To create an online course: write the one outcome it delivers, work backwards to the shortest path there, sell it before you build it, teach that first cohort live, and only then record what worked. That order is the whole method, and it is close to the reverse of how most first courses get made. The seven steps below are that sequence in detail, with the two decisions people get wrong (length and price) taken from the outcome rather than from a guess.
We make Drry, a platform people host and sell courses on, so we are not a neutral party. Almost all of this guide is about what goes in the course, which is true wherever you put it. The one section about hosting says plainly where we are an option and where we are not.
Search this and you get two kinds of answer, neither of which is for you. One kind is instructional design written for universities and corporate training departments: learning objectives, assessment rubrics, accessibility standards. The other kind is a software company explaining that step one is to sign up for their builder.
Both skip the part that actually decides whether the course works. Making a course is not hard because recording is hard. It is hard because deciding what to leave out is hard, and nobody can do that for you with a template.
1. Write the outcome sentence before anything else
One sentence, in this shape: "By the end of this, you will be able to ___." It has to end in something the student can do, that they could not do before, that somebody else could verify.
- "Understand photography" fails. Nobody can tell whether it happened, including the student.
- "Shoot and edit a set of portraits good enough to charge for" passes. It names an artefact, and the student knows on the last day whether they have one.
If you cannot finish that sentence, you do not have a course yet. You have a topic, and a topic is what produces the eleven-hour version, because without an end condition there is no principled reason to stop adding.
2. Work backwards, and cut the syllabus
With the outcome fixed, the structure is not a creative decision any more. Ask what somebody has to be able to do immediately before the last step, then before that, until you hit what they already know walking in. That chain is your course. Everything not on it is a different course.
This is where the real work happens, and it is subtraction. The material you cut hurts to cut precisely because it is the material you find interesting, which is a fact about you rather than about the student. The history of the field, the debates, the edge cases, the thing you got wrong in 2019: all genuinely interesting, none of it on the chain.
3. Decide the length from the outcome, not the price
The instinct is to make it longer to justify the price. It backfires twice: the course takes months longer to build, and the buyer looks at twelve hours of video and quietly decides to start it next month. What they are buying is the result, and every extra hour between them and the result makes the purchase feel worse rather than better.
- Lessons: one idea each, and short. Five to fifteen minutes is where most good ones land. If a lesson is running to thirty, it is usually two lessons that have not been separated yet.
- Modules: a module should end somewhere the student can stop and have gained something whole. Three to six lessons is typical, and the boundary is the milestone rather than the count.
- The whole thing: short enough that a motivated person could finish it in a week. That is the constraint that keeps everything else honest.
Completion is not a vanity metric here. People who finish are the people who get the result, and people who get the result are the ones who tell somebody else and who buy the next thing. A course nobody finishes earns once.
4. Sell it before you build it
The most valuable thing in this guide, and the most commonly skipped. Take money for the course before any of it is recorded. Announce it to whoever you can already reach, with the outcome sentence and the module list, at a founding rate with the full price named next to it, starting on a date. Naming the full price is what keeps the discount a discount: what your first buyers anchor on should be the number you will settle on, not the launch one.
- You learn whether anybody wants it while changing it is still free.
- The people who pay tell you what is missing, and they are right far more often than your own outline.
- You are paid during the build instead of after it, which is what makes finishing it realistic.
- A start date is a deadline, and a deadline is the only reliable cure for a course that is perpetually ninety percent done.
If nobody buys, you have learned that for the cost of an email rather than a season of your life. That is the good outcome of the two, even though it does not feel like it.
5. Record, last, and stop shopping for gear
Recording comes last because it is the expensive step to redo, and everything above changes what you record. It is also the easiest step, which is why it attracts so much anxiety: buying a camera feels like progress in a way that cutting your favourite module does not.
- Audio is the one that matters. Viewers forgive a soft picture and leave over bad sound within seconds. A modest dedicated microphone is the only purchase that reliably pays for itself.
- Light beats camera. Face a window. A recent phone in good light looks better than an expensive camera in bad light.
- Script the first line and nothing else. Fully scripted lessons sound read; unplanned ones wander. Knowing your opening sentence and your closing point is usually enough.
