Are online courses still worth selling in 2027?

Why are people questioning whether online courses are still worth selling?

Online courses are still worth selling in 2027, but selling a generic prerecorded course has become much harder than the size of the online-learning market suggests.

The confusion comes from two trends happening at the same time.

Paid online learning is still a very large business. Coursera completed its acquisition of Udemy in 2026, creating a combined platform with 1.66 million paid consumer subscribers in its latest reported quarter, up 44% from the comparable figure a year earlier. The combined company generated $158.6 million of consumer revenue during that quarter.

Independent course businesses are still processing serious money too. Thinkific reported $71.5 million of payments processed through Thinkific Commerce in its latest quarter, 10% more than a year earlier.

But the weakest part of the market is increasingly easy to identify.

Before the acquisition, Udemy's consumer revenue fell 9% in 2025, largely because fewer people were buying individual courses. Thinkific's latest numbers show another version of the same divide: revenue from its self-serve business slipped 1%, while its larger-customer Thinkific Plus business grew 14%.

Meanwhile, course supply keeps growing. Class Central counted around 272,000 Udemy courses at the end of 2025, with half attracting fewer than 243 enrollments.

So today's online-course market is hardly dead. Publishing useful information just no longer gives a creator much of an advantage. The money is moving toward courses that offer something harder to get from a YouTube video, an AI chatbot or another $20 course.

What we see now	Evidence	What it tells us
Paid learning is still large	1.66M paid Coursera-Udemy subscribers	People still pay for online learning
Direct course businesses are active	Thinkific processed $71.5M in quarterly payments	Independent selling still works
Individual course purchases are weaker	Udemy consumer revenue fell 9% in 2025	The old marketplace model is under pressure
Larger education businesses are doing better	Thinkific Plus revenue grew 14% in its latest quarter	Higher-value offers are holding up better
Competition is extreme	About 272,000 Udemy courses	Publishing alone creates almost no moat

Did the online-course boom end after the pandemic?

The pandemic course boom is over, but online learning itself has settled into a much larger market than it had before the boom.

Pandemic-era conditions were unusually favorable for course sellers. Millions of people suddenly had more time at home, companies rushed training online and many subjects still had relatively little good digital content.

Those conditions disappeared.

Udemy's numbers show how far the consumer market has normalized. Its consumer revenue fell from $308.3 million in 2023 to $292.1 million in 2024 and then to $265.8 million in 2025. That is a 14% decline in two years.

Yet people did not suddenly stop paying to learn online. Udemy's enterprise revenue moved the other way, rising from $420.6 million in 2023 to $524.1 million in 2025. And after Coursera acquired Udemy, the combined company reported 1.66 million paid consumer subscribers, 44% above its harmonized year-earlier figure.

The market looks much more mature now. Consumers are becoming selective, subscriptions are taking share from individual purchases, employers are paying for workforce training, and recognized credentials have become more important.

The broad opportunity survived. The unusually easy conditions did not.

Are people still willing to pay for online courses today?

People are still willing to pay a lot for online courses when the course gets them closer to a result they actually care about.

Current pricing across the market makes this clear.

At one end, Udemy trains customers to expect heavy discounts on a huge catalog of self-paced courses. At the other, Maven currently recommends roughly $800 to $1,200 for cohort courses containing six to eight live hours and at least one project. Its guidance rises to around $1,800 to $2,450 when students receive more live instruction, multiple projects and a capstone.

That gap is revealing.

Customers have access to more free educational material than ever, yet some are still paying four figures for learning. The higher-priced products usually include access to an expert, applied work, feedback, accountability or a professional outcome.

Career learning is particularly strong. The World Economic Forum says AI and big data are currently the fastest-growing skills globally, followed by networks and cybersecurity. Employers expect 39% of workers' core skills to change by 2030.

When learning can help someone earn more, get promoted, pass an exam, launch a business or perform a difficult job, willingness to pay remains high.

A $2,000 program does not need to compete on price with a $20 Udemy course if the $2,000 program includes expert review of the student's actual work and the $20 product provides recorded videos.

What the buyer gets	Typical pricing power
Recorded explanations available elsewhere	Low
A well-organized specialist curriculum	Low to moderate
Templates, systems and practical workflows	Moderate
Projects with expert feedback	High
Live access to a recognized expert	High
A credential or clear career outcome	High

Is the online-course market too saturated now?

