211 Learners. One Question. Zero Assumptions.

We posted an open-ended survey across online communities. The groups ranged from 10,000 to 190,000 members. Two questions:

“What is the one thing you love in an online course?”
“What is one thing you hate or think can be improved?”

211 learners responded. Some with a single line. Some with detailed paragraphs about their frustrations and what they wished existed.

No demographics were controlled. No categories were pre-set. We wanted to hear what learners actually care about on an open canvas.

Responses were categorized, scored by frequency, and analyzed. The findings were published at a Springer AI/Data Engineering conference and won the Best Paper Award in 2022.

The loudest signal: learners want courses built around real-world projects. Not toy examples. Not textbook exercises. Software that resembles what they’d encounter on the job.

Specific demands:

• Production-grade code. Proper formatting, file naming, folder structures. Not tutorial shortcuts.
• Development environment setup. Keyboard shortcuts, tooling, documentation navigation. The unglamorous infrastructure that makes real engineers productive.
• Interview readiness. Courses should map directly to job interview patterns. Learners take these courses to get jobs.

The broader research backs this up. Over 70% of MOOC participants are job-motivated. Courses built around real cases show significantly higher completion rates.

What this meant for us: Every course at Invact Metaversity is built around applied projects. The project is not the capstone. It is the spine.

The second-strongest signal was not about what’s taught. It was about how.

Video length: 5 to 10 minutes. Respondents were clear that long lectures lose them. A study of 6.9 million video sessions confirmed shorter videos are significantly more engaging.

Instructor on screen. Always. Strong preference against voice-over-slides. Learners want to see a person. Platform feedback showed the same: participants prefer watching a hand write an equation over staring at static text.

Simple English, beginner-first. Non-native speakers consistently asked for plain language. Research confirms they watch videos at a slower pace because they struggle with language, not content.

Energy and personality. Survey responses included: “should not be monotonous,” “at least try to be funny.” The instructor is the single most important factor in online course retention, with the largest positive effect on outcomes.

What this meant for us: Invact does not replicate the lecture hall. Sessions are live interaction, real-time doubt clearing, proof-of-work showcases. The format respects how people actually learn on screens.

Respondents were vocal. They want exercises after every section. Mini projects between modules. Major capstones at the end. They want to build.

The research is unambiguous. Students benefit as much as six times more from doing things than from watching video alone. Retrieval practice (quizzes, exercises, applied challenges) cements learning in long-term memory. Watching is passive. Doing creates retention.

Respondents also flagged a critical frustration: being tested on material not covered in the course. A design failure common in MOOCs where assessment and content are developed separately.

A suggestion that appeared multiple times: teaching via games. CSS Battle, Flexbox Froggy, and similar platforms were cited as effective. Gamified learning increases participation and makes automated grading scalable.

What this meant for us: Invact uses proof-of-work as the primary assessment. Students build, present, and receive peer and mentor feedback. Real work, reviewed by real people.

Across 211 responses about what makes online courses work, almost nobody mentioned community or forums.

Not because they don’t matter. Because the tool they forgot to mention was the tool they were using. They were responding on community forums. Community was so foundational it was invisible to them.

But the research is clear. Students need help channels to ask for support. Without them, they stop participating entirely. Forums are students’ most preferred tool, yet too many channels “hinder the real learning process” as participants get lost in noise.

The survey identified what learners consciously want. The deepest insight was what they unconsciously depend on: community infrastructure that feels so natural they forget it exists.

This became the central thesis. Not “add a forum.” Build an environment where community emerges from the architecture itself. Serendipitous encounters. Spatial proximity. Shared spaces. The way it works on a physical campus.

This finding is why Invact Metaversity exists.

Published as “Data-Driven Strategies Recommendation for Creating MOOCs for Effective Online Learning Experience.” Springer, 2022. Best Paper Award at the International Conference on Computing, Intelligence and Data Analytics.