Why Are Course Sales Down? How to Tell AI Pressure From a Funnel Leak
Oct 13, 2026
By Barbara Amador, Course Conversion Strategist · Published
Some course creators are seeing sales fall in 2026, and AI is a real part of the reason. Buyers are more hesitant to pay for skills that feel uncertain, and a chatbot can now explain a topic for free. That pressure lands hardest on courses that mainly deliver information. But a market shift rarely explains a sales drop on its own. It pushes on the whole business and shows up first at the weakest layer. To know which problem you have, compare your funnel numbers stage by stage against your last launch.
Knowing AI is part of the story doesn't tell you what to fix. Knowing where it's hitting your business does. Here is how to tell the difference.
Is AI really hurting online course sales?
For some creators, yes, and there is a public, documented case.
In mid-2026, developer educator Josh Comeau shared that his newest course was on track to sell about a third of a typical launch. His two older courses were down too. He named AI as the biggest cause. He also said other course creators he had talked with were seeing revenue fall by half or more.
One creator's account isn't an industry study. But it's specific and public, and it matches what a lot of creators are quietly wondering about their own numbers.
What changed for online course buyers?
Two things changed at once. Buyers got more hesitant, and they got a free alternative that can answer back.
The first shift is hesitation. Comeau's buyers weren't sure their developer jobs would look the same in a few years. People think twice before paying to learn a skill they're not sure will last. Any field that feels unsettled by AI carries some of that doubt.
The second shift is the free alternative. A chatbot can explain a concept, answer follow-up questions, and adjust to the person asking. For free. At 2 a.m.
Courses have always competed with free content: books, YouTube, blog posts. But free content couldn't answer back. Now it can. If your course mainly delivers information, it's up against something that delivers information on demand, tailored to each person, at no cost.
That pressure is real. It isn't in your head.
Which courses does AI pressure hit hardest?
Courses that mainly deliver information feel it most. Courses built around a specific result, real feedback, and accountability feel it least, because those are the parts a chatbot can't fully provide.
This is a reasoned read of where the pressure lands, not a measured one. But the logic holds up when you put the two side by side.
| Course trait | More exposed | Less exposed |
|---|---|---|
| What it sells | Information on a topic | A specific result for a specific person |
| Who it's for | Anyone curious about the subject | A defined group with a shared problem |
| How people learn | Watch and read | Do the work, get feedback, adjust |
| Who answers questions | No one, or a forum | A person with real judgment |
| What keeps people going | Their own willpower | Structure, deadlines, accountability |
| What buyers compare it to | A free chatbot answer | Very little |
Look at the first two rows. "What it sells" is your offer. "Who it's for" is your niche. Those are the two layers at the top of a course business, and every layer below inherits from them. A broad niche and an information-only offer were weak spots before AI. Now they're the first places a free alternative can step in. How to choose a course niche that actually sells and how to build an offer people can't ignore cover both in depth.
None of this means information courses are finished. A clear, well-organized path through a topic still saves people time. But "I explain this well" is a harder promise to sell when the free option also explains things well. "I'll get you to this result, and I'll be there when you get stuck" is a different promise.
Notice what sits in the right column: judgment, feedback, someone who understands your specific situation. Those are the things people still pay a human for. That holds for your buyers, and it holds for you when you're trying to read your own numbers.
Is it AI, or is something in my funnel broken?
Often it's both, and the AI pressure tells you where to look. A market shift pushes on the whole business at once. A funnel leak shows up at one stage first.
Think of it like water behind a wall. The pressure is everywhere, but the water gets out at the weakest spot. AI raised the pressure. Your numbers show where the wall is weakest.
These patterns help you tell the two apart.
| What to check | Points to market pressure | Points to a funnel leak |
|---|---|---|
| Shape of the drop | Small dips at every stage | A sharp drop at one stage while others hold |
| Timing | Gradual, across your whole topic area | Started after a specific change: a new freebie, price, or page |
| Warm vs. cold buyers | Both groups slowed down | One group, or one step, fell off |
| What people say | "I can probably figure this out myself" or "Not the right time" | Confusion, mismatch, or "This isn't what I expected" |
Here is the hard part. When sales fall, the natural move is to react. Cut the price. Run ads. Rebuild the course around AI. Each of those is a guess about which layer broke, and each one costs time or money to find out.
It's also hard to read your own business clearly. You wrote the sales page. You picked the niche. You know what you meant, so you read what you meant. The symptom you notice first is rarely where the problem started. The symptom is rarely the problem goes deeper into why.
