Case Study 2: Finding the Story — Structuring a Documentary Short from Raw Material
This is a from-scratch production walkthrough — an illustrative, constructed teaching example labeled throughout. Where Case Study 1 took apart a finished film, this one builds a story from nothing: a messy pile of interview fragments and observed moments, structured into a three-act spine with a hook, shown evolving through several drafts. It is set in the book's talking-head / interview setup and points straight at your Project 2.
The brief and the constraints
You have been asked (by yourself — this is your Project 2) to make a 3-minute documentary short about a person with a craft and a purpose. You have found your subject: a volunteer who runs a small community bike shop out of a rented garage, taking in donated bikes, fixing them up, and giving them to neighborhood kids who need a way to get around. One person. One interview. B-roll of the work. No narrator planned. A phone and a clip-on mic would be enough to shoot it.
Here is the trap, and almost every beginner falls into it: you are tempted to go shoot first — get the interview, grab pretty footage of hands on wrenches — and "figure out the story in the edit." Sometimes you find one. Usually you come home with three interesting hours and no spine, and you spend a week in the timeline discovering you never captured the one beat the story needed. This case study is the cheaper path. Before a single frame is shot, we are going to find the story on paper — which is the entire lesson of Chapter 17 and the deliverable of its Production Checkpoint.
The raw material below comes from a pre-interview: a no-camera conversation you have with your subject beforehand (a technique you will formalize in Chapter 19) to learn what the story might be. You take notes. What you get is ore, not metal — unordered, uneven, and full of buried gold.
The raw material (the ore)
Here is what you wrote down, roughly in the order it came up — which is to say, in no useful order at all. (Fragments are paraphrased; this is a constructed example.)
FROM THE PRE-INTERVIEW (unordered notes)
- "I've been doing this about eight years now."
- "We take in maybe 200 bikes a year. Fix what we can. Give them to kids who need them."
- "I learned to wrench from my dad, in our garage, when I was ten."
- "My first bike — that was the first place I could go on my own. That's the whole thing, really."
- "There was a stretch, maybe two years ago, I almost shut it down. Landlord tripled the rent.
I lost the space. I had forty bikes stacked in my own living room for months."
- "A kid came by during that time. His bus route got cut — he couldn't get to his after-school
job anymore. I built him a bike in one night. He kept the job. That's when I knew I couldn't quit."
- "Now we've got a real space, three volunteers, and a waitlist of names on the wall."
- "The mechanical part is the easy part. It's always the parts, the money, the space."
- "People say I fix bikes. I don't. I hand a kid the keys to their own neighborhood."
OBSERVED MOMENTS (B-roll candidates you noticed)
- His hands truing a wheel, plucking spokes like strings, listening.
- A wall of donated frames, each tagged with a paper label.
- A whiteboard with a waitlist of kids' first names.
- Him arriving before anyone, unlocking the roll-up door, coffee in hand.
- A bin of mismatched parts being sorted by feel.
- A kid standing over a bike that finally fits — the exact moment it becomes theirs.
Read that pile and you can feel the problem: there is obviously a story in here, but the notes are a transcript, not a structure. Some lines are facts (200 bikes a year). Some are gold (the kid who kept his job). One is a whole thesis (the keys to their neighborhood). If you shot this in the order it was said, you would have a list. Our job is to find the arc.
Phase 1 — Draft 0: the chronology (what not to do)
The instinct is to arrange the material in the order events happened. Let us do that, honestly, so we can watch it fail.
DRAFT 0 — THE CHRONOLOGY (a timeline, not a story)
1. He learned to wrench from his dad at ten.
2. His first bike gave him independence.
3. Eight years ago he started fixing donated bikes for kids.
4. Two years ago he nearly lost the space to a rent hike.
5. A kid needed a bike to keep a job; he built one overnight.
6. Now there's a real space, volunteers, and a waitlist.
7. "I hand kids the keys to their neighborhood."
This is not terrible — the events are all here — but it is flat, and you can say exactly why using §17.6's test: every beat connects to the next with "and then." He learned to wrench, and then he got a bike, and then he started the shop, and then rent went up, and then a kid came by. Nothing causes the next thing; nothing is at stake early; and the single most gripping moment (the kid who kept his job) is buried at position five, where no one impatient will ever reach it. Worst of all, it opens on the least gripping possible beat — a man's childhood — and asks the viewer to wait through biography for a payoff it hasn't promised. Draft 0 is a chronology dump. We can do far better with the same nine facts.
Phase 2 — Draft 1: impose three acts (find the change)
Stop asking "what order did things happen?" and start asking the two Chapter 17 questions: what changes across this story? and what are the three acts? Read the ore again for a narrative arc — a start-state and a different end-state.
