How many steps a tutorial video should have

A tutorial video should have four or five steps. That is what every one of the five finished how to videos on this site uses, counted on 13 September 2026: 23 steps across five videos, with each step on screen for a median of seven seconds and narrated in a median of ten words. Add an intro and an outro and the whole video runs 34 to 52 seconds.

Five videos, a small but exact set

This is a small corpus and the post says so plainly. The site holds 20 example videos, and five of them are tutorials: a password reset, an expense claim, moving a deal through a CRM pipeline, a first day in an HR portal and a dashboard tour built around one question. Each is stored with its script, an ordered list of scenes with a role, a narration line and a duration, so the counts are exact even if the sample is not large.

A step here means a scene with the tutorial role, the scene that shows one action on one screen. The other scenes, an intro, an outro, and in three videos an overview or a checklist, are counted in the length but not as steps. All five are in English, landscape at 16:9, read by an instructional voice. Every business in them is fictional and every screen is a hand built mock.

Two things limit the reading. Five videos cannot show a trend, only a pattern, and several of the timing rules come from the video pipeline that built them: tutorial steps are planned at 5 to 9 seconds, intros and wrap ups at 3 to 5, and narration at 10 to 25 words a step. What the count adds is what finished, edited examples actually settle on inside those rules. The set will grow and the numbers will be recounted.

Steps by kind of task

The table below this article gives the steps, timing and narration for each video. The step count barely moves: three videos have five steps and two have four.

What does move is what surrounds the steps, and that depends on the task. The password reset video is the leanest: four steps, an intro and an outro, 34 seconds, and nothing else, because a reset has no structure to explain and nothing to attach. The expense claim video adds a checklist after its five steps, listing the receipt, the cost centre and the reason a claim needs, because that is where claims actually bounce.

The CRM pipeline video goes the other way and puts its extra scene first: a drawn overview of the three pipeline stages before any screen appears, so every step that follows has somewhere to land. The HR portal video adds two scenes after its steps, a checklist of what to finish by Friday and a three card scene on who to ask. The dashboard tour replaces the overview with a question in its intro, why did sign ups drop last week, and makes every step part of the answer.

The rule the set suggests is simple. Keep the steps to four or five, and add no more than two scenes around them, to orient the viewer before them or to consolidate what they did after them.

How long each step stays on screen

Across 23 steps the median is seven seconds, and every step falls between five and nine. Sixteen steps are exactly seven seconds. That uniformity helps: a viewer following along settles into the rhythm and knows when to look for the next action.

The exceptions are informative. The longest step in the set, nine seconds, is the CRM step that teaches two facts at once, where the attached quote appears and the paperclip that marks the deal card. The eight second steps carry either a longer instruction or a close up that needs reading: the password reset spends eight seconds on the email step, and the HR portal gives eight to the step about acknowledging three policies. The five second steps are the simple closing actions, such as setting an alert at the end of the dashboard tour.

Intros and outros both run three to five seconds in all five videos. A tutorial that spends fifteen seconds introducing itself has already lost the viewer who came for step one.

When the camera zooms in

Every step in these videos is a still screenshot with one marked click point, and the scene zooms toward it. The video builder offers three zoom levels for a marked point: no zoom, zoom in at 1.5x, and a close up at 2.1x. The example pages describe how each video uses them.

The standard is 1.5x. The CRM video uses it on every step, and the expense claim and dashboard tour use it for most steps. The 2.1x close up is used sparingly, and on the point where the task is most likely to fail. The password reset saves it for the email step, where the message lands in a crowded inbox, the button is small and the link expires. The dashboard tour saves it for the campaign row where the answer to its question sits. The expense claim keeps it for the moment a status badge needs reading.

The dashboard tour shows the approach inside a product demo. Its five steps run seven seconds each except the last, and each line pairs an action with what the screen reveals: the overview shows sign ups down twelve percent, filter to last week, break it down by channel and see that paid is the one that fell. The close up waits for the fourth screen, the paused campaign that explains the drop.

The pattern is worth copying: zoom a little on every step so the eye finds the control, and zoom hard only once, on the step that decides whether the task succeeds.

How many words each step needs

The 23 steps carry 225 words over 160 seconds, 1.4 words a second, which is slower than conversation. The median step is 10 words and the longest is 15: attach the quote PDF; it shows under Files, and the deal card gets a paperclip.

Eleven of the 23 steps use fewer than 10 words, below the pipeline’s planning floor, and none comes close to its 25 word ceiling. Short lines work because the screen carries half the meaning. Open the deal from the board is six words, and with the marked click on the board it needs no more.

The longer lines add a reason, not more instructions. The HR portal’s bank details step says where to add them and that they are checked before the first pay run, which is the reason to do it today. One action plus one reason is the shape of a good step line.

A step plan for your own video

Write the task as a list of actions, one per screen, and stop at five. If the list runs longer, split the task. Give each step seven seconds and about ten words, name the control, and add a reason only when skipping the step has a cost. Put a short intro before the steps that names the task or asks the question it answers, and one scene after them, a checklist or an outro, that tells the viewer they are done.

How to use this

Capture or upload one screenshot per step, mark the control, and write the line for each before building. The AI tutorial video maker page explains why a tutorial holds one layout per step and notes that tutorial videos are a paid feature. The guide to making a how to video with AI walks through the build. For writing the narration itself, read how to write a tutorial video script, and for making videos without a camera or a presenter, see what a faceless video is.