The online-course market is heavily saturated, and the numbers are worse for new creators than headline enrollment totals make them look.

Class Central's latest large analysis of Udemy counted about 272,000 courses and more than 908 million cumulative enrollments.

Those 908 million enrollments sound impressive until we examine how unevenly they are distributed.

The average Udemy course had about 3,331 enrollments. The median had only 243. The top 20% of courses captured 91.1% of all enrollments.

Half of the catalog therefore failed to reach 243 students even after being available inside one of the largest course marketplaces in the world.

Supply continues to arrive quickly as well. Udemy reported that instructors published an average of more than 6,300 courses per month during 2025.

That makes broad topics especially difficult. A new course called “Learn Python,” “Digital Marketing Masterclass” or “Introduction to AI” enters a market with thousands of substitutes.

Much narrower problems can still be poorly served.

“How to automate financial reporting with Python for a five-person finance team” targets a smaller audience, but the buyer immediately understands whether the product is relevant.

Today, specificity often matters more than total market size.

Udemy catalog measure	Latest large dataset
Courses	~272,000
Cumulative enrollments	~908M
Average enrollments per course	3,331
Median enrollments per course	243
Enrollments captured by top 20% of courses	91.1%

Has AI made online courses too easy to create?

AI has made average online courses much easier to produce, which raises the bar for anyone trying to sell one.

Course creators can now generate an outline, lesson scripts, exercises, quizzes, summaries, translations, worksheets, landing-page copy and email sequences with tools that cost almost nothing.

The platforms themselves are building this into their products.

Thinkific now offers AI-assisted course creation and an AI teaching assistant called Thinker. Coursera includes AI-powered course-building and learner features. Teachable has added tools for curriculum generation, quizzes, subtitles and translations.

For creators, this lowers production cost considerably.

It also means thousands of competitors receive the same advantage.

A polished 30-video curriculum once required enough time and skill to create some natural scarcity. Today, the production process can be compressed dramatically. Even the average length of Udemy courses fell from 7.7 hours in early 2023 to 3.79 hours in 2025, while thousands of new courses keep entering the platform every month.

Creating the files has become the easy part.

The harder questions are whether anyone trusts the creator, whether the material comes from genuine experience, whether students can apply it, and whether the creator already has some way to reach buyers.

Those parts are much harder to generate with one prompt.

Can ChatGPT and AI tutors replace paid online courses?

ChatGPT and AI tutors can already replace a large part of what basic informational courses used to sell, especially when learners mainly need explanations or help solving isolated problems.

Imagine someone learning Excel.

A few years ago, getting stuck on an INDEX-MATCH formula might mean reopening a course, finding the right lesson and scanning through the video.

Today, the learner can paste the formula into an AI assistant, explain what is going wrong and ask follow-up questions until the problem is solved.

Programming, software tutorials, language learning, marketing basics, statistics and countless other subjects now work the same way.

Online-learning companies clearly expect this behavior to spread. Coursera recently invested $100 million in LearnVector, an AI-native education company founded by Andrew Ng. Its stated goal is personalized one-to-one learning that adapts to each student. Thinkific's Thinker similarly lets course businesses provide AI support trained around their own learning content.

Generic explanation is therefore losing value quickly.

Structured practice, original examples, projects, feedback, credentials and expert judgment are much safer. AI can also improve those courses by giving students instant support between lessons.

The likely winner in 2027 is the course that uses AI to make learning better while giving the learner something a general-purpose chatbot cannot easily reproduce.

What are online-course buyers actually paying for now?

Online-course buyers increasingly pay for a faster route to an outcome, especially when the course reduces uncertainty or gives them access to expertise they cannot easily find elsewhere.

Maven's current pricing data is useful here because its marketplace focuses heavily on professional cohort courses. The platform recommends charging more as live interaction and applied project work increase. Its own marketplace data also found that courses priced at $950 or more generated 50% to 100% more revenue per marketplace landing-page visitor than cheaper courses.

That does not mean expensive courses automatically perform better. Maven attracts a specific kind of professional learner and instructor.

But it tells us where pricing power still exists.

Someone may refuse to pay $100 for 30 hours of generic marketing videos and happily pay $1,000 to have an experienced growth leader critique the campaign they are launching next week.

Professional credentials create another source of value. Coursera's catalog increasingly emphasizes Professional Certificates, employer-created programs and degree pathways because learners can show those achievements outside the platform.

