"I want to sell courses online" splits into three different businesses with three different platforms. Choosing the model first — then the technology — is the difference between a course business and an expensive video library. Here is the map.
The three models
Self-paced academy. Recorded content, quizzes, certificates; sold individually or by subscription. Lowest operational load, hardest retention: completion rates for pure self-paced hover in single digits industry-wide. Wins on evergreen topics with strong SEO.
Cohort-based courses. Batches with schedules, live sessions, assignments, community. Completion and pricing power both multiply — cohorts routinely charge 5–10x self-paced prices — at the cost of instructor time per batch. The dominant model for career-outcome education in India.
Marketplace. Many instructors, revenue share, discovery. A two-sided business where the platform problem (supply, quality, discovery) dwarfs the software problem. Do not start here unless the marketplace is the business thesis.
What each needs technically
| Model | Core stack | Indicative build |
|---|---|---|
| Academy | Content delivery, payments, progress, certificates | ₹5–10L |
| Cohort | + batches, live integration, assignments, community | ₹8–16L |
| Marketplace | + instructor onboarding, payouts, discovery, moderation | ₹15–35L |
(Ranges track our LMS cost guide; SaaS platforms like Graphy/Teachable serve early validation well — the build-vs-subscribe math flips around 500–2,000 active learners.)
Non-negotiables across models in 2026: mobile-first delivery (Indian learners study on phones — app economics), UPI/EMI payments for high-ticket courses, WhatsApp for nudges and reminders, and basic video protection (signed URLs, watermarking — accepting that determined piracy is unstoppable and casual piracy is what you prevent).
The 2026 differentiator: AI inside the course
Grounded doubt-solving turns recorded content into something closer to tutoring — answering from your curriculum at 11pm, escalating gaps to instructors. For self-paced especially, this measurably moves completion, which moves refunds, reviews, and referrals.
Economics honesty
Platform cost is the small number. Content production, instructor time, and customer acquisition dominate course-business P&Ls. Validate with a cohort on rented tools before building anything — the validation discipline applies doubly to education.
Building an academy or edtech product? Tell us the model and we will scope it honestly — or get a range from the calculator first.
Frequently asked questions
How much does it cost to build an online course platform?
By model in 2026: a self-paced academy runs ₹5–10 lakh, a cohort platform with live sessions and community ₹8–16 lakh, and a multi-instructor marketplace ₹15–35 lakh. Validate on SaaS tools first; build past 500–2,000 active learners.
Which online course model should I choose?
Self-paced academies win on evergreen SEO topics but suffer single-digit completion. Cohort courses charge 5–10x more with far better completion, at instructor-time cost — the dominant Indian model for career outcomes. Marketplaces are a two-sided business; start there only if that's the thesis.
How do I protect course videos from piracy?
Signed URLs, domain-locked playback, and watermarking stop casual sharing — the realistic goal. Determined piracy is unstoppable industry-wide; pricing, community, and freshness are the actual moats.
What increases course completion rates?
Cohort structure with schedules and community, mobile-first delivery with WhatsApp nudges, and AI doubt-solving grounded in your curriculum answering at 11pm. Completion moves refunds, reviews, and referrals — it is the metric to obsess over.