- Design every lesson as one of four types, concept, demo, exercise or recap, each with its own length and framing rule, instead of treating every video as a one-off.
- Lock the window size, zoom, microphone and lighting for the whole course, and record each module in one session, so the course looks like one production.
- Read the completion curve lesson by lesson and diagnose the drop with a fixed set of causes: length, density, pace or an unclear goal.
Online course video rarely fails because the idea was weak. They fail at lesson fourteen, where the framing shifts, the window is a different size, the audio gets thinner and the instructor is suddenly typing faster than the student can follow. The student doesn’t articulate any of this. They just stop, and a few weeks later they’re one of the people who bought the course and never finished it. Course video is a consistency problem before it’s a quality problem.
This playbook is about running a course as a system. It covers the four kinds of lesson that cover almost everything a course needs, how to keep the frame identical across dozens of videos, where students most often quit, and how to read the completion data so you fix the right lesson for the right reason.
Think in lesson types, not videos
Most course creators plan their material as a list of topics and then record each topic as a video, with a length and style decided on the day. That produces forty different videos that feel like forty different people made them. A better approach is to decide, in advance, what kinds of lessons the course needs, and then to design each kind once.
Almost every lesson falls into one of four types.
A concept lesson explains an idea that the student has to understand before they can do anything with it. It’s often talk over a short slide or a simple screen. The job is to build a mental model, so the pacing is slow, the examples are concrete and there’s room for a pause.
A demo lesson shows the instructor doing something on screen. The student watches a workflow from start to finish and learns the sequence. The framing is tight, the zoom follows the cursor, and the narration describes what the student is seeing at each step.
An exercise lesson asks the student to do the thing themselves. The instructor sets up the task, then steps back. These lessons work best when they’re short and end with a visible checkpoint, so the student knows whether they succeeded before moving on.
A recap lesson pulls the module together. It’s the shortest type, usually under two minutes, and it names the two or three things the student should be able to do now.
Here’s how the four types differ in practice. The lengths are targets for a typical software or skills course, and you should adjust them to your subject.
| Lesson type | Target length | Framing | What the student should leave with |
|---|---|---|---|
| Concept | 4 to 8 minutes | Slide or simple screen, instructor visible if you use a camera | A mental model they can explain in their own words |
| Demo | 5 to 10 minutes | Tight zoom on the active control, one workflow | A sequence they could repeat with prompting |
| Exercise | 2 to 4 minutes, then a pause | Instructor sets the task, then the screen goes to the student’s side | A finished result and a way to check it |
| Recap | 60 to 120 seconds | Summary card over the final result | Three named skills and a link to the next lesson |
The point of the table isn’t the exact numbers. It’s that every lesson of a given type behaves the same, so the student learns what to expect. A student who knows that an exercise is short and ends with a checkpoint will attempt it. A student who meets an unexpected twelve-minute exercise will start skipping.
Keep every online course video in the same frame
Students track layout far more than they realize. When the same product always fills the same region of the frame, the eye goes straight to the active control without searching. When the window size or zoom changes between lessons, the viewer has to re-learn where things are, and that re-learning eats attention they were using to understand the material.
Lock the variables before you record the first lesson. Decide the window size, the browser zoom, the font size in the product, the microphone and its distance, the room lighting if you’re on camera, and the demo data. Write them on a single sheet and treat that sheet as a contract. Anyone recording for the course, including you on a bad day, works from the sheet.
This matters even more for courses recorded over months. Software updates, your own equipment changes, and a new room can each shift the look of the course in ways that are obvious to a viewer who watches three lessons in a row. If you must change a setting partway through, do it at a module boundary and say so in the module’s opening lesson, so the change feels deliberate.
Don't change the window or zoom mid-module: a shift in size or zoom between two lessons creates a visible jump that viewers feel even when they can't name it. Pick the settings before the first recording and hold them until the course is finished.
Record each module in one session
Batch recording is the other half of consistency. Recording the lessons of one module in a single session keeps the software state, your voice, the room and your energy level the same across the whole set. Recording one lesson on Monday and the next on Thursday invites small differences in all of those, and they accumulate.
A workable order for one module looks like this:
- Write the scripts for every lesson in the module first. Each script should name the lesson type, the target length and the one thing the student should leave with. Having all of them in hand before you record prevents the common problem of improvising the second lesson while still recovering from the first.
- Set the software state once. Use a reset routine, the same way you would for a software tutorial, so the demo data is the same for every take in the module.
- Record the demo lessons first. They’re the easiest to get right, and they establish the rhythm the other types should match.
- Record the concept lessons next, then the exercises. Exercises depend on the workflow the demos showed, so it helps to have the demo footage in mind when you set them up.
- Record the recaps last. They summarize what the module just covered, so you’ll know exactly what to name.
- Log the session. Note the date, the software version, the zoom and any problems. A short log is the difference between a course you can update in a year and one you have to re-shoot.
Name every file with a fixed pattern, such as module number, lesson number and type, for example 03-02-demo-filter-records. The pattern lets you find, replace and re-order lessons without opening them, and it makes the course index easy to generate.
Put the exercise pause where students can finish it
The exercise lesson is where courses quietly lose students, and the failure is almost always structural. A demo shows a workflow of six steps. The exercise asks the student to reproduce it. If the exercise gives no clear checkpoint, the student has to re-watch the demo to recover the steps, or they guess, get a different result and decide the course is broken.
