Context
Teacher activation was Prodigy’s core growth lever — but it was stuck. The 7-day activation rate (teachers with at least 2 students answering questions within a week of signup) sat around 22% through Q1 2022. I ran a multi-quarter experimentation program to move it.
Over six months, my team tested contextual onboarding with images (+1.07% activation, student engagement gains downstream), onboarding with video (inconclusive — low watch rate, teachers signing up in classrooms can’t easily watch a video), and guided empty states with sample data (neutral, but revealed that where we land teachers matters less than what they can do when they get there).
Each experiment eliminated a hypothesis and sharpened the next one. The Home Screen Redesign was the payoff.
The Problem
The Teacher App homepage was getting traffic but doing nothing with it. Around 80% of teachers clicked to the homepage after landing on their dashboard — but fewer than 10% took any action once they got there.
The page had 12 CTAs pulling in 12 directions. It was supposed to surface product updates and relevant announcements, but without the resources to keep it fresh, the content had gone stale. Teachers arrived, found nothing useful, and left.
It was a banner graveyard.
What I Did
I led the experiment to redesign the homepage from the ground up. The hypothesis was simple: reduce the CTAs from 12 to 3, built around the three core actions teachers actually use Prodigy for.
Based on design research, those three actions were:
- Evaluating students — running the Placement Test to gauge grade level
- Assessing students — creating assignments, test preps, and plans
- Motivating students — sending classroom rewards and goals
Each card on the new homepage was dynamic rather than static, showing teachers live data about how their students were performing across each feature. Instead of stale announcements, the page became a real-time snapshot of their classroom.
The experiment ran for 14 days (Nov 24 – Dec 8, 2022) across ~18,700 teachers, split 50/50 between control and treatment.
What Happened
The hypothesis was validated with statistical significance.
| Metric | Control | Treatment | Change |
|---|---|---|---|
| Teachers clicking ≥1 CTA on homepage | 8.53% | 26.51% | +210% (stat sig) |
| Teachers clicking homepage again another day | 4.34% | 10.16% | +134% (stat sig) |
| Teachers creating assignments | 26.09% | 28.85% | +10.6% (stat sig) |
| Teachers sending rewards | 26.02% | 27.43% | +5.4% (stat sig) |
The 210% lift in homepage engagement wasn’t a novelty effect — the repeat engagement metric confirmed teachers were coming back to the page, not just clicking once out of curiosity.
Feature usage followed. Teachers in the treatment group were more likely to create assignments (+10.6%) and send classroom rewards (+5.4%), both statistically significant. In both groups, teachers who used these core features were associated with more active students (28 vs. 22) and more questions answered per student (48 vs. 39).
Projected downstream impact: Based on the relationship between teacher engagement, student practice time, and membership conversion observed across Prodigy’s funnel, the activation pattern from this redesign was estimated to contribute meaningfully to ARR growth over the following two quarters — though that attribution is modeled, not directly measured from this experiment.
What I Learned
The answer wasn’t more context — it was less friction. Six months of onboarding experiments tested images, videos, and guided empty states on the assumption that teachers weren’t activating because they didn’t understand the product. None of those moved the needle significantly. The real problem was decision overload. Twelve CTAs on a homepage isn’t a content problem. It’s a focus problem.
Dynamic data beats static content. The control homepage had content teachers had already seen. The treatment homepage showed live classroom data — how many students were on the placement test, what assignments were pending. Relevance at the moment of visit matters more than the richness of the information.
Surface the right action at the right moment. Teachers who used the core features (assignments, rewards) had more active students — 28 vs. 22 — and students answering more questions on average (48 vs. 39). The homepage redesign didn’t just improve engagement with the page; it created a path to the features that drive real classroom outcomes.
The one open question: weekly return rate didn’t improve significantly. Activating teachers once is a solvable problem. Keeping them coming back is a different one — and became the next workstream.
Prodigy Education is a game-based math learning platform used by 100M+ students and teachers across North America.