The change is there once you look for it. At the start, this is a guy fixing bikes as a nice hobby. By the end, it is a mission he knows he cannot quit — and the thing that changed him is the kid who kept his job. That is the arc: from a hobby to a calling. The turning point is not the rent hike (that is a complication); it is the moment the stakes became human. Now the three acts almost write themselves:
DRAFT 1 — THREE ACTS (arc found: hobby → calling)
ACT 1 (setup + inciting problem)
- He fixes donated bikes for kids out of a garage. 200 a year. [the normal]
- Two years ago the rent tripled; he lost the space; forty bikes in his living room. [the problem]
ACT 2 (development — it gets harder, then the turn)
- He nearly shut it down. The mechanical part was never the hard part — it's parts, money, space.
- THE TURN: a kid's bus route got cut; he'd lose his job; the volunteer built him a bike overnight.
The kid kept the job. "That's when I knew I couldn't quit."
ACT 3 (resolution — the new normal + the meaning)
- Now: a real space, volunteers, a waitlist of names on the wall.
- "I don't fix bikes. I hand kids the keys to their own neighborhood." [the change landed]
Already this holds where Draft 0 sagged. There is a problem in Act 1 that makes us worry (he might lose it), a genuine turn in Act 2 (the kid), and a resolution in Act 3 that lands a changed man (hobby → calling). Notice what we did with his childhood and his dad's garage — we did not throw them away; we set them aside, because they are backstory, not the spine. We will find a home for them in a moment, as the B-story. Notice too that we broke chronology on purpose: the rent hike (which happened after years of quiet operation) now sits in Act 1 as the inciting problem, because that is where it does the most structural work.
One decision in Draft 1 deserves a name, because beginners get it wrong constantly: the turn is the human beat, not the logistical one. The rent tripling is dramatic, but it is a circumstance — it happens to the subject. The kid who kept his job is a choice — it is the moment the subject decides who he is going to be ("that's when I knew I couldn't quit"). Audiences follow decisions, not circumstances. So when you hunt for your Act 2 turn, look past the biggest event and find the biggest choice: a choice is where a person changes, and change is the whole arc.
Phase 3 — find the hook (move the gold to the front)
Draft 1 still opens on "he fixes bikes for kids," which is fine but not gripping. Somewhere in the ore is a stronger first beat. Scan for the line that opens the biggest loop. Two candidates jump out: the thesis line ("I hand kids the keys to their own neighborhood") and the crisis image ("forty bikes stacked in my living room").
The thesis line is beautiful but it is a conclusion — leading with it would front-load the ending and close the loop before we start (the mistake from §17.1). Save it for the payoff. The crisis image, though, is a perfect hook: it is concrete, strange, and it opens a question — why is there a wall of bikes in this man's living room? Lead with that, in his own voice or in a shot, and the viewer has to stay to find out.
DRAFT 2 — HOOK MOVED TO THE FRONT (in medias res)
HOOK "For six months, I had forty kids' bikes stacked in my living room." [open the loop]
ACT 1 Back up: who he is, what the shop is, 200 bikes a year for kids who need them. [setup]
ACT 2 Why the living room: the rent tripled, he nearly quit — and then the kid who kept his job. [turn]
ACT 3 Now: the real space, the waitlist. "I don't fix bikes. I hand kids the keys to their neighborhood."
This is the in-medias-res move: open at the most gripping moment, then circle back to fill in the setup. The living-room image is not where the story starts chronologically — but it is where the story starts best, because it opens the widest gap between what the viewer knows and what they want to know. We found the hook by hunting the ore for the beat that asks the loudest question, exactly as §17.3 prescribes. And — the practical payoff — we found it on paper, so on the shoot we will know to make sure we capture that line cleanly and grab a shot that shows it.
Phase 4 — braid the A-story and the B-story
We have a spine, but it is running on the A-story alone — the what: bikes fixed, space lost, space regained. The piece will inform, but to move, it needs its B-story — the why underneath. Look back at the two fragments we set aside: "I learned to wrench from my dad in our garage" and "my first bike was the first place I could go on my own." There it is. The B-story is about what a bike means: freedom, independence, a kid's first taste of the world on their own terms. That thread is the emotional engine, and it explains why this particular man does this particular work.
Now we braid: at each act, an A-beat (the shop's fortunes) carries the facts, and a B-beat (what a bike means) carries the feeling — and they illuminate each other.
DRAFT 3 — A-STORY AND B-STORY BRAIDED
A-STORY (the what: the shop) B-STORY (the why: what a bike means)
HOOK "Forty bikes in my living room." ────────── (a promise: this matters more than it looks)
ACT 1 the shop; 200 bikes a year ├─────── "My first bike was the first place I could
for kids who need them go on my own. That's the whole thing."
