Format guide
Explain it like the viewer is smart but busy
Explainer videos fail when visuals decorate instead of clarifying — motion for motion's sake while the script rambles. This guide covers script-first editing, visual simplification, kinetic text discipline, and the sound and pacing choices that carry viewers to the call to action.
From the Mellius product team · Updated September 28, 2026
Conspiracy theorists, you may hate this
AI Director
›
Just type what you need
- ›Remove all silent parts
- ›Add bold kinetic captions
- ›Only have Bill Ackman speaking
- ›Cut to highlights only
- ›Add cinematic transitions
- ›Match cuts to the beat
The short answer
Lock a short script built as problem, mechanism, proof, and next step, then cut visuals that show each sentence. Keep one idea per shot, animate text for key terms only, pace narration near 140 words per minute with pauses at transitions, and mix voice far above music so every word lands.
Script supremacy
Every visual exists to clarify a specific sentence; nothing moves just because motion is available.
One idea per shot
Viewers grasp each concept before the next arrives, with transitions that signal topic changes.
Effortless audio
Voiceover is clean, paced, and clearly above music so comprehension never requires effort.
Edit from the script, not the asset pile
Record or lock scratch narration first, then cut the timeline to its sentences: sixty to ninety seconds suits most product explainers, which means roughly 140 to 200 words at a comfortable pace. Break the script into problem, mechanism, proof, and next step, and give each its own visual chapter with a distinct transition so viewers feel progress. Delete any sentence the visual already communicates — narration should add what pictures cannot say.
Simplify relentlessly: one chart per point with at most three labeled elements, icon-driven diagrams instead of screenshots where possible, and screen recordings cropped tight to the relevant control with cursor movement smoothed. Hold each visual two to five seconds so the eye finishes reading before the cut. In Mellius you can ask to restructure around a shorter script or regenerate pacing conversationally, keeping every graphic and caption editable on the timeline with undo for each experiment.
- Lock narration before animating anything beyond placeholders.
- Give each script chapter its own visual treatment and transition.
- Reduce charts to three labeled elements maximum.
- Crop screen captures to the control that matters.
Animate text sparingly and mix voice first
Reserve kinetic text for the terms viewers must remember — the product name, the core mechanism, the numbers — and keep body explanations in stable, readable captions. Animate with one consistent grammar: same easing, same entrance direction, same timing grid, so motion feels designed rather than sampled from presets. Sync graphic entrances to the word that introduces them, landing within a few frames so eye and ear confirm each other.
Mix the voice to peak around -6 dB with music ducked ten to twelve decibels beneath, then check the whole piece on laptop speakers and a phone. Add pauses of half a second at chapter transitions; they cost runtime and repay attention. End with a single on-screen call to action held three full seconds — URL or button, spoken aloud and shown as text — because split or rushed endings waste the comprehension you just built.
Compare the workflows
| Style | Best fit | Trade-off |
|---|---|---|
| Animated motion graphics | Abstract concepts and products without filmable UI | Requires illustration and animation time |
| Screen-recorded walkthrough | Software products viewers must see working | Raw captures need tight cropping and pacing |
| Presenter plus graphics | Trust-led topics where a face carries credibility | Shoot quality and presenter delivery set the ceiling |
A practical sequence
Put it into practice
- 01
Lock the short script
Write problem, mechanism, proof, and next step in under 200 words, then record scratch VO.
- 02
Visualize sentence by sentence
Show each sentence with one simple graphic, holding long enough to read before cutting.
- 03
Finish for comprehension
Animate key terms only, mix voice above music, and hold the call to action for three seconds.
Frequently asked questions
How long should an explainer video be?
Sixty to ninety seconds fits most product and concept explainers, holding attention through one clear arc. Complex B2B topics can stretch to two minutes if each chapter earns its time. Anything longer usually wants splitting into a series rather than a single heavier cut.
Should I use voiceover or on-screen text in explainers?
Use both with distinct jobs: voiceover carries the narrative while on-screen text anchors the terms worth remembering. Verbatim full-script captions suit accessibility and muted feeds. Never show a wall of text identical to the narration while both compete — summarize visually instead.
How fast should explainer narration be?
Around 130 to 150 words per minute with deliberate pauses at transitions feels confident and clear. Faster delivery suits hype-driven spots but hurts comprehension of new concepts. Record with headroom for pauses, then tighten dead air in the edit rather than rushing the read.
Continue reading
Primary sources
Product capabilities change. These official references support the comparisons above.
Describe the edit. Keep the timeline.
Mellius turns conversational direction into cuts, captions, graphics, motion, color, and sound you can still inspect and change.
Start editing free