Blog

Recording a YouTube tutorial on Mac: a real workflow

Recording a YouTube tutorial on Mac — workflow
A 9-step, 90-minute workflow for producing a 7–10 minute YouTube tutorial on Mac — outline through thumbnail, one screen recorder, no dedicated editor.

The workflow below produces a 7–10 minute tutorial in about 90 minutes, end to end. It assumes you've already chosen a topic and have something interesting to show. The stack is one screen recorder, one microphone, and one image editor for the thumbnail. That's it.

Step 1 — Outline (5 min)

One screen. Three columns: hook, walkthrough, payoff. Under each, three bullets. The whole outline fits on a Post-it. If it doesn't, your topic is too broad.

Hooks for tutorial videos are almost always one of three shapes: "Here's the thing most people get wrong", "Here's a faster way to do X", or "Here's the thing I had to learn the hard way". Pick one. Write the literal first sentence you'll say into the mic.

Step 2 — Script (15 min)

Expand the outline into a teleprompter-ready script. Short sentences. Contractions. One idea per line. Mark the words you want emphasised in bold — you'll stress them without thinking on the take.

Length: aim for 1100–1500 words for a 7-minute talking-head + walkthrough video. Read it aloud once. Cut anything that tastes wrong in your mouth.

Step 3 — Pre-shoot setup (5 min)

The boring checklist that saves a re-record:

  • Close Slack, Mail, Messages. Hide the Dock.
  • Enable Do Not Disturb. Disable notification banners.
  • Set the screen to a sensible scaled resolution. "More Space" looks unreadable on YouTube.
  • Clear the Desktop and empty the Trash. Both. People notice both.
  • Point a window at your face. The built-in webcam is fine if the light is right.

Step 4 — Talking-head intro (10 min)

Record the first 15–30 seconds as a webcam-only or webcam-prominent shot. This is the hook. Use the teleprompter. A built-in one keeps your eyes on camera.

Two takes is usually enough if the script is right. If you're past three takes, the script is the problem — go back to step 2.

Step 5 — Screen walkthrough (20 min)

One continuous take. Talk over every action — describe what you're about to do before you do it. This kills dead air without you having to cut around it.

If your recorder has automatic zoom on click, every action gets a cinematic push-in for free. More on why this matters for YouTube specifically here. If it doesn't, you're either keyframing in Final Cut after this step or accepting that phone viewers can't see your buttons.

Pause for 1.5–2 seconds between sections. Those pauses become your cut points in step 6 and they make the upper bound of your edit time predictable.

Step 6 — Edit (15 min)

If your recorder has a real editor inside it, this is mostly trim + delete. Three operations, in order:

  1. Cut the head and the tail.
  2. Cut the obvious flubs. The pauses you planned in step 5 make this fast.
  3. Audit the auto-zooms. Delete the ones on non-meaningful clicks. Add one on any non-click moment that deserves emphasis.

If you need a real editor (multi-cam, B-roll, music ducking), this step grows. For a single-take tutorial, it doesn't.

Step 7 — Captions + export (5 min)

Generate captions on-device. Scan for misspelt brand names — the transcriber sometimes guesses brand spellings. Fix in-line. Export the video, burnt-in, plus an .srt sidecar.

Upload the .srt to YouTube alongside the video. YouTube will use your captions instead of its auto-generated ones, and your captions become search-indexable. More on on-device captions here.

Step 8 — Thumbnail (10 min)

Don't skip this. A YouTube tutorial's click-through is largely thumbnail + title. The thumbnail rules of thumb:

  • Big readable text — 3–5 words max.
  • One face (yours) or one screen — not both.
  • Contrast — bright background or a big colour block.
  • The frame must read at 240 pixels wide. That's how it appears in the YouTube sidebar.

"The thumbnail must read at 240 pixels wide. That's how it appears in the YouTube sidebar — and that's the size most viewers decide your video by."

Figma, Sketch, or any image editor works. Save as PNG, 1280×720.

Step 9 — Title, description, tags (5 min)

Title: lead with the keyword, end with the hook. "How to record a product demo on Mac (without editing software)" beats "A guide to producing better product demos".

Description: the first sentence is what appears in search snippets. Make it count. Below that, paste the transcript — YouTube uses it for search relevance.

Tags: 5–10. Specific over generic. "screen recorder Mac" not "video".

The total

~90 min

total to produce a 7-minute YouTube tutorial — when the recorder is the editor and step 6 collapses to trim-only

5 + 15 + 5 + 10 + 20 + 15 + 5 + 10 + 5 = 90 minutes. For a 7-minute tutorial video. Faster than it sounds because step 6 collapses to almost nothing when the recorder is the editor.

If you ship one of these a week, the saved time compounds. The tool that makes that possible isn't more software; it's fewer pieces of software.

Frequently asked questions

How long should a YouTube tutorial be?
7–10 minutes is the effective range for a single-topic tutorial. Under 5 minutes and you're cutting too much to be useful. Over 12 minutes and you likely need chapters and a clearer hook. The script word count: 1100–1500 words maps to 7 minutes at 130–140 wpm speaking pace.
Should I include chapters in a 7-minute tutorial?
Only if there are 3 or more distinct sections a viewer would want to jump to. For a single-workflow tutorial that runs linearly, chapters add navigation weight without viewer benefit.
What size should a YouTube thumbnail be?
1280×720 (16:9), PNG or JPG, under 2MB. Critically, the text and key visual must still read at 240 px wide — test it by resizing your thumbnail to that width before uploading.
Do YouTube's auto-generated captions affect SEO?
Yes, negatively compared to uploading your own .srt. YouTube auto-captions are not indexed as reliably as uploaded caption tracks, and accuracy on technical vocabulary is worse. Always upload the .srt your recorder generated.