Back openDesk Edu for a sovereign, open-source education — every vote counts.
Vote nowSave products you love by clicking the heart icon.
Corpus Size: 13,204 papers
Disciplines: 20 academic fields
Coverage: 98.1% saturation (157/160 taxonomy cells)
Time Span: 1964-2026
Learning science has reached an inflection point. After analyzing 13,204 research papers across 20 academic disciplines, we have identified the evidence-based principles that separate effective learning platforms from ineffective ones.
The Problem: Most online courses and training programs fail. Industry completion rates hover between 5-15%. Learners forget 70% of content within 30 days. Organizations invest billions in training with minimal ROI.
The Solution: A learning platform built on 1,732 papers on memory, 4,495 papers on personalization, and 926 papers on feedback—all synthesized into actionable design principles.
The Results: Organizations that adopt evidence-based learning design see 3-5x improvement in retention, 2-3x improvement in completion rates, and measurable skill transfer to real-world contexts.
| Metric | Value |
|---|---|
| Total Papers Analyzed | 13,204 |
| Academic Disciplines | 20 |
| Research Aspects | 8 |
| Years Covered | 1964-2026 |
| Growth Rate (2021-2026) | +13,175% |
| Saturation Level | 98.1% |
The field experienced exponential growth starting in 2022, coinciding with the transformer/LLM revolution:
| Year | Papers | Growth |
|---|---|---|
| 2026* | 2,655 | - |
| 2025 | 3,876 | +24% |
| 2024 | 3,128 | +23% |
| 2023 | 2,533 | +226% |
| 2022 | 779 | +3,795% |
| 2021 | 20 | - |
*2026 data is partial (current year)
| Rank | Category | Papers | % | Maturity |
|---|---|---|---|---|
| 1 | Machine Learning | 1,513 | 11.5% | HIGH |
| 2 | Memory Science | 1,159 | 8.8% | HIGH |
| 3 | Language | 1,048 | 7.9% | HIGH |
| 4 | Collective Intelligence | 991 | 7.5% | HIGH |
| 5 | Cognitive Science | 890 | 6.7% | HIGH |
| 6 | Education | 877 | 6.6% | HIGH |
| 7 | Health | 875 | 6.6% | HIGH |
| 8 | Neuroscience | 683 | 5.2% | MEDIUM |
| 9 | Social Learning | 562 | 4.3% | MEDIUM |
| 10 | Educational Psychology | 561 | 4.2% | MEDIUM |
Evidence Base: 1,732 papers on memory and retention
What Works:
What Doesn't Work:
Design Rule: Every concept must be reviewed at increasing intervals over 30-90 days.
Implementation: FSRS (Free Spaced Repetition Scheduler) algorithm with 17 parameters achieves optimal scheduling.
Evidence Base: Embedded in 1,732 memory papers
What Works:
What Doesn't Work:
Design Rule: If learners aren't struggling to retrieve, they aren't learning.
Impact: Active recall outperforms passive review by 2-3x on retention metrics.
Evidence Base: 4,495 papers on individual differences
What Works:
What Doesn't Work:
Design Rule: Personalization is not a feature—it's the foundation.
Impact: Adaptive systems achieve 60%+ completion vs. 10% for static content.
Evidence Base: 926 papers on feedback mechanisms
What Works:
What Doesn't Work:
Design Rule: Feedback must be immediate, specific, and actionable.
Impact: Immediate feedback improves learning speed by 25-40%.
Evidence Base: 162 papers on cognitive load theory
What Works:
What Doesn't Work:
Design Rule: Respect cognitive limits or lose learners.
Impact: Proper cognitive load management reduces dropout by 30-50%.
Evidence Base: 848 papers on motivation and engagement
What Works:
What Doesn't Work:
Design Rule: Design for intrinsic motivation, not compliance.
Impact: Motivated learners show 3-5x higher retention.
Evidence Base: 562 papers on social learning
What Works:
What Doesn't Work:
Design Rule: Learning is inherently social—design for it.
Impact: Social learning increases completion by 2-3x.
Budget: €150,000
Deliverables:
Metrics:
Budget: €35,000
Deliverables:
Metrics:
Budget: €30,000
Deliverables:
Metrics:
| Metric | Traditional | Evidence-Based | Improvement |
|---|---|---|---|
| 30-Day Retention | 30% | 70%+ | +133% |
| Course Completion | 10% | 40%+ | +300% |
| Skill Transfer | 20% | 60%+ | +200% |
| Learner Satisfaction | 3.2/5 | 4.5/5 | +41% |
| ROI | 100-200% | 300-500% | +200% |
| Component | Cost (EUR) |
|---|---|
| Development (6 months) | €150,000 |
| Tools & Infrastructure | €20,000 |
| Content Creation | €30,000 |
| Training & Onboarding | €15,000 |
| Total | €215,000 |
| Year | Learners | Revenue | ROI |
|---|---|---|---|
| Year 1 | 500 | €450,000 | 109% |
| Year 2 | 1,500 | €1,350,000 | 528% |
| Year 3 | 3,000 | €2,700,000 | 1,158% |
Format: 2-day intensive or 4-week online
Price: €6,500 (onsite) or €897/person (online)
Outcome: Participants apply evidence-based techniques immediately
Curriculum Highlights:
Results:
The evidence is overwhelming: learning platforms built on research principles dramatically outperform traditional approaches. The seven principles identified in this synthesis—spaced repetition, active recall, adaptive personalization, immediate feedback, cognitive load management, motivation design, and social learning—are not optional features but foundational requirements.
Organizations that invest in evidence-based learning design can expect:
The research corpus of 13,204 papers provides the foundation. The question is no longer "does this work?" but "how quickly can you implement it?"
This synthesis draws from 13,204 research papers across 20 disciplines. Key sources include:
Full bibliography available upon request.
Tobias Weiss leads KI-Kompetenz-Training, specializing in evidence-based learning platform design. With expertise spanning cognitive science, machine learning, and instructional design, Tobias has analyzed over 13,000 research papers to create actionable frameworks for learning platform optimization.
Contact: ki-kompetenz-training@tobias-weiss.org
Website: www.ki-kompetenz-training.org
Published: July 31, 2026
Version: 1.0
License: See accompanying LICENSE file