The anatomy of a tutorial people finish
A tutorial has one job: the viewer can do the task without rewinding. PageToVid's tutorial goal writes to the shape instructional designers use.
- The goal, in one sentence. "By the end you will have a shared dashboard."
- A prerequisite, only if it cannot be avoided. One line, not a lecture.
- One scene per step. Where the control is, what to do, what you see afterwards. Each step is numbered in the captions.
- A check. How to tell that it worked.
- The next task. Where someone who finished this one goes next.
A timeline can recap the steps at the end; it never replaces them. Left to itself, the tutorial goal writes 90 to 180 seconds — long enough to show every click — and you can ask for any length from 15 to 180 seconds.
Shots that perform the step
A recording that only scrolls teaches nothing. Each recorded scene can perform one action in a real browser before its frame is captured:
- Click a button, a tab or a menu item, and show what opens.
- Type a real query into a search or filter field.
- Hover to reveal a menu or a tooltip.
- Spotlight one control while the rest of the screen dims.
The narration says what the cursor does, in the present tense, and slows slightly on each step. The camera never types names, emails or passwords into a page, so a step that needs personal data is shown on a demo account.
Filming inside your product
Your marketing site is public; the screens a tutorial needs usually are not. A capture session, registered once from the VS Code or browser extension, lets the camera film behind your login. It is bound to one site, expires within three days, and is a Scale-plan capability. Register it with a demo workspace full of sample data, and the tutorial shows a realistic product without exposing a customer.
For a product with no public page at all, a tutorial can also be built from a storyboard you write, with animated scenes for each step — but a recording of the real interface is what lets a viewer follow along.
A library that stays true
Tutorial libraries rot one release at a time. Because each film is generated from the live product, rendering it again captures today's interface and keeps the narration, the structure and the timing: 40 credits instead of a recording session. When a step itself changes, edit that scene's line and action in the storyboard, then render again.
Read what is a tutorial video for the format, how to make a tutorial video for the method, and automated vs manual screencasts for when a recording made by hand is still the better choice. Writing help-center answers rather than product training? See help-center videos.