Home / Course Platforms

How to Sell an Online Course: The Complete Setup Guide

By Akira Endo · Published 2026-08-27

Most guides to selling an online course get the order wrong. They start with the platform, because the platform is the thing with an affiliate program attached. But the platform is roughly the fifth decision you make, and picking it first is how people end up paying $179 a month to host a course that nobody validated and three people bought.

This guide runs the steps in the order they actually matter. Some of it is unglamorous — the money questions in particular arrive later than you’d like and hit harder than you’d expect.

Some links in this article are affiliate links — see our affiliate disclosure. It doesn’t change what we recommend.

What we’ve actually run. While building our platform comparisons we opened trials on Kajabi, Thinkific and LearnWorlds, built a course skeleton inside each, and read the checkouts before cancelling. We also run systeme.io’s free plan for our own funnel. Where a figure below comes from one of those accounts, we say so. Where it comes from a vendor’s published pricing, we say that instead — those were read off official pricing pages on 21 August 2026, and prices move, so confirm before you commit. Teachable and Podia we have not run, and we don’t pretend otherwise.

The order of operations

1. Validate that someone wants this
2. Outline the curriculum
3. Record the thing
4. Choose where it lives
5. Price it
6. Build the checkout
7. Launch to whoever you already have
8. Handle the costs that only appear after launch

Steps 1 to 3 cost nothing but time. Step 4 is where the subscriptions start. Reversing that order is the single most common and most expensive mistake in this category, because a course platform is genuinely hard to leave once students are enrolled — you’re migrating content, moving student accounts and breaking links, usually mid-launch.

Step 1: Validate before you build

The cheapest version of validation is to sell the outline before the course exists. Write the curriculum as a sales page, put a price on it, and see whether anyone pre-orders. If nobody does at $99, they will not appear at $299 once you’ve spent six weeks recording.

Three signals worth more than enthusiasm:

• Someone has already paid for adjacent help. People who have bought a book, a template, or a coaching hour in your topic are proven buyers. People who’ve only read free content are not.
• The problem recurs. One-time problems get solved by a YouTube video. Recurring problems justify a structured course.
• You can name the transformation in one sentence. “Go from no email list to your first 100 subscribers” sells. “Learn about email marketing” does not.

If you can’t get a pre-order, that’s information, not failure. It’s much cheaper to learn it here than after a year of platform fees.

Step 2: Outline the curriculum

This is the step where the platform AI tools are genuinely useful, and it’s worth being specific about what they do and don’t produce, because the marketing implies more than you get.

We ran LearnWorlds’ AI course builder from a single sentence of description, and it returned a full structure — sections with activities mapped underneath, ready to edit. Kajabi has a comparable AI outline feature and we ran that too. Our honest read after using both: they’re good at structure and useless at substance. You get a sensible skeleton — a logical progression, sane module boundaries, the obvious topics in the obvious order — in about a minute, which genuinely saves an afternoon of staring at a blank page.

What they can’t do is know what you know. The generated outline is the median of everything written about your topic. Every specific example, every number from your own experience, every “here’s where this goes wrong in practice” — that’s the part that makes someone finish the course and recommend it, and none of it comes out of a generator.

Use the AI for the scaffold, then throw out the sections you can’t teach better than a blog post. A tight six-module course you can teach with authority beats a twelve-module one padded to look substantial.

We go deeper on what LearnWorlds specifically generates in our LearnWorlds review.

Step 3: Record without a studio

You need clear audio and readable screens. That’s the whole bar for a first course.

• Audio matters more than video. Viewers forgive a webcam and abandon bad sound. A basic USB microphone in a room with soft furnishings beats an expensive mic in an echoey one.
• Screen recordings carry most course types. If you’re teaching software, a spreadsheet, or a process, your face is optional and often a distraction.
• Record in short takes. A five-minute lesson re-recorded is cheap. A forty-minute lesson re-recorded is a lost evening, so you’ll leave the mistake in.
• Don’t script word for word unless you’re comfortable reading aloud; most people sound stilted. Bullet points per lesson, then talk.

Batch it. Recording four lessons in one sitting takes far less than four separate sittings, because most of the cost is setup and warm-up.

Step 4: Choose where it lives

Now the platform. Three structural questions decide it, and feature lists mostly don’t.

Do you need the audience tools bundled? Some platforms include email, funnels and checkout; others are a course player you attach to tools you already have. Bundled costs more per month and less in total if you’d otherwise pay for three subscriptions. Unbundled is cheaper and more work.

