Real school and university innovation programmes rarely adopt a new pedagogical method or classroom technology in one leap. They run it through a stage-gate process: an idea is proposed, built into a small prototype, tested as a limited classroom pilot, and — only if it clearly works — scaled and institutionalized across the school or district. At every gate, a review board decides whether the idea advances or is shelved.
Each gate is modelled as an independent Bernoulli trial with its own pass probability. The chance a fresh idea survives the whole pipeline is the product of the three gate probabilities:
P(institutionalized) = p1 · p2 · p3
p1 = P(Idea → Prototype)
p2 = P(Prototype → Pilot)
p3 = P(Pilot → Scale-up)
This is why innovation funnels look so lopsided: even three "coin-flip-ish" 50% gates in a row only let 12.5% of ideas through, which is exactly why real institutions submit far more ideas than they expect to keep — a low per-idea success rate is normal, not a sign of failure.
- New ideas per year — how many proposals enter the funnel; a higher intake rate feeds more attempts through the same gates.
- Gate 1–3 pass rates — how strict each review stage is; tightening an early gate (Gate 1) filters cheaply, while tightening a late gate (Gate 3) wastes the effort already spent on piloting.
- Avg. time to scale — the mean simulated time (in years) an idea takes from submission to full institutionalization, averaged only over ideas that actually made it.
Real-world relevance: this is the same funnel logic behind corporate R&D stage-gate systems (Cooper's model) and behind why education systems that want more classroom innovations either widen the funnel mouth (more pilots) or loosen an early, cheap-to-run gate rather than the expensive final one.