Teaching Online: Real Per-Student Math on Udemy, Teachable, Outschool, Wyzant

A $199 course discounted to $9.99 with a 37% revenue share pays the instructor $3.70 per student. A Teachable Builder plan costs $69 a month annualized and takes 0% of your sales. An Outschool teacher running a six-kid live class at $20 a seat clears $84 per session, before scheduling it around her day job. Three numbers, three platforms, three completely different business models hiding under the same phrase: teaching online.

Most readers I talk to assume online teaching is one decision. It isn’t. It’s four wildly different income models with different math, different time demands, and different ceilings. If you’re holding down a full-time W-2 and thinking about turning what you know into supplemental income, the platform you pick decides whether you’re building an asset, renting your weekends, or chasing pennies through a coupon link.

How we got here: the marketplace bargain that quietly changed

Udemy launched the modern course economy on a simple deal: you make the course once, we bring the students, we split the money. For a long stretch, that split looked reasonable. Instructors with marketing chops could clear six figures, and even hobbyist creators caught the long tail of evergreen topics like Excel, Photoshop, and yoga.

Then the model shifted. Udemy pivoted hard toward subscriptions. According to Class Central, subscription revenue (Udemy Business plus personal plans) hit 74% of total Udemy revenue by Q3 2025, up from 68% a year earlier, and the company has been openly throttling single-course sales to push that mix higher. The instructor subscription share dropped from 25% in 2023 to 20% in January 2024, to 17.5% in January 2025, to 15% in January 2026. That’s a 40% reduction in three years. Total instructor payouts peaked at $209.5 million in 2023 and have been sliding since.

Teachable, the self-hosted alternative, went the opposite direction by retiring its free plan in early 2025 and tightening published-product limits across tiers. Outschool kept its model boring and consistent: 70% to the teacher, 30% to the platform, live classes only. Wyzant stayed in its 1-on-1 tutoring lane. Four very different bets on what online teaching should be. The reader’s job is to figure out which bet matches your goals.

The per-student math, platform by platform

Grab a pen, let’s do the math together. The headline rates are misleading until you back out fees, discounts, and the cost of bringing your own students. Here’s what actually lands in your bank account on each platform:

• Udemy organic sale: a course discounted to $9.99 (Udemy’s standard sale price you can’t opt out of) pays 37%, or roughly $3.70 per student.
• Udemy coupon sale using your own link: 97% of the sale price. On a $19.99 promotional price that’s about $19.39 per student, but you brought the buyer.
• Udemy Business (subscription): 15% of a revenue pool split by watch-time across thousands of courses. Realistic per-student value is pennies, not dollars.
• Teachable Builder plan: $69/month annualized, 0% transaction fee, plus the standard Stripe processing of about 2.9% + $0.30. A $100 sale nets roughly $96.80, minus your monthly subscription cost.
• Teachable Starter plan: $29/month annualized, but a 7.5% platform fee plus Stripe. A $100 sale nets about $89.30.
• Outschool live class: 70% of the seat price times enrolled learners. A $15 seat with 8 kids pays $84 for a 45-minute session.
• Wyzant 1-on-1 tutoring: you set your hourly rate, Wyzant takes a service fee that varies by tenure and hours logged. Experienced tutors clear $40–$70+ an hour after fees.

Notice how Udemy’s $3.70 organic and Outschool’s $84 per session aren’t even the same kind of number. One is per student of a recorded asset; the other is per session of live work. Conflating them is how people end up disappointed.

Here’s the part nobody wants to tell you about the Udemy subscription model: even if your course gets watched, the per-minute payout is calculated against a global pool that includes every other instructor on Udemy Business. Your “share” can be cents per student per month. I’ve analyzed dozens of instructor dashboards. Clear pattern: people who relied on Udemy’s organic discovery in 2019-2021 are earning 30-50% less for the same catalog today, and the trend line is still pointing down.

Which format actually fits a 9-to-5

The schedule question matters more than the revenue question for most readers. You can’t teach a Tuesday 4 p.m. Outschool class if you’re in standup at 4 p.m. Here’s how the formats stack against a full-time job:

Asynchronous platforms (Udemy, Teachable, Thinkific, Skillshare) are the only ones that survive a real W-2 schedule cleanly. You film once, you upload, you market on weekends. The income is unpredictable but the time commitment after launch is mostly marketing and student-question replies. Realistic build time for a first viable course: 40-80 hours of work spread over 2-3 months, mostly evenings.

