How To Launch An Online Course With A Pilot
In March 2020 we shut down our education business.
Some background. In 2018 my co-founder and I started an IT company. Alongside it, we ran an educational project that taught people how to build and promote online schools. We used other learning platforms to run it, kept running into their limits, and at some point realized we knew exactly what we wanted from an LMS. In 2019 we launched our own.
For about two years we ran both. Then the pandemic arrived, demand for online learning jumped, and we made the call: close the education project, let part of the team go, put everything into the platform.
I still think it was the right decision. What I want to talk about is how we got to it, because the way we got there is the same mistake I now see in a lot of course launches.
We Deliberated Instead Of Testing
For those two years we never really tested which of our two businesses was the business. The education project brought in revenue. The platform didn't yet. By the evidence we had, the education project was the business. We kept both alive because finding out felt risky.
So when we finally decided, we decided under pressure, all at once, with people's jobs attached. A small, deliberate test a year and a half earlier would have given us the same answer at a fraction of the cost.
That was one half of the pattern. We treated big decisions as things to think about until they became obvious. It felt responsible. Mostly it was a way to avoid being wrong in public.
The other half was the opposite. When we did believe in something, we went in at 100%. Full team, full budget, full commitment, on the theory that half-hearted execution gets half-hearted results. The theory isn't foolish. It's just expensive, because it turns every uncertain bet into a full-size one. And most bets are uncertain.
Slow to test, then all-in at once. A typical full-size course launch has exactly the same shape.
The Full-Size Course Launch Is The Same Bet
Here is how it usually goes. An expert spends months building a complete course: every module recorded, every worksheet designed, a landing page, maybe a launch webinar. Then they open sales and find out, for the first time, whether anyone wants it, at what price, and whether the people who buy can actually get through it.
That's the 100% bet. All the cost goes in before any information comes out.
The alternative is to find the smallest version of the course that produces real answers, run it, and let the results make the decision.
What A Pilot Cohort Looks Like
A pilot is a small, real, paid run of the course before the full build. A few things make it work.
It's small enough to count. A group you can name individually. The point is to watch each learner, not to hit a revenue target.
It's live, or close to it. Deliver the first version in live sessions or with close contact: calls, comments, quick replies. You'll learn more from one confused question in week two than from any survey at the end.
It's paid. Free pilots attract people who are curious. Paid pilots attract people with the actual problem. Price it lower than the planned full course and tell participants openly why: they are getting a discount in exchange for being early and giving feedback. That trade buys you honesty.
It's built only as far as needed. Outline the whole path, but produce the lessons just ahead of the group. If module four changes because of what happened in module two, you haven't wasted anything.
What The Pilot Should Tell You
Go in with specific questions, written down before you start. Otherwise you'll read the results the way you want to.
Will people pay? Not "would you buy this?" in a poll. An actual payment. How hard was it to fill the group, and who said no and why?
At what price? A pilot is a good moment to try more than one price, or more than one package. If you haven't thought about how to set this up, here is a practical piece on testing more than one price without extra tools.
Where do people get stuck? Watch which lesson produces the most questions, the longest silences, the most "I'll catch up later." That's your redesign list for the full version.
Do they get the result? Not the certificate. The outcome the course promises to produce. If pilot learners don't get there with you in the room, recorded lessons won't get them there either.
What do they ask for that you didn't plan? Some of the best modules come from questions nobody in the pilot expected to ask.
When To Stop, Change, Or Scale
Decide your thresholds before the pilot starts. What would make you stop? What would make you change the format? What would make you build the full version?
This is the part we got wrong for years. Without a line drawn in advance, every result looks "promising enough to keep going," and you drift. With a line, the pilot actually makes the decision for you. That is uncomfortable. It's also the entire point.
A pilot that tells you to stop is the cheapest bad news you will ever get. Compare it with the alternative: eight months of production, a launch, and the same answer at many times the cost.
For L&D Teams
The same logic works inside companies. Before rolling a new program out to the whole organization, run it with one team. Pick the questions in advance: do people show up without being chased, where do they stall, does the behavior you are training for actually change on the job? Rolling out at 100% and discovering the problems in the completion report is the corporate version of the full-size launch.
What I Would Do Differently
If I could go back, I wouldn't change the decision to go all in on the platform. I would change how long we waited to test it, and how much we put at stake each time before we knew.
Small tests, clear thresholds, and the discipline to act on what they tell you. It works for a company. It works even better for a course, because a course is much cheaper to test.