given up. Dave, the only keyholder, was stuck on a bus that never came. Maria understood what that silence meant.
Now the audit, line by line.
| Draft line | What it asserts | What the record holds | Who could say it | A version that survives |
|---|---|---|---|---|
| "By nine, the workshop should have been rattling with chains and pumps." | That a normal session is crowded and loud by nine, and that the narrator knows what it ordinarily sounds like | A posted nine o'clock start. That's it — no interior footage, no established normal. | Nobody in this record | "The schedule says the session starts at nine." |
| "Instead, the door was locked." | That the side door was locked at opening time | M1, directly | Reporting presence, or an external guide citing the footage | "Anyone arriving at nine finds the side door locked; the handle doesn't turn." |
| "The Tuesday session had been abandoned months ago — the team had given up." | A collective decision, dated | Nothing. Maria's assumption is about one morning, not a history | Nobody | Cut. |
| "Dave, the only keyholder, was stuck on a bus that never came." | That Dave holds the only key, that no bus came, that he was aboard it | M2 records Dave's account of a delayed bus. "Only" isn't in the record, and his account doesn't put him on the bus. | Dave, in his own voice, as his account | "Dave holds a key. Two days later, he told an interviewer his bus didn't come and that he got there late." |
| "Maria understood what that silence meant." | That Maria knew, and knew correctly | M3 records an assumption she said she made | Maria, as an assumption | "Another volunteer said she had assumed the session was called off." |
Four of five lines fail the same way: the voice reaches past the material for motive and explanation. Notice how tempting the repair is. Say the same thing in a lower register, cut the adjectives, and you have the same problem with better rhythm. Emotionally convincing sentences still need a source for their factual content.
One line, the second, is nearly fine. "The door was locked" is something the camera saw. It needs a small adjustment — "at nine" rather than "instead," so the line stops implying a comparison to a session nobody filmed — and then it can stand.
A word about what silence in a record means. "The notes don't contain it," "the person declined to say it," and "it didn't happen" are three different sentences. The record here is silent about what Dave found at the door. That is not the same as Dave refusing to describe it, and it is not the same as there being nothing to find.
Give the narrator a knowable position
A voice is a position. For factual work you have three, and the pitch should say which one you're proposing.
A participant speaks from experience. Dave can carry his own Tuesday: the bus, the lateness, whatever he saw when he reached the door. He can't carry the interior of the building, the reasoning of other volunteers, or the history of the session. On his own account he wasn't there at nine — that's one of the few things his interview establishes about his whereabouts, and only because he said it. And the moment a participant's account becomes a claim about other people — "nobody else has a key," "the others had stopped caring" — it needs attribution or checking. First person doesn't exempt you; it just moves the source into the frame.
A reporting presence explains an investigation from inside it. This is the person who stood at the door at nine with the shot. They can say I was there, it was locked, I called Dave after that. They cannot say what happened while they were elsewhere, which here is almost everything. This is the voice most likely to borrow authority it hasn't earned, because being at the location feels like being everywhere.
An external guide organizes supported material. It can carry the schedule, the attributed accounts, and the stated gap. It has no built-in signal to the viewer, and that's the trap: an unmarked guide voice defaults in the audience's ear to "the program knows." So the guide has to mark its sources in the sentence itself.
| Position | How the viewer learns it | What earns trust in the line |
|---|---|---|
| Participant | In their own voice and body on camera, talking about their own hours | The account, taken as testimony |
| Reporting presence | Shown at the location, asking the questions, or narrating in first person from the scene | Their own senses, and only those |
| External guide | Nothing, unless the line says so | Names: "the schedule says," "Dave says," "the notes don't record" |
Drift is the usual failure. A passage that starts with "I stood outside the door" and continues into "the team had quietly stopped opening on Tuesdays" has moved from a reporting presence into a position no one on the production occupies. If your series shifts between positions across an episode, that's workable — but the audience should be able to tell which one is talking in any given stretch.
This is also how you answer the pitch question in writing. A model line for the deck: Narration is an external guide. It names its sources — "the schedule says," "Dave says" — and it never describes a room we didn't shoot. Where the record has a hole, the narration says so or the edit carries the jump. That single paragraph does more for a commissioning reader than a page of tone description, because it shows a voice with a knowable position instead of an all-knowing one bolted on to cover thin material.
Write bridges that connect rather than invent
A bridge is narration whose job is to move the viewer from one supported piece of material to the next. It can do four things: stamp a time and place, supply established context, deliver an attributed account, or mark a gap. It cannot supply the missing scene. That's the whole discipline, and it still leaves room to write.
| The bridge's job | The line that invents | What the record holds | The line that holds |
|---|---|---|---|
| Set the scene | "The workshop should have been in full swing." | A posted nine o'clock start | "Tuesday, nine." |
| Link door to keyholder | "Dave's bus didn't come, so the session never opened." | A locked door at nine (observation); Dave's late-bus account (testimony) | "The door was locked at nine. Dave says his bus didn't come." |
| Account for the morning | "The session was called off." | Forty seconds of tape; nothing afterward | "The next thing on the tape is Thursday." |
| Account for the others | "The team had given up." | One volunteer's assumption | "Another volunteer says she assumed it was off." |
The third row is the one people find hardest, and it's the most useful. Where a scene is missing you have three honest replacements. An attributed account — Dave describing his morning, in his own words, on camera — is the strongest, provided he was actually there to see what he's describing. A stated gap is the plainest: the narration says the tape ends here, or a card does. A narrower transition does the least work and costs the least: name the next supported thing, and let the cut carry the movement. "The next thing on the tape is Thursday" is a bridge. It orients the viewer, it makes no claim about the interval, and it doesn't need a single new fact.