ACT 2 the rent tripled; he nearly quit ├─────── the kid whose bus route was cut — a bike was
THE TURN: the kid who kept his job the difference between keeping a job and losing it
ACT 3 the real space; the waitlist └────── "I don't fix bikes. I hand kids the keys
to their own neighborhood." (A and B become one line)
Look at what the braid does to the ending. "I hand kids the keys to their own neighborhood" is the exact point where the A-story (fixing bikes) and the B-story (freedom) fuse into a single sentence — which is why it is the perfect final beat. The whole film has been quietly building both threads so that this line lands as a summation the viewer has already felt, not a slogan. That is how you earn a thesis line: braid its two threads for three minutes, then tie them in the last beat. The dad-and-garage detail now has a home too — it belongs in the B-story of Act 1, explaining where his belief came from without derailing the spine.
Phase 5 — the but/therefore test, the spine, and the logline
One last pass to make sure we have a story and not a decorated list. Run §17.6's connective-tissue test: can we link the beats with "but" and "therefore" instead of "and then"?
He runs a shop giving bikes to kids, but the rent tripled and he nearly lost it, therefore he was ready to quit — but then a kid needed a bike to keep his job, therefore he built one overnight and understood he couldn't quit, therefore he rebuilt the shop for good.
It passes. Every beat now causes the next; the outcome is in doubt through Act 2; and the change (ready-to-quit → can't-quit) is real. Compress the whole thing into a logline:
A volunteer mechanic who nearly lost his community bike shop rebuilds it around a single truth he learned from one kid: a bike isn't transportation — it's a child's first taste of freedom.
That one sentence contains the subject, the stake (nearly lost it), the turn (one kid), and the change (a truth learned). If a stranger read only that sentence, they would know what the film is about and want to see it. That is how you know your spine is sound — and we have it before spending a minute of shoot time.
The capture plan: shoot to fill the beats
Now — and only now — the spine tells you exactly what to shoot. This is where Chapter 17's paper work hands off to the interview craft of Chapter 19 and the B-roll craft of Chapter 20. Because you know your beats, you know your questions and your shot list. You are not "getting an interview"; you are capturing the specific soundbites and images each beat needs.
⚙️ Settings Box — the interview capture (talking-head setup, previewing Ch.19).
Element Starting point Why Framing Medium close-up, subject on a third, off-axis eyeline just beside the lens Intimate; leaves room for lower-third name later Lens / aperture ~50mm-equivalent, f/2.8 Natural perspective; soft background separation Resolution / fps 4K, 24 fps, 180° shutter (1/50 s) Cinematic motion; room to reframe in the edit Light Window as soft key camera-left; bounce fills the shadow side Free, flattering, motivated by the space Audio Lav on the subject plus a shotgun as backup; monitor on headphones Sound is half the picture; two-mic safety Room tone 30 seconds, everyone silent You will need it to smooth every cut A starting point, not a recipe — adjust to your room. The real point: you know these settings serve specific beats you already outlined.
Your questions now target beats, not topics. To capture the hook, you ask something that produces the living-room line as a full sentence: "Walk me through the winter you almost lost the shop — where did all the bikes go?" To capture the turn, you ask about the kid directly and wait for the full-sentence answer (Chapter 19's core skill). To capture the B-story, you ask, "What did your first bike mean to you?" To capture Act 1's normal, you ask for a number and a picture at once: "How many bikes come through here in a year, and where do they all come from?" To capture the resolution, you invite the thesis in his own words: "When someone says you just fix bikes, what do you tell them?" — and if the "keys to their neighborhood" line arrives, your closing beat is on tape. Notice that every question is aimed at a specific beat you already know you need; you are not fishing, you are filling a spine. And for each beat, your B-roll list already exists, straight from the ore's observed moments:
BEAT → B-ROLL (the show-don't-tell shot each beat needs)
HOOK (the crisis) → a wall/stack of tagged frames; the roll-up door opening on them
ACT 1 (the work) → hands truing a wheel, listening to the spokes; the whiteboard waitlist
ACT 1 (B-story) → an old photo or a worn tool; his hands doing what his dad's hands did
ACT 2 (the turn) → a single bike being built late, one lamp on; an empty bus-stop sign
ACT 3 (the payoff) → a kid at the exact moment a bike fits and becomes theirs
Every one of those shots shows what an interview line says — the feeling on the audio, the proof in the picture (§17.4). You are shooting for the edit before you have shot a frame.
How it will cut
Because the story is found, the edit is now mostly assembly. In Part VI you will build this as a paper edit (Chapter 30): pull the select soundbites, lay them in the spine's order, then cover every cut with the B-roll you gathered. Here is the shape it will take — a beat map that is already, in effect, your timeline.
