Why tutorial-style hooks and loops don't fit, what to use instead, plus a worked example on why time seems to have sped up. Research brief, November 2026.
Most YouTube scripting advice, like Humble&Brag's script template, is built for a promise → delivery video: "you have a problem, stick around, here are 5–7 payoffs, ascending in value." That works when the viewer's question is "how do I…?". A documentary viewer's question is different: "what's really going on here?". The payoff isn't a tip. It's a shift in understanding.
One central question → the answer everyone believes → the fact that breaks it → a better explanation, built step by step → what it means for the viewer. Each act is linked by but / therefore, never and then, and anchored to a person or a concrete scene rather than a list of facts.
This isn't one guru's system. Script-tool guides [1], documentary teachers [2][3], Veritasium's research on teaching [4], Johnny Harris's process [5], and analyses of investigative YouTube [6] all arrive at the same shape from different directions.
| Tutorial / self-help (Humble&Brag) | Documentary / video essay |
|---|---|
| Engine: "What will I get?" curiosity loops around tips | Engine: "What's the real answer?" one big question kept open across the whole video [1][6] |
| 5–7 independent Setup-Tension-Payoff loops | 3–5 acts that each overturn the last one. "If an act only adds information, it is a segment rather than an act, and segments are where long films sag." [1] |
| Ascending value: save your best tip for last | Ascending understanding: save the reframe for late in the video. The essay "says what it all means only late" [6] |
| Hook: contrarian one-liner, "here's what no one tells you" | Cold open: a scene, artifact or moment shown before it's explained [6][5] |
| Intro: "By the end of this video you'll have…" | Harris avoids telling viewers what he's going to tell them. The first line drops into action [5] |
| Summary recaps 3–5 takeaways | Ending is reflection: what changed, what stays unknown, what it means [7][1] |
| Expert-to-student | Curious friend investigating alongside the viewer [5][8] |
Some of it still applies: write the payoff before the setup, cut anything that doesn't serve the spine, read it aloud, and match the title promise early. Harris reinforces the thumbnail's promise "within the first minute" and answers it by the end [5]. What changes is the structure in between.
Bernard's core distinction is between a subject and a story: "Not the subject: the story." A subject is "time perception." A story has a beginning, middle and end, and it pulls the viewer into wanting to know what happens next [2]. She also says to start from questions, not a thesis to prove. Films that open with "I intend to prove…" tend to play only to people who already agree [2].
Before you outline, write these four lines (adapted from [1]):
Everything that doesn't move the viewer from line 2 to line 3 is a candidate for cutting [1].
This is the most YouTube-specific version I found [1]. It lines up with Derek Muller's finding that the misconception has to come first. In his PhD research, presenting the facts directly made students more confident in their wrong ideas. Showing the misconception first and then working through it improved their test scores, even though students said they felt confused [4].
| Act | Job | ~Runtime |
|---|---|---|
| 1. The case (cold open + question) | A concrete scene, the question in one sentence, and one line on why the obvious answer fails | ~10% |
| 2. The accepted answer | Present the best version of what everyone believes, fairly and with sources. This is Muller's misconception step | ~20% |
| 3. The crack | A study, date, person or detail the accepted answer can't explain. This is CBC's "destabilization," where the story turns out to be about something different [7] | ~30% |
| 4. The rebuild | The new explanation, assembled step by step. Each step answers the objection the previous one raised | ~25% |
| 5. The cost / meaning | What the answer changes for real people, what remains unknown, and your reading of the answer. Burns calls the ending "send them home safe" [3] | ~15% |
In a doc, a re-hook tells the viewer what the next act will overturn. "'So that is the official story' offers a place to stop; 'the official story explains everything except the invoice' does not." [1] Harris does the same thing through foreshadowing, hinting that more is going on without saying "stick around" [5].
The runtime percentages come from a script-tool blog and are guidance, not a rule. The sequence is what matters.
This replaces the Setup-Tension-Payoff loop. The rule comes from the South Park creators: if the words "and then" belong between your beats, the story is boring. The link between beats should be "but" or "therefore" [9][8]. Randy Olson adapted it for science communication as And / But / Therefore. Most experts talk in "this AND this AND this," when the audience needs "this AND this, BUT the problem is this, THEREFORE…" [10].
Write every beat on its own line. Between each pair, write the connecting word. Each and then is a place viewers are likely to drop off. Rewrite it as a complication (but) or a consequence (therefore), or cut the beat. Kallaway calls this "the dance" between context and conflict: conflict opens a loop and context closes it [8].
An analysis of investigative YouTube essays finds the sentence-level rhythm to be "claim → evidence → implication" [6].
For a topic like time perception, the protagonist can be the viewer (second person: "Think about 2021. Can you name three things that happened?"), one real person whose years blurred, or you as the investigator.
Harris describes his mode as contemplative rather than explainer. He alternates between two voices, which he calls "sweet and sour" [5]:
| Explainer voice | Poetic voice | |
|---|---|---|
| Precision | Exact, technical | Softer, less precise |
| Offers | Firm answers | Open questions |
| Who acts | Faceless forces | People in gray areas |
| Viewer keeps | The facts | What they felt |
A topic like time perception fits this well: explain the memory science, then pause and let the viewer feel it.
