Course Platform Landscape: What to Steal, What to Skip
Context: You're a solo operator — your own backend, frontend, Stripe, Supabase, Clerk, running on a single VPS (4vCPU/8GB/160GB). You're the only instructor and the only support person. That's a fundamentally different shape than any of these three platforms, so the lens here isn't "copy their feature list" — it's "which of their mechanisms solve a problem you actually have."
1. Kajabi — the all-in-one business infrastructure model
What it actually is: Not a course host. A full creator-business stack — courses, email marketing, funnels, landing pages, checkout, community, podcasts — sold as a single $179–$499/month subscription with 0% revenue cut.
Business model: Flat subscription, no revenue share, tiered by contact list size and product count. As of Jan 2026 they raised prices and tightened contact limits, pushing serious creators toward higher tiers.
What actually drives outcomes on Kajabi (per their own data across ~32,000 courses): per-lesson discussion features correlate with a completion-rate jump from ~42.6% to ~65.5%. That's a genuinely useful data point — discussion at the lesson level, not just a global forum, is what keeps people finishing.
Mechanisms worth studying:
- Products vs. Offers separation. A course is a product; how you package/price it (one-time, payment plan, bundle, upsell) is an offer. Same content, multiple commercial wrappers. This is a clean data model — worth mirroring in your schema even at small scale (don't couple
coursetoprice). - Cohort courses with deadlines. Structured start/end dates and pacing rules measurably improve completion vs. pure self-paced "watch whenever."
- Post-purchase automation. The moment payment succeeds, a sequence fires: welcome, first-lesson nudge, "here's what to expect." This isn't marketing fluff — it's onboarding, and onboarding is retention.
- Gamification (points, badges, streaks) tied to actions that matter (finishing a module, submitting an assignment) rather than vanity actions (logging in). They explicitly warn against over-rewarding low-value behavior.
What to ignore: the funnel/email-marketing/affiliate machinery, the multi-tier pricing complexity, anything about "winback campaigns" — that's built for creators acquiring cold traffic at scale. You're not running lead magnets and abandoned-cart sequences; you likely already know your students or reach them through your existing content brand.
2. Udemy — the marketplace model (informative mainly as a cautionary tale)
Business model: Free to publish, revenue-shared per sale — 97% if the student came through your coupon/referral link, only 37% if they found you via Udemy's own marketplace/ads. A separate subscription pool (Udemy Business) pays instructors by share of watch-time minutes, and that pool's payout rate has been cut from 25% (2023) → 20% → 17.5% → 15% as of Jan 2026. Udemy also runs unstoppable platform-wide sales that can discount a $199 course to $9.99, and instructors can't opt out on the standard agreement.
Why this matters for you, even though you're not on a marketplace: this is the clearest illustration of the core risk of not owning your platform — no email list, no pricing control, no student relationship, and the platform can and does change the deal underneath you unilaterally. You already avoided this by building your own stack. Worth naming explicitly as validation of that decision, not just a side note.
What's actually worth taking:
- Two-tier attribution for revenue channels (your traffic vs. platform-discovered) — even on your own platform, tracking which channel a sale came through (blog post, YouTube, direct, referral) is valuable analytics you can build cheaply with Stripe metadata + UTM params in Supabase. You don't need Udemy's complexity, just the underlying idea: attribute every sale to a source.
- Consumption-based engagement tracking. Minutes-watched, quiz completion, exercise attempts — Udemy weights all of this as "consumption," not just video playback. If you ever build analytics for your own course quality, track engagement across every interactive surface, not just "% video watched."
- Instructor payout minimum thresholds — a practical ops pattern (batch payouts, don't process every $2 transaction) if you ever add revenue-sharing (e.g., guest instructors, affiliates).
What to ignore entirely: the revenue-share mechanics, the marketplace discovery algorithm, coupon systems designed to route around platform fees — none of it applies when you are the platform and keep 100% minus Stripe's ~2.9%+$0.30.
3. Skillshare — the engagement/royalty model
Business model: Subscription-only for students (~$168/yr). Teachers are paid from a rotating royalty pool (~20–30% of subscription revenue) split by share of total minutes watched — not per-sale. As of 2026 they've added stricter qualification gates (50+ followers, minimum monthly watch-time) to cut payouts to inactive/low-engagement teachers, reflecting a broader shift toward rewarding depth of engagement over catalog size.
What's genuinely instructive here, more than Kajabi or Udemy:
- Project-based classes outperform passive ones. Skillshare's structural bet is that every class ships with a hands-on project + teacher feedback loop. Video alone is not the product; video + an artifact the student produces + feedback is the product. If your course platform is video-lecture-only right now, this is the single highest-leverage structural change you could make — even as a solo instructor, a lightweight "submit your output, I'll leave one comment" loop dramatically changes perceived value and actual learning outcomes.
- Positive-review bonus incentives — rewarding "met/exceeded expectations" ratings specifically (not just any rating) is a subtle but smart signal design: it optimizes for outcome quality, not just star-rating volume.
- Royalty tied to watch-time, not enrollment — not directly applicable to your solo-instructor, one-time-payment model, but the underlying idea — measure engagement, not just conversion — should shape your own internal dashboard. A student who buys and never opens lesson 2 is a support/design problem, not a win.
What to ignore: the whole royalty-pool payout mechanic and follower-count gating — irrelevant when there's one instructor and no pool to split.
4. Since you're solo instructor + solo support: the real constraint
None of these three platforms are designed for a single person doing content, sales, delivery, and Q&A. Their feature sets assume either (a) many instructors sharing infrastructure (Udemy, Skillshare) or (b) one creator running a full sales/marketing operation with a team (Kajabi's ideal customer). Your actual bottleneck is different: your time, especially for Q&A/support.
That reframes which of the above matters most:
| Pattern | Why it matters for a solo operator |
|---|---|
| Lesson-level discussion (Kajabi data point) | Answer a question once, publicly, and it serves every future student — turns 1:1 support into 1:many |
| Project submission + feedback loop (Skillshare) | Async, batchable — you can review submissions in a weekly block instead of live chat |
| Cohort pacing with deadlines (Kajabi) | Concentrates your Q&A load into predictable windows instead of a constant trickle spread across self-paced students at random stages |
| Post-purchase onboarding automation (Kajabi) | Front-loads expectation-setting so fewer support questions are "how do I even start" |
| Attribution tracking (Udemy) | Tells you which content channel actually converts, so you spend marketing time efficiently instead of everywhere |
Concrete next step to consider: an FAQ/discussion thread attached to each lesson (not a generic community forum) is probably the single highest-leverage feature you're missing, given you're both the only instructor and the only support line. It directly converts repeated 1:1 support load into a compounding public asset — and it's a small, well-scoped feature to build on your existing Supabase schema (a lesson_id foreign key, basic threading, your own auth via Clerk — no new infra needed).
Sources
- Kajabi 2026 pricing/review coverage: builtbyfoundry.io, marketerschoice.com, learningrevolution.net
- Kajabi completion-rate data: ruzuku.com
- Udemy revenue share terms: support.udemy.com (official), classcentral.com (analysis)
- Skillshare royalty/teacher terms: help.skillshare.com (official)