Custom LMS or ready-made platform: what building from scratch really costs

Your school has grown: 800 students, five courses, three curators. Then someone at the weekly meeting says, "Let's build our own platform — I'm tired of paying for someone else's." What usually follows is six months of pain.

This is not an argument against custom development. It is an argument for doing the math before, not after.

What "your own platform" actually contains

School owners picture a website with video lessons. The real module list looks different:

• Video: upload, transcoding into several qualities, CDN, download protection, resume-from-last-position.
• Course structure: sections → modules → lessons, drafts, drip release, prerequisites.
• Access control: roles (student, curator, admin), plans, access periods, renewals.
• Payments: gateway, recurring charges, refunds, receipts, accounting reconciliation.
• Homework: submission, review queue, curator comments, resubmission.
• Communications: email sequences, reminders, Telegram notifications.
• Gamification: points, streaks, levels, leaderboards.
• Analytics: completion per lesson, drop-off points, cohorts.
• Admin panel — by volume a separate product, often bigger than the student side.
• Infrastructure: backups, monitoring, logs, personal data handling.

None of these is a one-week feature. The payment module alone — recurring billing, refunds, correct handling of failed charges — reliably eats 80–150 hours.

Doing the math

Take Ukrainian market rates: mid-level full-stack 25–40 $/hour, tech lead or architect 45–60, designer 20–30. A freelancer costs less, an agency more, an in-house team adds taxes and paid idle time.

A rough MVP estimate for a platform that can carry a real school:

• Architecture, auth, roles — 120–180 h
• Courses, lessons, access — 150–220 h
• Video and storage — 100–160 h
• Payments and subscriptions — 120–200 h
• Homework and curator review — 100–160 h
• Email and Telegram communications — 80–120 h
• Analytics and admin panel — 150–250 h
• Design and front-end — 180–260 h
• QA, deployment, post-test fixes — 150–250 h

Total: 1,150–1,800 hours. At 30 $/hour that is $34,000–54,000 before your first student. With no gamification, no landing page builder, no CRM.

Then the monthly bills: hosting, database and CDN run 80–400 $ depending on video volume; the payment gateway takes 2–3% either way; support and development need 25–40% of the annual build budget — another $10,000–20,000 a year just to keep the system from rotting.

For comparison, a ready-made platform costs roughly 1,000–5,000 UAH a month — 12,000–60,000 UAH a year. An order of magnitude apart.

The resource more expensive than money

Building an MVP takes 5–8 calendar months with a fully dedicated team, plus 2–3 months before it stops breaking on live students.

Half a year without launches, for a school earning 300,000–500,000 UAH per cohort, means millions of hryvnias in lost revenue. That is the real price of building your own, and it almost never makes it into the estimate.

The second hidden item is single-person dependency. The developer who wrote your LMS resigns, and you either pay whatever they ask or hire a team that spends three months reading undocumented code.

When custom development genuinely makes sense

1. The LMS is your product. You sell the platform to other schools, not courses. Then development builds an asset.
2. A unique learning mechanic at the core. A simulator, an auto-graded practice environment, an adaptive path built on your own data — something no ready-made system has and that is your main value.
3. Hard client requirements. Corporate buyers demanding on-premise deployment, integration with their HR stack and a security audit.
4. Very large scale. From 10,000–15,000 active students, when platform fees start to exceed the cost of an in-house team.

If none of these describes you, a custom LMS is the most expensive way to get something that already exists.

When it is almost always a mistake

• "We're missing one feature." Negotiating a change request or building an add-on over the API is cheaper.
• "We want our own design." Customisation, custom domains and branding ship with most platforms and cost zero development hours.
• "We don't want vendor lock-in." An in-house team is lock-in too — pricier and more fragile.
• "We ran the numbers, it came out cheaper." Such calculations almost always omit support, QA, infrastructure and the cost of lost time.

The hybrid most schools end up with

The healthiest scenario is a ready-made platform as the core, plus custom pieces where you are genuinely unique: a separate service running your simulator and pushing results back through the API; your own Telegram bot on top of the platform funnel; custom dashboards built on data exports; routine automation through an AI agent that operates the platform directly.

You pay a few thousand hryvnias for the core and spend development budget only on what makes money.

Seven questions that settle the debate

1. What specific job does no ready-made platform do?
2. What is that job worth in money per year?
3. Who maintains the system two years from now?
4. How many launches will you miss while building?
5. What happens to a thousand students' data during migration?
6. Who owns security for personal data and payments?
7. Are you prepared to fund platform development forever?

If the answer to the first question is "well, it's inconvenient there", you don't need a custom LMS. You need a different ready-made one.

Instead of a conclusion

We built CREO precisely as the core for that hybrid: gamification that keeps course completion at 70–80%, payments via WayForPay and Stripe, Telegram funnels, a sales page builder and a CRM. Support replies in about 15 minutes, and more than 100,000 students have already gone through the platform. Separately we are shipping MCP control, so an AI agent can create courses, enrol students and pull reports without manual admin work — beta testers get it first.

The beta starts in August 2026. Join the early list at platform.creo.ua.