Navigating the EdTech Landscape in 2026: What’s Evidence-Backed and What’s Just Expensive
There are more than 10,000 educational technology products on the market right now. Every week, EdTechHub’s landscape report documents new ones launching. Each claims to improve outcomes, reduce workload, or solve some specific gap in the education pipeline. School administrators who evaluate these tools for a living describe it as “the most confusing procurement decision in tech” — harder than enterprise software because the evidence is thinner and the stakes are someone’s child’s education.
A 2024 analysis by the EdTech Evidence Exchange found that only 12% of widely used EdTech products had published independent studies with control groups measuring actual learning outcomes. The other 88% have case studies, testimonials, and proprietary internal data — none of which tells you whether the product caused any learning improvement, or just got adopted by schools that were already well-run.
This matters because the tools look similar and the outcomes aren’t. A student using Anki for spaced repetition — based on decades of cognitive science — and a student using a generic flashcard app that doesn’t adapt difficulty are using products that appear identical but produce very different results over six months of study.
How to evaluate EdTech before buying
The Every Student Succeeds Act (ESSA) created a rough quality ladder for EdTech research evidence. Tier 1 requires randomised controlled trial evidence. Tier 2 requires quasi-experimental evidence. Tier 3 requires correlational evidence. Tier 4 requires a plausible logic model for why the product might work. Most EdTech products sit at Tier 4 or unclassified. Products with Tier 1–2 evidence — Khan Academy’s math interventions, DreamBox Learning, Read 180 — are notable partly because they’re uncommon.
For individual students and parents choosing tools without access to procurement databases, the practical heuristic is simpler: does this product embody evidence-based learning principles? Spaced repetition, active recall, immediate corrective feedback, worked examples, and deliberate difficulty progression are the mechanisms with the strongest research support. Gamification, social features, AI chat interfaces, and animated avatars may improve engagement, but engagement and learning are not the same thing. A product built around the former without the latter is producing enjoyment, not education.
The best tools by learning goal
| Learning goal | Evidence-backed tool | Why it works | Cost |
|---|---|---|---|
| K–12 maths | Khan Academy / DreamBox Learning | Adaptive difficulty, mastery progression, immediate feedback | Free / school licence |
| Reading and literacy | Lexia Core5 / Read 180 | Structured literacy, decodable text, phonics support with evidence | School licence |
| Language learning | Duolingo Max / Babbel | Spaced repetition, AI conversation practice, daily habit design | Free–$30/month |
| Coding and CS | Scratch / Code.org / Codecademy | Project-based, scaffolded difficulty, immediate execution feedback | Free–$20/month |
| Test prep | Khan Academy SAT / PrepScholar | Official College Board partnership, adaptive diagnostic, score tracking | Free–$397 |
The coding education expansion — and its blind spot
Code.org’s Hour of Code has reached over 100 million students. Scratch is used in more than 100 countries. State mandates requiring CS education have created a booming market for curriculum and teacher training. Coding education in K–12 has gone from optional enrichment to mainstream requirement in under a decade.
The quality problem hiding inside this success: most coding education focuses on syntax acquisition rather than computational thinking. A student who learns Python commands from a tutorial app but hasn’t grappled with how to decompose a complex problem into logical steps, manage ambiguity, or debug systematically hasn’t learned the transferable skill. The syntax is the least important part. CS50 at Harvard (free on edX) and the AP Computer Science A curriculum are standouts precisely because they treat problem-solving as the subject and syntax as the medium. Most commercial coding apps have the ratio reversed.
Self-directed learning: the options are genuinely good now
For adults learning independently, or students supplementing formal education, the 2026 landscape is remarkably strong. Khan Academy covers elementary through college-level material rigorously and for free. Coursera and edX offer university courses from MIT, Stanford, and Yale for free auditing, with paid certificates. YouTube has genuine experts teaching almost anything at depth. AI tutors are available 24 hours a day, can explain anything at whatever level you need, and will keep going until you’ve understood. A motivated learner with an internet connection in 2026 can access a better education than most universities provided twenty years ago.
The gap is structure. About 80% of people who start MOOCs don’t finish them. Self-directed learning drifts without deadlines, external evaluation, and community. The setups that work combine free online content with some accountability structure — a study group, a cohort programme, a mentor, or at minimum a public commitment with real consequences for backing out. The content problem is solved. The motivation and accountability problem is where self-directed learners still tend to fail, and technology hasn’t solved it in any reliable way.
The short version for anyone buying EdTech tools
Look for published independent research, not case studies. Try the free tier before paying a subscription. Prioritise products built around learning science principles over products with impressive-looking AI features. And be sceptical of any product whose primary evidence of effectiveness is its own marketing materials — that’s true for consumer apps, and it’s especially true when you’re choosing tools for a child’s education.