From a breakdown of investigative YouTube [6]:
"What unites them is that the cold open makes a promise about a payoff to come and deliberately withholds the answer." [6] Veritasium's Petr Lebedev: make the first minute "the most interesting thing you've made in your life" [12].
| Shape | How it works | Best for |
|---|---|---|
| Whale [7] | Setup with stakes → point of no return → destabilization → climax → reflection | Events unfolding in real time |
| Spiral [7] | Circles one event or idea. Each pass goes deeper until the story turns out to be about something else | Abstract topics. Each lap around "why does time feel fast?" adds a layer (memory, routine, stress, meaning) |
| Quest [7] | A protagonist searching for an answer, with obstacles, discovery and an ambiguous but satisfying end | Presenter-led investigation |
| Thriller arc [6] | Hook/inciting incident → rising action with a red herring → revelation | True crime, mysteries |
| Thesis arc [6] | Context → escalating evidence → analytical reframe → thesis | Systems and culture (Wendover, Folding Ideas) |
| Chronology [5][3] | History gives the backbone, so the work is curation. Burns calls chronology "a compass" | History pieces (about a third of Harris's videos) |
Many top videos mix these. Harris uses thriller pacing on thesis content [6]. For abstract, idea-driven topics, a good fit is the 5-act frame with spiral layers inside Act 4.
The late-reveal structure is persuasive whether or not the evidence holds up. "The place to judge one is the evidence before the turn." [6] For a science topic, make sure the crack in Act 3 is a real finding and not a dramatized one.
A draft spine to react to, not a script. I only pulled a handful of time-perception sources to test the structure, so a real episode would still need its own research pass.
| Act | Draft beat |
|---|---|
| 1. Case | Second-person cold open: "Name one thing that happened in 2021." The viewer can't, and that's the hook. Question: "Why do the last few years feel like they vanished?" |
| 2. Accepted answer | "You're just getting older." Present it fairly: the proportional idea, and Wittmann & Lehnhoff's survey of 499 people aged 14–94, where older people felt decades passing faster [14]. Add "phones and the pandemic" as the second popular explanation. |
| 3. The crack (BUT…) | Age doesn't fully hold up. Weeks, months and even a year didn't feel faster with age in that same survey [14], and a Christmas/Ramadan study found no link between being over 55 and feeling time pass faster [15]. The pandemic doesn't fit neatly either: people rated that period as slower, longer and further away [16]. So why does it feel gone? |
| 4. Rebuild (spiral layers, THEREFORE…) |
Layer 1, memory density: the brain encodes new experiences, not routine ones, and we judge past durations by how many new memories we made. This is Claudia Hammond's "holiday paradox" [14]. Layer 2, in the moment vs looking back: those are two different clocks [14][16]. Layer 3, collective disruption: in a US national sample, over 65% reported blurring days and uncertainty about the future six months into the pandemic, and half said time sped up or slowed down [17]. Mood was the main predictor of distortion [16]. Blursday measured this across 9 countries [18]. Layer 4, the twist: periods remembered as most meaningful are often the ones that felt fastest [19]. |
| 5. Meaning | The review ends where an episode like this can: understanding why life feels fast "may help clarify what a fast life signifies—and when it reflects loss, fulfillment, or some combination of the two." [19] End on a reflective note rather than a list of tips. One "make new memories" idea can close it, but it shouldn't be the point. |
We feel time sped up, and everyone blames age. But age doesn't explain recent years, and the lockdown felt slow while it was happening. Therefore the speed must be coming from memory, not the clock. But fast years aren't always empty, because meaningful ones fly too. Therefore a fast life can be a loss or a sign that you were living well.
Spine (write first): Question · Belief at start · Belief at end · Evidence that forces the change
Writing order: Ending line → Act 3 crack → Act 4 rebuild → Act 2 → cold open last
Act 1, Case (~10%): Scene, artifact or second-person moment → the question in one sentence → why the obvious answer fails. Match the title's promise within the first minute.
Act 2, Accepted answer (~20%): The strongest, fairest version of what people believe. Re-hook: what's about to break it.
Act 3, Crack (~30%): A specific, quotable fact or person the accepted answer can't explain.
Act 4, Rebuild (~25%): New explanation in layers. Each one answers the objection the last one raised. Use one human anchor per layer.
Act 5, Meaning (~15%): What changes, what's unknown, your reading of the answer, and a callback to the cold open.
Every beat: linked by BUT or THEREFORE · claim → evidence → implication · plain who-did-what · varied sentence length · two columns (words | visuals) · red until cited.
Evidence note: the craft sections draw on practitioner advice (Harris, Burns, Bernard, Kallaway, Parker & Stone), one PhD-based teaching finding (Muller), and two script-tool or analysis blogs (Prepublish, Canopticon). Prepublish's act percentages are guidance, not tested benchmarks. The time-perception studies in section 11 were read at abstract or summary level only and need a full pass before scripting.