Which limit will you hit first? This is the question nobody asks and everybody regrets. Each platform meters something different, and the entry plans bite in different places — contacts on one, video bandwidth on another, active students on a third, email subscribers on a fourth. The right question isn’t “which plan is cheapest” but “which ceiling do I hit at 200 students, and what does the next tier cost.”

How badly does the fee structure punish early revenue? Transaction fees on entry plans ranged from 0% to 7.5% across the five platforms we compared. At low volume a percentage fee is trivial; the moment you have a launch, it isn’t.

One finding that surprised us enough to flag: none of the five major platforms currently offers a permanently free plan. Several used to. Every one is trial-only now, so budget for a paid subscription from month one rather than planning around a free tier that no longer exists. If free is a hard requirement, we cover what genuinely remains in free course platform options.

The full breakdown of entry prices, trial lengths, fees and first-ceiling limits is in our course platform comparison.

Step 5: Price it

Price on the outcome, not on the runtime. Nobody buys four hours of video; they buy the result at the end of it.

A workable starting frame:

• Under $50 is hard to sustain. You need volume you probably don’t have yet, and cheap courses attract the least committed students, who ask the most support questions.
• $99–$299 is where most first courses from solo creators land, and where a small audience can produce meaningful revenue.
• Above $500 generally needs either a strong existing reputation or a live component — cohort calls, feedback, community — to justify the gap.

Run the arithmetic before you commit, because platform costs are fixed and your sales aren’t. At a $179/month plan, a $99 course needs roughly two sales a month just to cover hosting before any fee. At a $29/month plan it needs one. That difference is irrelevant at 50 sales a month and decisive at three.

Annual billing is usually 20–25% cheaper than monthly across these platforms, but it commits you before you know whether the course sells. Pay monthly through your first launch, then switch once there’s revenue to justify it.

Step 6: Build the checkout

We read the checkout flow on both Kajabi and Thinkific during our trials, and it’s worth doing on whatever you pick, because this is where a surprising amount of revenue leaks.

Things to verify with your own eyes before launch day:

• Does it take the payment methods your buyers use? Card is table stakes; PayPal is not universal, and regional methods vary.
• What does the buyer see after paying? A confirmation that doesn’t clearly say “here’s how you get in” generates support email on day one.
• Is there an order bump or upsell slot, and do you want one? These meaningfully raise average order value, and they’re easier to set up before launch than retrofit.
• What does the refund process take? Find out before someone asks.

One practical note from our own trial signups: read what the trial actually enrols you in. Kajabi’s checkout offered us a 14-day trial where 30 days was advertised, and Thinkific’s 30-day trial runs on the $109 Start plan rather than the $54 Basic plan you may be intending to buy. Neither is hidden, exactly, but neither is on the pricing page either. Both required a card up front.

Step 7: Launch to whoever you already have

You do not need an audience to launch. You need a list of people who have the problem — and that can be twenty people.

The sequence that works with a small list:

1. Tell the people closest to the problem first, individually, before any broadcast. These are your first sales and your first testimonials.
2. Open with a founding-member price and be honest about why: the course is new, you want feedback, the price goes up after.
3. Email three times, not once. Announcement, then a mid-window email answering the objection you’ve heard most, then a closing email. Most sales land on the third.
4. Collect testimonials immediately from the first cohort, while they’re still enthusiastic. You’ll need them for launch two.

If you’re building the list at the same time as the course, the lead magnet should be a piece of the course itself — the outline, one lesson, a checklist. It self-selects for buyers.

Step 8: The costs that appear after launch

Three that reliably surprise people:

Payment processing on top of platform fees. The platform’s transaction fee and Stripe/PayPal’s processing fee are separate charges. A “0% transaction fee” plan still costs you roughly 2.9% + $0.30 per sale to the processor. Vendors describing themselves as fee-free are almost always talking about their own cut only.

Usage tiers you cross quietly. Email contacts, video bandwidth and active-student counts all increment without an alert until you’re over. The jump to the next tier is rarely small.

Refunds and chargebacks. Budget a few percent. Processing fees on a refunded sale are often not returned.

Taxes on digital goods. Depending on where your buyers are, VAT or sales tax may apply, and platforms differ in whether they handle it for you or leave it entirely to you. Check this before your first international sale rather than after.

A realistic first 90 days

The compression most people attempt — platform in week one, recording in week ten — is what produces an expensive, unsold course. Delaying the subscription until week five costs nothing and is the single easiest saving available.

What to do next

If you’re at the platform-choice step, the comparison of entry prices, real fees and first ceilings is in our course platform comparison. If you’re weighing the two most commonly shortlisted options against each other, we ran both and wrote it up in Kajabi vs Thinkific.

And if you’ve read this far without a validated topic, go back to step one. It’s the only step that can’t be fixed later.