Projects add another layer. A learner who finishes with a portfolio, working automation, business plan or software application has something concrete to show for the money.

The further we move from “watch these videos” toward “finish this with something useful,” the stronger the product becomes.

Can a prerecorded online course still make money?

Prerecorded online courses can still make money, especially when the seller already has distribution, but the passive-income version of the model is much weaker today.

Self-paced delivery still has wonderful economics. Once the product exists, another student can join without requiring another live teaching hour.

The difficulty is everything surrounding the videos.

Learners frequently struggle to finish self-paced education. Academic reviews of large open online courses continue to find extremely high dropout rates, often above 90%. A paid specialist course should have better engagement than a free MOOC, so we should not transfer those percentages directly. The underlying problem remains familiar: access to information does not guarantee that someone will use it.

Course maintenance is another issue.

A course on watercolor painting can remain useful for years. A course showing a specific AI workflow, advertising dashboard or software interface can become outdated within months.

Marketing also keeps running after the recording stops. Search positions change. Email lists need new subscribers. Paid ads need optimization. Competitors release new products.

A good self-paced course these days is better described as leveraged income than truly passive income. The teaching can scale without your time scaling at the same rate, but the business still needs attention.

Are cohort-based courses better than prerecorded courses now?

Cohort-based courses have become more attractive in subjects where students need feedback, accountability or expert access, because those benefits are much harder to commoditize than video lessons.

Maven is the clearest example of the model.

Its pricing recommendations currently run from about $800 to $2,450 for live cohort programs, depending largely on the number of live hours and the depth of project work.

That would be difficult to sustain if buyers valued only access to information.

The product is closer to a temporary learning environment. Students show their work, ask questions, receive feedback and progress alongside other people trying to solve the same problem.

Some instructors have built very large businesses this way. Maven has documented instructors with more than $1 million in course revenue, although these are clearly exceptional examples rather than something a new seller should use as a forecast.

There is a trade-off.

A cohort requires the instructor to show up. Running four cohorts per year is less passive than selling an evergreen video course every day.

But a creator who can sell 20 seats at $1,000 produces $20,000 from a cohort. A $100 prerecorded course needs 200 purchases to reach the same gross revenue.

When expertise is scarce and the audience is relatively small, that trade-off can work extremely well.

Is corporate online training a better business than selling courses to individuals?

Corporate online training currently looks healthier than mass-market individual course sales, particularly in fast-changing professional fields.

The pre-acquisition Udemy numbers make the split unusually clear.

Udemy Business revenue grew from $420.6 million in 2023 to $494.5 million in 2024 and $524.1 million in 2025. Consumer revenue went in the opposite direction, falling from $308.3 million to $292.1 million and then $265.8 million.

Thinkific is showing a similar pattern at a much smaller scale. In its latest quarter, Thinkific Plus subscription and commerce revenue grew 14%, while self-serve revenue fell 1%.

Corporate buyers have a practical reason to keep spending.

The World Economic Forum says 50% of workers in its latest employer research had already completed training as part of long-term reskilling strategies, up from 41% two years earlier. Employers expect 59 out of every 100 workers to need training by 2030.

AI makes that need more immediate. Companies cannot wait for universities to redesign degree programs every time a new model, coding tool or automation workflow appears.

A specialist creator can therefore sell the same underlying expertise in a very different format: private company workshops, employee academies, team licenses, implementation programs or recurring training.

One $10,000 company engagement can produce the same revenue as 100 consumers buying a $100 course.

The sales process is harder, but the economics deserve serious attention.

Are AI courses still worth selling in 2027?

AI courses are currently one of the strongest online-course categories, but creators have to accept that the material can age frighteningly fast.

Demand is easy to see.

Class Central found that newly launched AI courses attracted more than 2.7 million enrollments during 2025. Roughly one in five new Coursera courses in its analysis focused on AI. The World Economic Forum ranks AI and big data as the fastest-growing skills category among employers.

The problem is shelf life.

An AI-agent course recorded around a specific framework can look dated after a few releases. A prompt-engineering trick may disappear when the underlying model becomes better at understanding ordinary instructions. A software tutorial can break after an interface redesign.

This makes recurring formats unusually attractive.

Instead of selling one frozen “Complete AI Masterclass,” a creator can maintain a continuously updated membership, run quarterly cohorts, sell company workshops or update a specialist academy as the tools change.