Sequence and cause are separate claims, too. Two recorded events happening in order are not one causing the other. Narration that says "and then" is usually safe. Narration that says "because" is making an argument, and in this record the only person entitled to that argument is Dave, about his own lateness. If the sentence asserts causation as the program's finding, the program needs something beyond a locked door and a bus route. And be careful about what adjacency does on its own: cutting straight from the locked door to Dave's bus line makes a causal suggestion the words declined to make. Where that lands is a real editorial choice, not a loophole.
While you're at it: don't manufacture private decisions to smooth a transition. "Maria decided the session wasn't worth the trip" is invented. So is any line that explains an absence by supplying a feeling.
Keep disagreement and uncertainty in the chosen wording
The first thing to notice about these materials is that they don't contain a disagreement. Dave explains his own lateness. Maria reports a belief. Nobody in the record says whether the session happened. A pitch built on "two volunteers tell different stories" has invented a shape, and the invented shape will demand invented resolution later.
The second thing is that detail is not corroboration. The bus account is specific and human; the assumption is vague. Specificity feels like credibility, so the temptation is to let the detailed account become the working version and treat the vague one as a rumor to be cleared up. But a specific account can still be partial — Dave's says nothing about why no one else opened the door — and vagueness isn't falsity. Each sentence keeps its own attribution either way.
Then the wording itself. "Says," "told an interviewer," "assumed," "according to," "the notes don't record" all hold a claim at the strength the record supports. Watch for the verbs that quietly upgrade: understood, realized, knew, confirmed. "Maria understood" converts one volunteer's guess into knowledge, and it does it in a word.
Ask of every line what further evidence would change it. The door being locked at nine: later footage, a second angle, a time-stamped photo. The session being called off: a message to the group, a sign on the door, another volunteer's account. Dave holding the only key: the organization's key list. Dave arriving late: a witness at the door, or records, if the series wants to state it as fact rather than as his account. Maria's assumption: nothing would change it, because it's already fully supported as an assumption — and it will remain the least load-bearing line in the passage, since it's a fact about her belief rather than about the morning.
Last, recheck the words once the images are assembled. A careful line acquires implication from its neighbor. "Dave says his bus didn't come" is defensible on its own. Placed immediately after forty seconds of a locked door, it starts to sound like the explanation for the morning, and the narration didn't say that. Read the passage in position before you defend it on the page.
Which lines stand, which need more evidence, which go
Can stand: the schedule's nine o'clock start, the locked door at nine, the two attributed statements, and the jump from Tuesday to Thursday.
Need more evidence before they can be stated as fact: anything about the rest of the morning, the claim that Dave holds the only key, any statement about whether the session happened, and any causal link between the bus and the closed door.
Should disappear: the months of abandonment, the team's decision, and the meaning of the silence.
And one practice note for the real thing. This example is fictional, and a mere page audit isn't a reliability check. Before describing an actual narrator as accurate, walk the authorized script against the source material with the person who shot it, and route the passage through factual review. Delivery will never tell you which line is true.
Here is the passage with its bases still visible:
Tuesday, nine. The schedule says the session starts then. Anyone arriving finds the side door locked; the handle doesn't turn.
The next thing on the tape is Thursday. Dave, who holds a key, tells an interviewer about his Tuesday: his bus didn't come, and he got there late. Nothing in the record says what he found.
Maria, another volunteer, says she assumed the session had been called off. Nobody in the record says anyone told her so.
Frequently asked questions
What is the central rule for writing factual-series narration?
Every narrated sentence is also a claim about how the narrator knows it. Before choosing a voice, mark what each sentence rests on—something the camera shows, something a person said, context the production has established, or the production's own interpretation. Then ask which of those the narrator can honestly carry.
Why can a well-delivered, calm line still be unsupported?
Confident delivery does not improve evidence; it only makes an unsupported sentence easier to approve. Rewriting the same claim in a lower register or cutting adjectives leaves the same factual content. The line still needs a source for what it asserts.
What are the three narrator positions, and what can each carry?
A participant speaks from experience—their own hours and testimony—but not the interior of a building or other people's reasoning. A reporting presence can say what they saw at the location, such as standing at a locked door, but not what happened elsewhere. An external guide organizes supported material, but it must name its sources in the sentence; otherwise viewers default to 'the program knows.'
How do you write a bridge without inventing the missing scene?
A bridge can stamp a time and place, supply established context, deliver an attributed account, or mark a gap. It cannot supply the missing scene. Honest replacements include an attributed account from someone who was there, a stated gap, or a narrower transition like 'the next thing on the tape is Thursday.' Sequence and cause are separate claims: 'and then' is usually safe, while 'because' is an argument.
What should happen when the record is silent or seems to contain disagreement?
The notes not containing something, a person declining to say it, and it not happening are three different sentences. In the example, nobody says whether the session happened; building a premise on 'two volunteers tell different stories' invents a shape. Keep each sentence at the strength the record supports, and watch verbs like understood, realized, knew, or confirmed, which upgrade a claim.