The Value Gap: Why EdTech Unicorns Fail Learners
How Perceived Value Built a $10B Industry That Teaches Nothing
There’s a peculiar phenomenon in EdTech: companies are worth billions while their users learn almost nothing. This isn’t a bug in the system. It’s the system working exactly as designed.
The gap between what learning feels like and what learning actually is has created an entire industry optimised for the wrong outcome.
Understanding this gap explains why your Duolingo streak means nothing, why that $2,000 coding bootcamp didn’t land you a job, and why EdTech valuations keep climbing while educational outcomes flatline.
The Experience Economy Meets Education
Joseph Pine and James Gilmore’s seminal work on the experience economy revealed something crucial: people don’t buy products or services—they buy experiences. In entertainment, hospitality, and retail, this insight revolutionised business models.
EdTech took note. But education has a fatal quirk that makes the experience economy dangerous: the best learning experiences often feel terrible in the moment.
As cognitive scientist Daniel Willingham has extensively documented, effortful thinking is inherently unpleasant. The struggle that creates durable learning (what Robert and Elizabeth Bjork call “desirable difficulties”) feels like failure to the learner. Smooth, pleasant experiences that generate immediate satisfaction are precisely the conditions that undermine long-term retention and transfer.
This creates a toxic dynamic: EdTech companies can grow faster by making learning feel valuable rather than making it be valuable.
The Two-System Problem
Daniel Kahneman’s distinction between System 1 (fast, intuitive) and System 2 (slow, effortful) thinking illuminates why this gap persists. Learners use System 1 to evaluate whether they’re learning. It feels good, therefore I’m learning. It feels hard, therefore it’s not working.
But actual skill acquisition happens almost entirely in System 2. As Anders Ericsson’s research on deliberate practice demonstrates, expertise requires focused attention on weaknesses, immediate feedback, and repeated attempts to do things just beyond current capability. None of this feels good.
Learners, making System 1 judgments about System 2 processes, consistently mistake fluency for mastery. This is what Philip Guo at UC San Diego calls the “illusion of competence”: the deadly gap between feeling like you understand something and actually being able to apply it.
EdTech platforms can exploit this gap deliberately or accidentally, but the commercial incentive is identical: maximise perceived value while minimising the cost of delivering experienced value.
The Metrics That Betray Learning
Justin Reich at MIT has spent years studying why digital learning interventions consistently fail to scale learning outcomes despite scaling user bases beautifully. His conclusion is damning: EdTech companies measure what’s easy to measure, not what matters.
Engagement metrics like time on platform, completion rates, daily active users are the currency of venture capital and public markets. These metrics capture perceived value brilliantly. They say nothing about capability development.
As Ben Williamson at University of Edinburgh argues, the datafication of education has created what he calls “computational governance,” where platforms shape behaviour toward what can be measured rather than what should be learned. When retention depends on engagement rather than mastery, the rational business decision is to optimise for the former.
Neil Selwyn takes this further, arguing that EdTech’s obsession with efficiency and scale fundamentally misunderstands education as a human practice. Learning requires time, struggle, relationship, and failure.
None of these scale elegantly or show up in monthly active user charts.
The AI Acceleration
Large language models have turbocharged this problem. They can generate explanations so immediately satisfying that they create unprecedented perceived value while potentially destroying experienced value.
As Sherry Turkle warned in her work on technology and human connection, we’re increasingly vulnerable to substituting simulation for authentic experience. AI tutors can simulate the experience of understanding without requiring the cognitive effort that creates actual understanding.
Ethan Mollick at Wharton has been tracking this in real-time with students using ChatGPT. He observes that AI can make students feel more confident while making them less capable, what he calls “the delegation trap.” Students experience the value of having problems solved for them without experiencing the value of solving problems themselves.
The Valuation Paradox
Here’s where it gets darkly fascinating: the worse EdTech is at delivering experienced value, the faster it can grow, at least initially.
The jobs-to-be-done framework helps explain this. Learners aren’t hiring EdTech products to learn; they’re hiring them to feel like they’re learning. They’re hiring them to check a box, satisfy a social obligation, or reduce anxiety about falling behind.
Audrey Watters has been documenting this for over a decade. EdTech, she argues, is more about surveillance, credentialing, and the performance of learning than about learning itself. Platforms that deliver on the performance while skipping the learning capture more market share, not less.
The venture capital model amplifies this. A company with 50 million users learning nothing is worth more than a company with 50,000 users developing genuine capability.
The Retention Cliff
But here’s the problem: perceived value without experienced value has a shelf life.
BJ Fogg’s behaviour model shows that behaviour requires motivation, ability, and trigger in the same moment. EdTech companies are excellent at providing triggers (notifications, streaks, gamification). They’re decent at lowering ability barriers (smooth onboarding, bite-sized content). But they systematically fail to build real ability, which means motivation eventually dies.
This is why EdTech companies show spectacular growth curves followed by catastrophic retention problems. The gap between perceived and experienced value can be papered over with marketing, social proof, and sunk cost fallacy but only for so long.
The Signal Design Challenge
Herminia Ibarra’s work on professional identity offers an unexpected insight. People don’t change because they learn new information; they change because they get new evidence about who they are and what they can do.
This is the crux of the value gap: perceived value is about information and feeling. Experienced value is about identity and capability.
For experienced value to overtake perceived value as a growth driver, EdTech needs to make capability development visible and social. Learners need evidence they can share with themselves and others that they’ve genuinely changed.
The Accountability Moment
We’re approaching an inflection point. The industrial model of education (and by extension, EdTech) is finally colliding with the demands of a world that values creativity, adaptability, and authentic capability over credentials and completion certificates.
AI makes this collision inevitable and immediate. When anyone can generate a perfect essay or explanation, the credential economy collapses. The only thing that matters is what you can actually do.
This creates a ruthless market dynamic: EdTech companies can no longer hide behind information delivery as a proxy for learning. The gap between perceived and experienced value becomes obvious the moment learners try to apply what they “learned.”
The Path Forward
Several researchers offer provocative solutions:
Make transfer visible: Michelle Miller’s work on teaching and learning suggests that platforms should require learners to demonstrate capability in novel contexts, not just within the app’s controlled environment.
Embrace productive failure: Manu Kapur’s research shows that struggling with problems before receiving instruction produces deeper learning than smooth, scaffolded experiences. EdTech needs to reward struggle, not hide it.
Build social proof around capability: Reuben Lim’s work on badging and credentials suggests that social recognition needs to attach to demonstrated capability, not time spent or content consumed.
Align revenue with outcomes: John Hattie’s meta-analyses on learning effectiveness point toward a brutal truth: what works for learning often looks nothing like what scales commercially. Until business models reward actual capability development, the gap persists.
The Uncomfortable Truth
The EdTech decacorn that teaches nothing isn’t an accident.
It’s the logical endpoint of a system that rewards perceived value and can’t measure experienced value.
The companies that survive the next decade won’t be those with the smoothest AI tutors or the most engaging gamification. They’ll be the ones whose learners can demonstrably do things they previously could not and whose business models depend on that outcome.
The gap between perceived and experienced value is an arbitrage opportunity for founders and a moral hazard for society. AI eliminates the excuses for not closing it.
The question is whether we’ll choose to.
Kate Busby is Founder of Vibe Combinator, CoFounder of Quiet Edge and a Fractional CMO based in Barcelona, Spain, catch her on X and Linkedin. Subscribe to Substack to receive all articles on AI, edtech and impact accountability straight to your inbox.