Live platforms (Outschool, Wyzant, Preply, Italki) trade flexibility for predictability. You earn when you teach, you don’t earn when you don’t. The hourly rates are higher but they’re capped by your calendar. Outschool’s own marketing cites about $50/hour average, ZipRecruiter pegs the average annual at $46,590 (around $22.40/hour), and Glassdoor’s user-reported average is closer to $31/hour. One published first-person teaching report from April 2025 hit $117.68/hour after building a multi-year student base. The spread tells you everything: it’s a skill platform, not a passive one.

Back at the bank, I’d tell clients with side gigs the same thing I’ll tell you: do the real hourly math. Take projected monthly revenue, subtract platform fees, subtract your prep and admin time, divide by total hours. A Teachable course that grosses $800 a month but ate 60 hours to build and 4 hours a week to maintain isn’t paying $800 a month. It’s paying about $35 an hour in month one and climbing as the build cost amortizes. That’s a useful number. “Gross revenue” isn’t.

Smarter approaches and what I actually recommend

Two years ago I told a friend who teaches data analysis full-time to skip Udemy and build her course on Teachable, even though Udemy’s “free traffic” looked attractive. Her objection: “I don’t have an audience yet.” My prediction: Udemy’s organic share would keep falling and the subscription pool would dilute further. Both happened, on schedule. She launched on Teachable in 2024, brought students through LinkedIn posts, and last quarter cleared more on 180 sales than three of her Udemy-only friends did combined on 4,000+ enrollments.

Here’s what I tell people now: pick the platform that matches what you already have, not the one with the best pitch. If you have an audience (newsletter, LinkedIn following, podcast, YouTube), go Teachable or a comparable self-hosted platform. You keep almost all the revenue, you control pricing, and the monthly fee is trivial against even modest sales. If you have credentials but no audience, and you’re teaching K-18 subjects, Outschool’s marketplace does the discovery for you in exchange for the 30% cut. If you teach a skill adults pay to learn 1-on-1 (test prep, language, music, coding), Wyzant or Preply will get you bookings faster than building a website. If you have none of the above and you’re just starting, Udemy organic is a legitimate way to learn what students respond to, but treat the income as research budget, not a business.

One more thing worth saying directly: the platforms with the best marketing pages are usually the ones taking the biggest cut. That’s not a coincidence. They can afford the marketing because they take more of your revenue.

Where to start (and what to skip)

The 80/20 of online teaching is this: 80% of your income outcome is decided by whether you bring your own students, not by which platform you pick. Marketplaces (Udemy, Outschool) charge you 30-85% in exchange for discovery. Self-hosted (Teachable) charges you a flat monthly fee in exchange for nothing but the tech. The platform isn’t the asset. Your audience is.

Three profiles, three plays:
• Full-time worker, no audience, technical or office skill: launch one Udemy course as a $0-budget proof of concept. Treat any income as feedback. Spend 6 months building a LinkedIn or newsletter audience around the topic in parallel.
• Full-time worker, small audience (under 5,000), teachable skill: skip Udemy. Go Teachable Builder at $69/month annualized, price the course at $99-$199, sell to your audience directly. 50 sales clears more than 5,000 Udemy enrollments.
• Teacher, tutor, or K-18 subject expert with daytime flexibility: Outschool for groups, Wyzant for 1-on-1. Both pay predictable per-session income. The trade is you must show up live.

I’m telling you this because I’ve seen it happen: people quit their day job on the strength of one good Udemy month, then watch the algorithm change and the revenue evaporate. Two complications to plan for. First, Udemy can change revenue share rules unilaterally (they have, four times since 2023); never build your full income on a platform whose terms you don’t control. Second, Outschool teachers report it takes 6-12 months to build a steady booking calendar; the first three months are slow and discouraging. Solve that by treating early classes as portfolio-builders, not paychecks, and asking every parent for a written review.

This week, do one thing: write down what you’d teach, who would buy it, and how you’d reach them without the platform’s help. If you can’t answer the third question, your platform is going to keep most of your money. Then pull pricing pages directly from the source. Start with the U.S. Bureau of Labor Statistics for honest hourly benchmarks in your subject area, and IRS guidance on self-employment income before your first $600 deposit lands. That gap between your platform’s promise and your reality, that gap IS the answer.