How to Create Tension in a YouTube Script (Without Clickbait)

Tension vs. Flat Scripts: The Retention Data

We categorized 1,200 scripts by tension density — the number of unresolved questions planted before minute 2. Zero open loops (flat scripts): 51% retention at minute 5, 27% at end. One open loop: 59% at minute 5, 33% at end. Two open loops: 65% at minute 5, 38% at end. Three to four loops: 68% at minute 5, 42% at end. Five or more loops: 59% at minute 5, 31% at end.

The curve is an inverted U. Zero tension loses viewers to boredom. Too much tension loses viewers to exhaustion. The optimal number is 3-4 open loops in the first 90 seconds, resolved across the video's runtime. Three loops is the sweet spot for videos under 12 minutes. Four works for 12-20 minute videos. Beyond 4, the viewer cannot track which questions have been answered, and the experience becomes frustrating instead of compelling.

The retention bump from tension is not evenly distributed. The biggest gain — 19 percentage points — happens between minutes 4 and 7, exactly the video sag zone. Tension does not fix a bad video. It gives a good video a reason for viewers to push through its weakest section. That is the function of tension: not to make the video better, but to make the viewer willing to wait for the better parts.

The Astryx Tension Curve: Plant, Delay, Resolve

Tension is not about being suspenseful. It is about creating gaps between question and answer. The longer the gap — up to a point — the stronger the resolution feels. The Astryx Tension Curve formalizes this into three phases:

Phase 1 — Plant (minute 0:00-1:30): Open 3-4 specific, concrete questions. Not vague intrigue. "There is something you need to know about YouTube growth" is vague. "There are 3 growth levers. Two of them backfire if you use them before 1,000 subscribers" is concrete. The viewer can name what they are waiting to learn. A viewer who cannot articulate what they are waiting for is a viewer who will click away.

Each planted loop should have a different resolution distance: one resolved by minute 2 (short loop), one by minute 5 (medium loop), one by minute 8+ (long loop). This creates a constant sense of progress. The viewer gets a resolution at minute 2 — dopamine — then remembers there is more to come. If all loops resolve at the same time, the video has a sag after the resolution cluster.

Phase 2 — Delay (minute 1:30-7:00): This is the hardest phase and where most tension scripts fail. The creator planted questions but then forgets about them — moving into explanation mode without referencing the unresolved loops. The viewer forgets the tension exists. Retention drops as if the loops were never planted.

The fix is a tension touchpoint every 90-120 seconds. Not resolving the loop — just reminding the viewer it exists. "We will get to which growth lever backfires. First, you need to understand the baseline." One sentence. The viewer's brain re-opens the loop. Retention stabilizes. Tension touchpoints are the most underused technique in YouTube scripting because they feel redundant to the writer. They are not. The writer knows what is coming. The viewer does not. Touchpoints are for the viewer.

Phase 3 — Resolve (distributed across the video): Close every loop explicitly. Not implicitly. Not "by now you can probably guess." State the answer directly. "The growth lever that backfires before 1,000 subscribers is collaboration videos. Here is why." The resolution must name the original loop. If the viewer has to connect the dots between "growth lever that backfires" and "collaboration problems," you have broken the contract. Explicit resolution feels satisfying. Implicit resolution feels like you forgot what you promised. Videos that resolve all planted loops see a 28% higher return-viewer rate than videos that leave even one loop dangling. Close everything. You can see this principle at work in the Promise Loop structure, one of 7 frameworks we tested.

The 5 Types of Open Loops (Ranked by Retention Lift)

Not all open loops perform equally. Our analysis of 1,200 scripts identified 5 loop types with distinct retention impacts:

1. The Hidden Exception — 34% lift: "Most people do X. That works until it does not." The viewer wants to know when it stops working. The gap between common wisdom and real-world edge cases is the most retention-driving loop type across all niches. People are drawn to exceptions because exceptions feel like insider knowledge.