The fast pace can actually help good instructors. Students have to keep learning.

It punishes anyone hoping to record the subject once and leave the product untouched for three years.

Do you need a big audience to sell an online course?

You do not need a huge audience to sell an online course, but you absolutely need a believable way to reach the right buyers.

The distinction between audience size and audience quality is easy to underestimate.

Five thousand email subscribers who all work in the profession addressed by a course can be more valuable than 500,000 entertainment followers who barely know why the creator is selling education.

Marketplaces can provide distribution instead.

That service is valuable enough that Udemy's revenue-share model changes dramatically depending on who found the customer. For a normal marketplace transaction generated by Udemy, instructors receive 37% of net revenue. When the instructor generates the sale through their own coupon or referral link, the instructor keeps 97%.

The gap tells us how valuable demand generation has become.

Course hosting itself is easy to buy from Thinkific, Teachable, Kajabi, Podia, Maven and many others. Video storage and checkout pages provide little competitive advantage by themselves.

Traffic and trust remain scarce.

That traffic can come from YouTube, an email newsletter, SEO, LinkedIn, a professional community, an existing consulting business, partnerships, affiliates or marketplace search.

Starting with zero audience can still work. Starting with zero distribution plan usually does not.

Should you sell an online course on Udemy or your own website?

Udemy is currently more useful for creators who need discovery, while selling independently makes more sense once the creator can generate demand without the marketplace.

The economics show the trade-off immediately.

An instructor keeps 97% of Udemy net revenue when the sale comes through the instructor's own referral link and 37% when Udemy produces the marketplace transaction.

Udemy has also moved increasingly toward subscriptions. In 2025, 72% of the company's total revenue came from subscriptions, compared with 66% a year earlier. Its instructor subscription revenue pool fell from 20% to 17.5% in 2025 and then to 15% in 2026.

That model can still be attractive if Udemy introduces thousands of students to an instructor who would otherwise have none.

Once a creator controls the audience, independent economics become much more appealing. The creator can set pricing, capture customer email addresses, bundle courses with services, sell memberships and decide what to offer next.

Maven provides a middle ground for premium professional education by combining marketplace discovery with creator-led cohort courses. Thinkific and similar platforms sit further toward ownership: they give creators the infrastructure while expecting them to generate most demand themselves.

There is no reason to be ideological about the choice. Early on, giving up margin to acquire customers can make sense. Once the audience belongs to you, continuing to give away most of the economics is much harder to justify.

Selling model	What you gain	What you give up
Udemy marketplace	Built-in discovery	Large share of marketplace-sale revenue
Udemy with your own referrals	Infrastructure and marketplace presence	You still generate the customer
Maven	Premium marketplace and cohort infrastructure	Less suited to cheap mass-market courses
Independent platform	Pricing and customer ownership	You must find buyers
Direct corporate sales	Much larger contracts	Longer and harder sales process

How much money can an online-course creator realistically make?

Online-course income is extremely uneven, and the typical course earns far less attention than creator success stories suggest.

Udemy gives us one of the clearest datasets.

As seen above, its average course had around 3,331 enrollments at the end of 2025, while the median had only 243. More strikingly, the top 20% of courses accounted for over 91% of all enrollments.

That is a heavily concentrated market.

At the upper end, major course platforms can point to instructors generating hundreds of thousands or even millions of dollars. Those businesses clearly exist. They tell us what is possible rather than what a new instructor should expect.

A more useful calculation starts with the offer.

A $100 course needs 100 sales to generate $10,000 in gross revenue.

A $500 specialist course needs 20.

A $1,000 cohort needs 10.

A $5,000 corporate program needs two clients.

That arithmetic explains why higher-value niches can beat huge audiences. Selling ten places to professionals with an expensive problem may be far easier than attracting thousands of people to another general course.

The realistic earning potential depends heavily on price, trust, distribution and how painful the customer's problem is. “How much do course creators make?” therefore has no useful single average.

Which online-course topics look strongest for 2027?

The strongest online-course topics right now solve expensive, fast-changing or professionally important problems.

AI is the obvious example. Employer demand is rising quickly, the tools change constantly and workers have a strong reason to keep their skills current.

Cybersecurity has similar characteristics. The World Economic Forum ranks networks and cybersecurity immediately behind AI and big data among the fastest-growing skills.