The fix is to design the handoff. An exercise lesson should open with the goal stated in one sentence, show the finished result for three seconds so the student knows what they’re aiming for, and then leave the screen to the student for a timed pause. At the end, show the checkpoint: what the correct state looks like, and how to check it. A student who can compare their result to a visible target will keep going.
End every exercise with a visible checkpoint: a screen that shows the expected result, plus one line on how to tell whether they got it. A student who knows they're on track is far less likely to quit the course at that point.
Keep the exercise short enough that a student can finish it in one sitting, even on a busy day. Three minutes of practice that a student completes is worth more than ten minutes they abandon. If the task really needs more time, split it into two exercises with a checkpoint between them.
Read the completion curve one lesson at a time
Once a course is live, the analytics show you something no pre-launch review can: where students stop. Most platforms report completion or watch-through per lesson. Read that curve from the start of the course to the end. A course usually loses students gradually, but the sharpest drops point to specific lessons, and those are the ones worth fixing first.
When you find a sharp drop, diagnose it with four questions, in order. Is the lesson too long for what it delivers? A twelve-minute demo that could be eight minutes will lose people around minute nine regardless of quality. Is the lesson too dense? A concept lesson that introduces three new ideas in one video overloads the viewer, and they’ll stop to recover. Is the pace wrong for the type? A demo that narrates every click at the speed of a concept lesson feels slow, and a concept lesson that rushes through the model feels like noise. Is the goal unclear? If a lesson’s opening doesn’t say what the viewer will be able to do afterward, many of them will decide it isn’t for them.
Fix one lesson at a time, and give the fix a few weeks before you judge it. Changing several lessons at once makes it impossible to tell which change helped. Replacing a lesson also means re-recording it with the same frame and setup as the rest of its module, so the fix doesn’t introduce the jump you were trying to avoid.
Rewatch data is worth reading too. A lesson that students replay several times is often one where the instructor moved too quickly through a step or where the on-screen result was hard to see. Replays aren’t a failure in themselves, but a lesson with a high replay rate and a low completion rate is usually asking the student to do something the video doesn’t make clear. Check whether the replayed moment is a demo, an exercise checkpoint or a concept, and handle each the way the lesson-type table describes.
Keep the course from drifting
Courses age. The software changes, a recommended tool is replaced, and an example that was current when you recorded it becomes wrong. A course that isn’t maintained slowly stops matching the product its students are trying to use.
Build the maintenance into the course’s structure from the start. Keep the lesson index as a single list with lesson types, lengths and the screens each lesson depends on. When a product update touches one of those screens, the matching lessons go on a list to re-shoot. Because each lesson is a self-contained file with a fixed name, you can replace one without disturbing the rest, the same way you would with a chaptered software tutorial.
Version the course as a whole, even if students only see the latest version. A course that says “updated for version 4.2” on its opening lesson tells students that the material is current, and it gives you a clear point to say what changed in the next release.
Keep the production honest
Course video has one more trap that’s easy to miss: the temptation to over-produce. A course that looks like a film production in the first module and a webinar in the last one will feel uneven, and the uneven part is what the viewer notices. Set a production level you can hold for the whole course, and then hold it. A consistent, clean, modestly produced course outperforms a stunning first module followed by thirty tired ones.
The same logic applies to length. Students who finish courses tend to finish them because the lessons are short enough to fit into their days, not because the course is comprehensive. A course with forty focused lessons is easier to finish than one with ten sprawling ones, even if they cover the same material.
The short version
An online course holds together when its lessons behave like a system. Design four lesson types with fixed lengths and framing, lock the window, zoom and audio for the whole course, record each module in one session, and end every exercise with a visible checkpoint. Read the completion curve lesson by lesson and diagnose each drop with the same four questions. Keep a lesson index that makes replacing one lesson cheap.
For the recording itself, RecordlyPro records in your browser with nothing to install and adds auto-zoom on every click, cursor effects, captions and chapters, which helps each demo lesson look finished without an editing session for every take. If your course includes long walkthroughs, the same chaptering approach from how to record a software tutorial that stays accurate keeps them easy to repair. And if you’re deciding how long each lesson should be, how to record a tutorial people finish covers the pacing and length trade-offs in more detail.
Keep reading
Async video: stop scheduling the meeting you could record
A hundred random links can do less than five relevant ones. Here's what makes a backlink pass real ranking value in 2026 — and what makes it worthless or worse.
Sofia Meyer
October 6, 20262 min readAuto-zoom in screen recordings: what it is and why your demos need it
A hundred random links can do less than five relevant ones. Here's what makes a backlink pass real ranking value in 2026 — and what makes it worthless or worse.
Sofia Meyer
October 13, 20262 min readHow to choose a screen recorder for product demos in 2026
A hundred random links can do less than five relevant ones. Here's what makes a backlink pass real ranking value in 2026 — and what makes it worthless or worse.
Sofia Meyer
October 16, 20262 min readDofollow vs nofollow in 2026: what still passes value
The link attributes changed, Google's treatment changed, and a lot of advice is out of date. Here's what dofollow, nofollow, sponsored and ugc actually do now.
The Recordly Team
October 10, 20262 min read