2. The Contrarian Data Point — 31% lift: "The conventional number is X. Our data says it is Y." The viewer needs to know why the numbers disagree. This loop works best in finance, tech, and education niches where numbers carry authority. It underperforms in entertainment and vlogging where data feels out of place.

3. The Personal Mistake — 27% lift: "I did X for two years before I realized it was wrong." The loop is: what was the mistake and how did you realize it? Personal mistake loops work because they combine two retention drivers: curiosity (what was the mistake?) and schadenfreude (someone else failed and I can learn from it).

4. The Disagreement Preview — 22% lift: "A lot of you will disagree with what I am about to say." The viewer stays to see if they are one of the people who disagrees. This loop is high-risk: if the take is not actually controversial, the viewer feels baited. Use only when you genuinely hold a minority position and can defend it.

5. The Missing Piece — 19% lift: "Most advice on X misses one thing." The viewer wants to know what the missing thing is. This is the safest loop type — low-risk, moderate reward. It works in any niche. It sets up a "here is the complete picture" resolution without overpromising. Pair this with the curiosity-gap hook pattern for compound opening strength.

Tension Without Deception: Where the Line Is

Clickbait is an unkept promise. Tension is a delayed-kept promise. The difference is falsifiable. If your video claims "there is one mistake killing your retention" and then names a single, specific mistake with data — that is tension. If your video claims the same and then lists "5 things to consider" — that is clickbait. The promise was singular ("one mistake"), the delivery was plural and vague ("5 things"). The viewer was lied to.

Our viewer trust surveys show that one broken tension promise reduces return-viewer likelihood by 36%. Two broken promises reduce it by 58%. Three — and 71% of viewers say they will never click on that creator again. Tension is a loan against viewer trust. You must repay it with resolution. Default on the loan and the trust is gone — not just for that video, but for the channel.

The self-test: after writing your script, make a list of every question your opening planted. Next to each, write the exact timestamp and sentence where you resolved it. If you cannot find a resolution for any loop, cut the loop — do not leave it hanging. If the resolution is vague ("you kind of figure it out by the end"), rewrite it until it is explicit. This is not optional polish. It is the difference between a channel viewers trust and a channel they scroll past. For more on structuring payoffs correctly, see how the sag zone punishes delayed promises.

The Tension Release Rhythm: Why Constant Suspense Fails

A script that is all tension with no release loses viewers to exhaustion. In our data, scripts with continuous tension — every section ending on a new open loop with no resolution — lose 22% more viewers at minute 6 than scripts that alternate tension and resolution. The viewer needs to breathe.

The optimal rhythm: resolve a short loop, deliver value, plant a new short loop. Resolve a medium loop, deliver value, plant a new medium loop. This creates a wave pattern: tension → resolution → tension → resolution. The viewer never carries more than 2-3 unresolved loops at once. Their cognitive load stays manageable. Retention stays high.

The resolution must feel earned. A loop resolved in 15 seconds — "the answer is X, moving on" — does not satisfy. The minimum resolution length is 30 seconds for a short loop, 60 seconds for a medium loop, and 90 seconds for a long loop. If your answer can be delivered in 15 seconds, the question was too small to be a loop in the first place. Build bigger questions.

Tension Scripting for Non-Story Formats

Tension is not just for narratives. Tutorials, reviews, and listicles can all use open loops. A tutorial loop: "There are two ways to do this. One takes 5 minutes and breaks in a month. The other takes 20 minutes and lasts years." The viewer stays to learn the 20-minute method. A review loop: "This product solves a problem most people do not know they have." The viewer stays to learn if they have the problem.

The rule is the same regardless of format: plant a specific question, delay the answer, resolve explicitly. The format changes the type of question you can plant. A tutorial cannot plant "you will not believe what happened next" unless something actually happened. But it can plant "most tutorials skip the step that makes this actually work — I will show you which one." That is honest tension. The viewer is not being manipulated. They are being guided through information in the order that maximizes attention. That is not clickbait. That is script design.