- Do not edit for perfection. Cut the dead air and the false starts, then stop. Students are there for the result, not the production.
6. Price it against the outcome
Running time is not a price signal, and pricing by the hour is how good short courses end up cheap. What the buyer is weighing is the value of being able to do the thing, against what it would cost them to get there another way.
A course that gets a freelancer their first client is worth a multiple of one four times its length that teaches a hobby, and the market already prices it that way. Start higher than is comfortable: discounting later is easy and raising a price your first buyers anchored on is not. How to price a membership covers the recurring version of the same question, and the reasoning carries over.
7. Teach it live the first time
Run the first cohort as sessions rather than a finished library. It costs you calendar time and it buys the one thing no amount of planning provides: the list of places real students got stuck, which is never the list you predicted.
Those sessions are also the recording. Teach it live, answer the questions as they come, then cut the result into the lessons you already outlined. The finished course is the second version of something that has already worked on people, rather than the first version of something that has worked on nobody.
Where should you host an online course?
On whichever platform takes the least of what you earn and lets you keep the student list, because on the things people shop for (the player, the uploads, the drip) they barely differ. The honest summary of the market: course platforms differ far less in their players and their uploads than in what they charge and what they let you keep. Look at the transaction fee on the plan you would actually be on, whether the student email addresses are yours to export, and whether the course can sit next to the other things you sell. Kajabi against Teachable and Teachable against Thinkific walk through those trade-offs on the best known names.
Free plans are worth reading carefully rather than counting, and on the course platforms above there is usually nothing to read: Kajabi, Teachable and Thinkific all run a trial rather than a free plan. Where a free plan does exist it is normally on a per-sale platform, free to open and then taking a percentage of every sale, which is a price rather than an absence of one. Why platform fees are a tax on growth does the arithmetic.
Drry hosts the course beside a community and whatever else you sell, takes 0% of what you sell on every plan including the free one, and pays into your own Stripe account. The honest limitation is the one every hosted platform shares: it brings you no students. Everything in this guide about finding the first ones still applies.
Questions coaches ask
How long should an online course be?
As long as the outcome needs and not one lesson longer. Most first courses are two or three times the length they should be, because the maker keeps adding everything they know rather than everything the student needs. A course that gets somebody to a real result in ninety minutes is worth more than one that takes them twelve hours to abandon, and it is far easier to sell, because the buyer can see themselves finishing it.
Can I create an online course for free?
You can make one for free. Recording is a phone or a screen recorder, editing has free options, and writing costs nothing. Where free runs out is delivery: hosting video, taking payment, and keeping the course away from people who have not paid. Most of the big course platforms no longer have a free plan at all, only a trial. The ones that are free to open are usually free because they take a percentage of every sale instead, so read what the free option charges rather than whether it exists.
Do I need expensive equipment to record a course?
No, and equipment is the most popular way to avoid starting. A recent phone shoots better video than most students will ever notice. The one thing genuinely worth spending on is a microphone, because people forgive a soft picture and switch off bad audio within seconds. Good light matters more than a good camera, and a window is good light.
How much should I charge for an online course?
Price against the outcome rather than the running time, which is the mistake the hourly framing encourages. A course that gets a freelancer their first paying client is worth more than one four times its length that teaches a hobby, and buyers already know this. If you are guessing, start higher than feels comfortable: discounting later is easy, and raising a price your first buyers anchored on is not.
Should I build the whole course before I sell it?
No. Sell it first, to a small group, before any of it is recorded. You find out whether anybody wants it while it still costs you nothing to change, the first students tell you which parts are missing, and you get paid during the build rather than after it. Building in private for three months and launching to silence is the single most common way a first course dies. Price that first group lower, and name the full price next to it: they are buying it as it is made, with the gaps still in it, and their notes are shaping it. Naming the number you will settle on is what keeps the first buyers from anchoring on the founding one, which is the trap in the pricing question above.
What if nobody buys it?
That is information, and it arrives cheapest when you pre-sell. A course that nobody buys usually failed at the promise rather than at the content: it was aimed at a topic instead of at a person with a problem they are actively trying to solve. Before rebuilding anything, go and ask five people who did not buy what they thought it was for.