FIGURE CS2.1 — The cut, as a beat-map timeline (previewing Ch.30)
V1 (B-roll) [door/bikes][ hands/wheel ][ old tool ][ late build ][ empty stop ][ kid + bike ]
A1 (sync) [ HOOK line ][ setup: the shop ][ B: first bike ][ the rent / the turn ][ payoff line ]
A2 (music) [ ................ soft bed, rising under the turn, resolving under the payoff .......... ]
▲hook ▲setup ▲B-story ▲Act2 turn ▲Act3 change
open the loop who/what why he cares stakes + decision loop closes
The interview audio (A1) is the spine; B-roll (V1) shows each beat and hides the cuts between
soundbites; music (A2) carries the emotional arc. This entire structure existed on paper first.
Notice that the edit's shape is the spine's shape — because the spine is the edit, drafted in advance. The living-room hook opens the loop on V1's stacked bikes; the B-story of the first bike plays under a worn tool; the turn lands on a single late-night build; and the thesis line closes the loop over the one shot you most needed to get — a kid receiving a bike. If you shot the beats, this cut nearly assembles itself. If you had gone out and "figured it out in the edit," you would be praying that shot exists. It does, because you planned the story first.
A note on holding the plan loosely
This spine is a plan, not a cage. You built it so you would have something to shoot against and so the shoot day would be calm — but nonfiction has a way of handing you a truer story than the one you outlined (§17.5's reality arc). Maybe, on the day, the volunteer says something about his own kid that reframes the whole piece. Maybe the waitlist of names on the wall turns out to be the real heart of it. Because you planned an arc, you will recognize the better one the instant it appears — an unstructured shooter films everything and notices nothing. And because your spine told you to shoot generous coverage of every beat, you will actually have the footage to tell the truer story if it shows up. That is the paradox of planning nonfiction: you structure it carefully precisely so you are free to abandon the structure for something better. Plan the arc; stay awake for a better one.
Discussion questions
- Draft 0 (chronology) contained all the same facts as the final spine. Precisely why did re-ordering them — not adding anything — turn a list into a story?
- We opened on "forty bikes in my living room" rather than the thesis line about "keys to the neighborhood." Defend that choice using the difference between a hook and a payoff.
- The dad-and-garage detail was cut from the spine, then restored as B-story. When is a true, interesting detail wrong for the main structure — and how do you decide?
- The logline names a stake ("nearly lost his shop") and a turn ("one kid"). Write an alternative logline that foregrounds the B-story (freedom) instead of the A-story (the shop). Which is the stronger film?
- Everything here was decided before the shoot. What is the risk of over-planning a nonfiction story, and how does the "reality arc" idea (§17.5) protect you from it?
- Look at the beat-to-B-roll list. Pick one beat and argue that a different B-roll shot would show it better. What is your shot showing that the listed one isn't?
Your turn: structure your own ore
Take your Project 2 subject through the exact five phases in this case study.
- Gather the ore. Do a no-camera pre-interview (or use notes you already have). Write down fragments in whatever order they come — quotes, facts, observed moments. Do not organize yet.
- Draft 0 → Draft 1. Write the chronology, then find the change and impose three acts. What is different about your subject (or their world) at the end than the start?
- Find the hook. Hunt the ore for the beat that opens the biggest question. Move it to the front; make sure it's a hook, not a spoiler.
- Braid A and B. Name the surface "what" (A-story) and the deeper "why" (B-story). Assign each fragment to one thread. Find the beat where they fuse — that's your ending.
- Test and compress. Run the but/therefore test. Fix any "and then." Write the one-sentence logline. If you can't, the story isn't found yet — keep digging.
Keep the result. It is your Production Checkpoint deliverable and the spine of everything you shoot in Chapters 18–20.
Key takeaways
- Find the story on paper, before the shoot. A pre-interview yields ore; structuring it into a spine is the cheapest, highest-leverage work in the whole project.
- Chronology is not story. The same facts, re-ordered from "and then" into "but/therefore," turn a flat timeline into an arc that pulls.
- Find the change first. The arc (here: hobby → calling) is the backbone; the three acts organize it; the turn is the human moment, not the logistical one.
- The hook is usually buried in the ore. Hunt for the beat that opens the biggest question (the living-room bikes), and lead with it — a hook, not a payoff.
- Braid the A-story and B-story, and end on the beat where they fuse into one line.
- A found spine tells you exactly what to shoot — the questions, the settings, and the show-don't-tell B-roll for every beat. You shoot for the edit before you shoot at all.
- The whole method is a constructed example, but the process is real and repeatable: ore → chronology → three acts → hook → braid → spine → logline. Run it on every nonfiction piece you make.