Software development, data analysis, sales, finance, specialized marketing, professional certification and industry-specific compliance can also work because competence has measurable economic value.

But a professional topic is not automatically good and a hobby topic is not automatically bad.

The useful distinction is how replaceable the course feels.

“Learn photography” competes with an ocean of free material.

“Underwater macro photography with a working wildlife photographer” has a clearer reason to exist.

“Learn Excel” is brutally crowded.

“Excel modeling for first-year private-equity analysts” speaks to a specific person who has a specific problem.

Narrowing a course reduces the theoretical audience. It can simultaneously increase the percentage of that audience willing to pay.

That is often the better trade in today's crowded market.

What makes an online course hard for AI and competitors to copy?

The hardest online courses to copy are built around real expertise, proprietary material, feedback, reputation, community or access rather than around the lesson videos alone.

A simple test works surprisingly well.

Give an AI model the public course title and curriculum. How much of the product could it reproduce?

If the answer is “most of it,” the course has a problem.

AI can already generate a decent explanation of Facebook advertising, Python loops, copywriting frameworks or basic financial accounting.

It cannot automatically reproduce ten years of mistakes made managing $50 million in advertising spend. It does not own your proprietary dataset. It cannot give a student access to your hiring network. It cannot certify someone on behalf of an institution. And it cannot replace informed feedback from an expert who understands the student's actual project in every situation.

Real-world proof also compounds.

Udemy reported that its top instructors updated their courses about five times a year on average in 2025. That frequency gives us a clue about what serious course businesses already require: useful courses increasingly behave like maintained products.

For a creator entering the market now, the course videos should probably be viewed as one layer of the product. The durable value sits in the things competitors cannot recreate over a weekend.

What usually kills an online-course business?

Weak distribution now kills more online-course businesses than the cost of creating the course.

AI has made production cheap. Course platforms have made hosting cheap. Payment processing is easy. None of those tools guarantee customers.

Suppose we sell a $100 course and spend $60 in advertising to acquire each buyer. Payment fees, refunds, taxes, affiliate commissions and customer support quickly eat into what remains.

A $1,000 specialist product has much more room to acquire a customer profitably, assuming buyers see enough value to pay the price.

Marketplace dependence creates a different risk. Udemy reduced the share of subscription revenue allocated to instructors as its business shifted toward recurring subscriptions. Creators inside somebody else's ecosystem ultimately operate under somebody else's economics.

Poor retention can also hurt recurring course businesses. Coursera's combined enterprise net retention rate was 91% in its latest reported quarter, down from 95% on the harmonized year-earlier measure. Even giant learning platforms have to keep proving their value to existing customers.

Then there is maintenance. A course that needs rerecording every six months has a very different labor profile from an evergreen course about drawing fundamentals.

The attractive headline margin on digital products can hide a lot of work.

Before creating a course, we would spend more time proving the distribution channel and willingness to pay than choosing cameras, microphones or course software.

Are online courses still worth selling in 2027?

Yes, online courses are still worth selling in 2027, but we would only enter this market with a course that solves a specific problem for a clearly defined group of people.

The evidence points in one direction.

People still pay for online learning. The combined Coursera-Udemy business has 1.66 million paid consumer subscribers. Thinkific customers continue to process tens of millions of dollars every quarter. Employers expect huge reskilling needs. Premium course platforms continue to support programs priced above $1,000.

At the same time, the easiest course model has clearly weakened.

Udemy's consumer revenue fell for two consecutive years before the Coursera acquisition. Half of its courses had fewer than 243 enrollments in Class Central's latest analysis. More than 6,000 new courses were being published there every month during 2025. AI can now explain many beginner topics instantly and help creators produce competing courses much faster.

Those facts make a broad prerecorded course from an unknown instructor a weak bet today.

The more attractive opportunities look different: narrow professional courses, cohort programs, certification preparation, continuously updated AI training, courses attached to communities, programs with project feedback, company academies and education products connected to consulting or other services.

A creator also does not have to choose one format. A free tutorial can attract the audience. A self-paced course can serve the entry level. A cohort can offer feedback. Corporate training can monetize the highest-value problems.

That gives us a much clearer answer than saying the course business is either “dead” or “booming.”

Online courses are still worth selling. Generic information is becoming harder to sell every year, while expertise that helps people get a real result can still command serious money.