How to make an explainer video people watch to the end
Most explainer videos fail in the same place. Not the animation, not the voice: the first ten seconds, where the video introduces itself instead of saying something the viewer wanted to know.
This is a working method rather than a theory. It applies whether you are drawing the video yourself, briefing an animator or generating it.
Where do you start?
Before writing anything, finish this sentence: after watching, the viewer should be able to ____.
Not "understand our product". Something they can do or decide. "Tell whether this applies to their team." "Set the thing up without asking anyone." "Explain it to their manager in one line." If you cannot finish the sentence, the video does not have a job yet, and no amount of production will hide that.
What are the five beats?
Almost every explainer that works has the same shape underneath.
1. The situation, in their words. Open inside the problem, described the way the viewer would describe it. Fifteen seconds at most. No company name yet.
2. Why it matters. What the situation costs: time, mistakes, the thing that keeps not getting done. One or two lines. This is what buys attention for the middle.
3. The mechanism. How the thing actually works. This is the part people skip, and skipping it is why so many videos feel like adverts. Show the steps in order, one idea at a time.
4. Proof. Something concrete: a before and after, a real example, a number you can stand behind. Skip it entirely rather than inventing one.
5. One next step. A single action. Two options is one too many.
If a section does not fit one of the five, it is probably there because someone in the room asked for it. Cut it.
How long should the video be?
Narration runs at roughly 150 words a minute. So:
- 90 seconds is about 220 words
- 3 minutes is about 450 words
- 5 minutes is about 750 words
Pick the runtime first, write to the word count, and you will not have to butcher a finished script later. For a landing page, 60 to 90 seconds. For a lesson, 3 to 5 minutes. Past 6 minutes, ask whether it is really two videos.
How should the script sound?
Read the script aloud, all of it. If you run out of breath, the sentence is too long. If a phrase feels like something you would type but never say, rewrite it. Contractions are fine, and dry humor survives narration better than jokes do.
One idea per sentence, one point per scene. When a sentence carries three claims, whatever illustrates it will be a crowded frame that nobody parses.
What should the visuals do?
The test: watch with the sound off. If you can still follow the argument, the visuals are doing work. If you cannot, they are decoration, and the video is a podcast with pictures.
What earns its place on screen:
- Relationships: what connects to what, what feeds what
- Sequence: this step, then this one, then the failure mode if you skip one
- Contrast: before and after, with and without, expected versus actual
- Scale: how big, compared to something the viewer already knows
What does not: logos, stock smiling, decorative icons and text that duplicates the narration word for word.
Hand-drawn video suits all four because a drawing appears in the order the idea is built, which is exactly the pace of an explanation. That is why it works for teaching and for onboarding, where the viewer is meeting the structure for the first time.
Why keep one look across videos?
If you are making more than one video, the look is part of your identity. Decide once and apply it to everything: the same style, the same voice, the same opening rhythm. A series that looks like a series gets more attention per video than the same videos would get individually. There are 18 styles here, and the useful move is to choose one that matches the subject and then stop revisiting the decision.
What mistakes cost the most?
- Opening with your logo or your company history
- Explaining the category before explaining the thing
- Three calls to action
- Narration that reads out the on-screen text
- A script written to be read rather than heard
- Shipping the first cut without watching it once with the sound off
How do you produce it?
Once the script is written, the production step is mechanical. You can brief an animator, assemble it yourself in a canvas tool, or hand the script to a generator. With explain.ink you paste the script and get it drawn scene by scene with narration, or start from a topic and let it write the first draft to edit. If the material already exists as a file, upload the document.
Whichever route you take, the five beats and the sound-off test are what separate a video people finish from a video people close.