How to create an online course
Recording forty lessons nobody asked for is the standard way to spend three months. The sequence below finds out first.
The short version
- Sell it before you build it. A pre-sale to ten people is the only validation that counts.
- One outcome, not a subject. \"Get your first five clients\" beats \"Freelancing 101\".
- Short wins. Two hours that get finished beat twelve that do not.
- Completion is the product. Refunds and silence both come from courses nobody finished.
- Price above a product, below coaching. A course competes with free video on price alone.
The order
Pick one outcome, pre-sell it, then record the smallest version that delivers it. Almost everyone does this backwards (three months of recording, a launch, and silence) because building feels like progress and selling feels like exposure. A course is the digital product where that mistake is most expensive, since it is also the one that takes longest to make.
The rest of this is that sequence, and the reasons each step is where it is.
Step 1: one outcome, not a subject
A subject is "Instagram marketing". An outcome is "a month of content planned in one afternoon".
The difference decides everything downstream: who it is for, what goes in, what the price can be, and whether anyone finishes. A subject has no end, so the course has no end, so it becomes forty lessons.
Test it: can you say what the buyer can do on the last day that they could not do on the first? If not, you have a subject.
| Subject | Outcome |
|---|---|
| Freelancing | Your first five clients |
| Notion | A client CRM you actually use |
| Photography | Product photos that sell on Etsy |
| Copywriting | A sales page that converts |
This is the same narrowing rule as what digital products sell best, and it bites harder here because the build is so much longer.
Step 2: pre-sell it
The step people skip, and the only real validation there is.
Announce the course with a date and a price, and sell it before recording. Ten buyers is enough to proceed; two means the offer is wrong and you have lost a week instead of a quarter.
- Say exactly what they get and when.
- Discount the pre-sale honestly: they are taking a risk on you.
- Deliver live if you can, week by week. The recordings become the course.
- Refund gladly if you do not run it.
Delivering live first is the strongest version of this: you find out where people get stuck while you are teaching, and the edited recordings are a better course than anything you would have scripted alone.
If nobody buys, you have learned the thing that three months of recording would have told you later and more painfully. How to validate a SaaS idea before building it makes the same argument for software; the logic is identical.
Step 3: build the smallest version
Not a rough version. A short, complete one.
| Length | Completion | Verdict |
|---|---|---|
| Under 1 hour | High | Often enough |
| 1–3 hours | Good | The sweet spot |
| 3–6 hours | Falling | Needs real structure |
| 6+ hours | Poor | Usually a subject, not an outcome |
Lessons of five to ten minutes. One idea each, with a title that says what it is: people navigate a course by skimming titles, and "Module 3, Part 2" tells them nothing.
Include the thing they will actually use: the template, the checklist, the swipe file. For many buyers that is the product and the video is the explanation of it, which is why Notion templates and courses blur together, and why the template alone is often a better first product.
Equipment: a screen recorder and a decent microphone. Audio matters, video mostly does not, and most good courses are a screen with a voice over it. Using AI to create digital products covers where a model helps: outlining and scripting, not deciding what to teach.
Step 4: price it
A course sits above a template and below coaching, and it has an awkward competitor: free video.
You are not selling information, which is free. You are selling order, compression and the fact that someone decided what not to include. Price accordingly, and not at the bottom: a cheap course signals thin content and attracts the buyers most likely to refund.
How to price a digital product has the arithmetic. The course-specific note: the price has to survive the comparison a buyer makes with a free playlist, and only specificity wins that.
Step 5: where it lives
For a first course, this matters far less than it feels. A delivery page with the videos and the downloads is enough: a Notion page, or a simple storefront delivering a PDF of links.
Where to sell digital products covers checkout. Dedicated course platforms are worth paying for once you have proved the course sells and you need progress tracking, drip content or a community. Buying that first is the classic way to have a monthly bill and no students.
The thing that decides the second course
Completion. It is the metric nobody tracks and the one that determines whether this becomes a business.
People who finish leave testimonials, buy the next thing and tell someone. People who do not, quietly ask for a refund three weeks later and never return. Long courses have poor completion, which is the practical reason to keep it short rather than an aesthetic one.
What raises it:
- Short lessons with honest titles.
- A win in the first lesson, not an introduction.
- One thing to do after each lesson.
- Tell them how long it is up front.
- Email them at day three, which is where people stall.
Then the second course for the same buyer sells far more easily than the first for a new one, the same compounding as bundles and upsells, and where an affiliate programme starts being worth opening.
What goes wrong
- Recording before selling. The expensive version of every other mistake here.
- Teaching the subject you know rather than the outcome they want.
- Length as a proxy for value. It is a proxy for abandonment.
- A platform before a product.
- Launching once and stopping. A course is evergreen; the selling has to be too (how to launch a digital product).
Frequently asked questions
How do I create an online course from scratch?
Pick one outcome someone will pay to reach, pre-sell it to a small group before recording anything, then build the smallest version that gets them there. Recording first is how three months disappear into a course nobody wanted.
How long should an online course be?
As short as the outcome allows: typically one to three hours of video split into lessons of five to ten minutes. Length is not value, and long courses have worse completion, which drives refunds and kills word of mouth.
Where should I host an online course?
For a first course, a Notion page or a simple storefront delivery is enough. A dedicated platform is worth paying for once you have proved people buy, not before.
Do I need an audience to sell a course?
You need somewhere the offer will be seen, which can be a small email list, a community you are already in, or search. A large audience helps; what you cannot skip is having anyone at all to pre-sell to.
What equipment do I need to record a course?
A screen recorder and a decent microphone. Audio quality matters far more than video quality for teaching, and most courses are screen recordings rather than a person on camera.
The system behind this, written down
Everything above is the map. The Income Loop is the work inside it: modules 0–6 from the problem you solve to the offer that pays for it, plus ten traffic paths: the deeper post banks, the content sales systems and the full software build sequence, in one place.
Get The Income LoopNot ready to pay for anything? The Basic Income Loop is free and includes a complete seven-day starter for Threads, Instagram or software, enough to find out which one suits you before spending anything.