How to tell AI pressure from a funnel leak
You tell them apart by comparing your own funnel numbers across two periods and looking for the first stage that broke. This is the numbers side of finding the bottleneck in your course business.
- Pick two comparable windows. Your last launch and the one before it, or the same month this year and last year.
- Pull your stage rates. Opt-in rate, early engagement, click-through to the sales page, sales page conversion, checkout completion, and refunds. How to measure each stage of your course funnel has the formulas.
- Split warm from cold. Look at list visitors and new visitors separately. A drop in one group tells a different story than a drop in both.
- Look at the shape. Small dips everywhere point to the market. One sharp drop points to a leak.
- Read the words. Go through replies, questions, and refund reasons. Listen for doubt about the topic versus confusion about the offer.
- Find the first stage that broke. Read the stages in the order the business is built: niche, lead magnet, list, offer, sales page, launch, traffic. Stop at the first one that is clearly off.
- Get a second read before you change anything. This is the step creators tend to skip. Someone who didn't build the business can read the numbers without the story you already have attached to them.
Step six is the discipline. You'll likely see more than one number you don't like. The earliest one gets fixed first, because it may be the reason the later ones look weak. What should I fix first in my funnel? covers what to do when two problems are both real.
What not to do when course sales drop
The costly moves are the ones that change something before you know which layer is leaking.
- Cutting the price first. A lower price makes a course cheaper, not more necessary. Why course creators underprice covers what that costs.
- Buying traffic to make up the gap. More visitors into a leaking funnel means more people lost at the leak. Traffic is the last thing to fix for a reason.
- Rebuilding the whole course. A rebuild answers a question you haven't asked yet. The course may be fine and the offer around it may not be.
- Adding "AI" to everything. Putting AI in the title or the bonus stack doesn't fix a broad niche or a vague promise.
- Changing five things at once. If the freebie, emails, price, and page all change together, you won't know what worked.
- Deciding it's all the market. If it's all AI, nothing is yours to fix. That feels better. It usually costs more.
The short version
AI is putting real pressure on online course sales, especially for courses that mainly deliver information to a broad audience. But pressure isn't a diagnosis. It pushes on the whole business and shows up first at the weakest layer. Compare your funnel stage by stage against your own last launch. Small dips everywhere point to the market. One sharp drop points to a leak you can fix. Find the first stage that broke before you change your price, your traffic, or your course.
If you want a faster read on where your business is leaking, the free Course Business Diagnostic tests all seven layers in order and stops at the first one that isn't holding.
Start the free Course Business Diagnostic
If you'd rather talk through your numbers with someone who diagnoses course businesses for a living, you can reach me here.
FAQ
Is AI killing online courses? There is no evidence that online courses are ending, but some creators are reporting steep drops. Developer educator Josh Comeau said in 2026 that his newest course was tracking at about a third of a typical launch, and he named AI as the main reason. The pressure lands hardest on courses that mainly deliver information. Courses built around a specific result, feedback, and accountability are harder for a free chatbot to replace.
Why are my course sales down this year? It may be market pressure, a leak in your funnel, or both. Compare each stage of your funnel to your last launch or a comparable month. If every stage dipped a little, the market is likely part of it. If one stage dropped sharply while the others held, something in your business changed, and that is the part you can fix.
Should I still create an online course in 2026? Yes, if you validate it first. The risk is higher for a broad, information-only course, because buyers can now get explanations for free. Before you build, confirm that a specific group of people will pay for a specific result, and that your course delivers something a chatbot cannot, such as feedback, structure, or accountability.
Which online courses are most affected by AI? Courses that sell information on a topic to a broad audience are the most exposed. Their main value, explaining things clearly, is now available for free on demand. Courses that promise a specific result for a defined group, and include real feedback and accountability, are less exposed.
How do I know if it's AI or my funnel? Look at the shape of the drop. Market pressure tends to show up as small dips at every stage at once. A funnel leak tends to show up as a sharp drop at one stage, often after a specific change like a new freebie, price, or sales page. Read the stages in the order the business is built and stop at the first one that broke.
Should I lower my course price if sales are down? Not as a first move. A price cut treats a value problem as a cost problem. If buyers are hesitating because the course feels replaceable, a lower price makes it cheaper but not more necessary. Find the layer that is leaking first. If the offer itself needs work, fix what it promises before you touch the price.
Should I spend more on ads when course sales drop? Not until you know where the funnel is leaking. Paid traffic multiplies whatever your funnel already does. If one stage is losing people, more visitors means more people lost at that stage, at a higher cost. Fix the first broken layer, then add traffic.